 
Last updated on June 19, 2018. This conference program is tentative and subject to change
Technical Program for Thursday June 14, 2018

ThA1 
Demetra 
Agents and Autonomous Robots 
Regular Session 
Chair: Frego, Marco  Univ. Di Trento 
CoChair: Pereira, Fernando Lobo  Porto Univ 

10:0010:20, Paper ThA1.1  
Efficient Re–planning for Robotic Cars 
Bertolazzi, Enrico  Univ. of Trento  PI IT00340520220 
Bevilacqua, Paolo  Univ. of Trento 
Biral, Francesco  Univ. of Trento 
fontanelli, Daniele  Univ. of Trento 
Frego, Marco  Univ. Di Trento 
Palopoli, Luigi  Univ. Di Trento 
Keywords: Autonomous robots, Automotive, Transportation systems
Abstract: We consider the problem of the reactive replanning of an optimal trajectory for autonomous vehicles subject to geometric and dynamic constraints. Reactive replanning is used when a vehicle following a planned trajectory encounters an unforeseen obstacle. In such a case, a new local trajectory that avoids the obstacle has to be generated, without violating any constraint and preserving optimality. The solution presented in the paper guarantees that the new trajectory rejoins the previously planned one shortly after the obstacle. Moreover, the transition between old and new trajectory is smooth up to second derivative (curvature), which makes it easy to track an emergency manoeuvre. Finally, our solution is efficient and can be implemented in realtime on lean hardware. In order to validate the approach, we show how the replanning can be executed in a few milliseconds (on a standard machine) for the realistic example of a racing car.


10:2010:40, Paper ThA1.2  
A Hybrid Systems Model Predictive Control Framework for AUV Motion Control 
Gomes, Rui  Porto Univ 
Pereira, Fernando Lobo  Porto Univ 
Keywords: Autonomous robots, Decentralized control, Cooperative autonomous systems
Abstract: A computationally efficient architecture to control formations of Autonomous Underwater Vehicles (AUVs) is presented and discussed in this article. The proposed control structure enables the articulation of resources optimization with state feedback control while keeping the onboard computational burden very low. These properties are critical for AUVs systems as they operate in contexts of scarce resources and high uncertainty or variability. The hybrid nature of the controller enables different modes of operation, notably, in dealing with unanticipated obstacles. Optimization and feedback control are brought in by a novel Model Control Predictive (MPC) scheme constructed in such a way that timeinvariant information is used as much as possible in a priori offline computation.


10:4011:00, Paper ThA1.3  
Autonomous Trajectory Design System for Mapping of Unknown SeaFloors Using a Team of AUVs 
Salavasidis, Georgios  National Oceanography Centre 
Kapoutsis, Athanasios  Democritus Univ. of Thrace 
Chatzichristofis, Savvas  Democritus Univ. of Thrace 
Michailidis, Panagiotis  Centre for Res. and Tech. Hellas 
Kosmatopoulos, Elias  Democritus Univ. of Thrace and CERTH, Greece 
Keywords: Autonomous robots, Iterative learning control, Cooperative control
Abstract: This research develops a new online trajectory planning algorithm for a team of Autonomous Underwater Vehicles (AUVs). The goal of the AUVs is to cooperatively explore and map the ocean seafloor. As the morphology of the seabed is unknown and complex, standard nonconvex algorithms perform insufficiently. To tackle this, a new simulationbased approach is proposed and numerically evaluated. This approach adapts the Parametrized Cognitivebased Adaptive Optimization (PCAO) algorithm. The algorithm transforms the exploration problem to a parametrized decisionmaking mechanism whose realtime implementation is feasible. Upon that transformation, this scheme calculates offline a set of decisionmaking mechanism’s parameters that approximate the – nonpractically feasible  optimal solution. The advantages of the algorithm are significant computational simplicity, scalability, and the fact that it can straightforwardly embed any type of physical constraints and system limitations. In order to train the PCAO controller, two morphologically different seafloors are used. During this training, the algorithm outperforms an unrealistic optimalonestepahead search algorithm. To demonstrate the universality of the controller, the most effective controller issued to map three new morphologically different seafloors. During the latter mapping experiment, the PCAO algorithm outperforms several gradientdescentlike approaches.


11:0011:20, Paper ThA1.4  
Underwater Coverage with a Mobile Robot of Limited Control Authority 
Aranda, Ernesto  UNED 
Cortes, Jorge  Univ. of California, San Diego 
Guinaldo, Maria  UNED 
Dormido, Sebastián  UNED 
Keywords: Coverage control, Optimization, Maritime
Abstract: This work considers the coverage of underwater areas with a mobile robot with constrained control and communication capabilities. While underwater, the robot can control its depth but it is subject to flow in the other directions. While on the surface, it can move (essentially) freely. The aim of the work is the coverage of the areas with the minimum waste of resources. For that, we propose a twopart algorithm, where one part is a genetic algorithm and the other part is an algorithm based on Netwton's method. Numerical simulations are provided to illustrate the efficiency of the algorithm.


11:2011:40, Paper ThA1.5  
SelfLocalization of Anonymous Mobile Robots from Aerial Images 
POULET, Olivier  Normandie Univ. Le Havre, LITIS 
Guinand, Frederic  Normandie Univ. Le Havre Univ. LITIS 
Guerin, François  Univ. Le Havre 
Keywords: Autonomous robots, Robotics
Abstract: This paper presents three methods for Anonymous mobile robots localization within a global frame. An aerial camera takes, at regular time intervals, pictures of the area in which robots are moving. The camera determines the coordinates of each robot. Each robot receives the whole set of coordinates extracted from each picture. Mobile Robots are all identical, they do not have any identifier and they can neither communicate with each other nor they can detect themselves. The first localization method is based on the analysis of the angular variation between two images. The second method relies on the analysis of the distances stemmed from three successive pictures. The last one determines if there exists an orientation allowing a specific robot to travel the path between two successive positions. A simulation plateform using augmented reality and the multirobot software PlayerStage are presented. This plateform is used for validating the different localization methods. Tests and results are presented and compared.


11:4012:00, Paper ThA1.6  
Distributed Constraint Optimization for Continuous Mobile Sensor Coordination 
Fransman, Jeroen  TU Delft 
Sijs, Joris  TNO 
De Schutter, Bart  Delft Univ. of Tech 
Dol, Henry  TNO 
Theunissen, Erik  NLDA 
Keywords: Agents and autonomous systems, Robotics, Distributed control
Abstract: Distributed Constraint Optimization Problem (DCOP) is a framework which can represent naturally distributed problems. However, it only allows discrete values for the decision variables. This limits its application for real world problems that can be modelled as Multi Agent Systems (MAS). In this paper, an extension of DCOP is investigated to handle variables with continuous domains. Additionally, an iterative anytime algorithm based on Distributed Pseudo tree Optimization Procedure (DPOP) is presented. The algorithm iteratively samples the search space in order to handle problems which are restricted by time and memory limitations. The performance of the algorithm is examined through a mobile sensor coordination problem. It was found that compared to DPOP with uniform sampling of the continuous domains, the proposed algorithm outperforms in both resource requirement as in performance.


ThA2 
Ares 
Microgrids and Electrical Power Systems 
Regular Session 
Chair: PRODAN, Ionela  Grenoble Inst. of Tech. (Grenoble INP)  Esisar 
CoChair: Shetty, Akhil  Univ. of California, Berkeley 

10:0010:20, Paper ThA2.1  
Optimal Energy Reserve Procurement 
Shetty, Akhil  Univ. of California, Berkeley 
Li, Sen  Univ. of California at Berkeley 
Poolla, Kameshwar  Univ. of California at Berkeley 
Varaiya, Pravin P.  Univ. of California at Berkeley 
Keywords: Energy systems, Emerging control applications, Optimization
Abstract: Uncertainties from renewables and demands create power imbalances in realtime electricity markets. This paper studies the problem of procuring reserve services in forward capacity markets from diverse resources to cover imbalance signals et. We consider the reserve procurement problem in two scenarios: (a) et reveals itself causally, (b) et is revealed all at once by an oracle. Each case induces an optimal resource procurement cost. The ratio between the costs in these two cases is defined as the price of causality. It captures the additional procurement cost from not knowing the entire imbalance signal in advance. An upper bound on the price of causality is derived, and the exact price of causality is computed in some special cases. The algorithmic basis for these computations is set containment linear programming. A mechanism is proposed to allocate the procurement cost to agents that contribute to the aggregate imbalance signal. This allocation is fair, budgetbalanced, and respects the cost causation principle. Our results are validated through simulation studies, where we explore the dependence of the price of causality on unit resource prices.


10:2010:40, Paper ThA2.2  
A Novel Design of Maximum Power Point/Droop Controllers for Photovoltaic Sources in DC Microgrids 
Krommydas, Konstantinos  Univ. of Patras 
Alexandridis, Antonios  Univ. of Patras 
Keywords: Energy systems, Decentralized control, Emerging control applications
Abstract: In this paper, we consider the challenging problem of operating photovoltaic (PV) sources in direct current (dc) microgrids at maximum power point (MPP) under normal conditions and in droop control operation when a sudden disconnection from the main grid occurs. Based on some recently proposed control design results, we present an augmented cascaded control scheme with improved stability properties that fulfills the following: 1) drives the PV sources energy production at MPP when the dc microgrid is connected to the main grid, 2) achieves accurate power sharing among the PV sources when the dc microgrid is disconnected from the main grid, 3) needs no communication between the PV sources, 4) automatically implements the transition between the MPP and the power sharing mode in a smooth, autonomous and seamless manner. In order to prove that the aforementioned aims are satisfied, a rigorous stability analysis is carried out on the accurate dynamic closedloop system. Furthermore, a tuning method is proposed for the complete control scheme in order to achieve the desired dynamical responses under the time scale separation principle between the cascaded controlloops. Finally, the excellent performance of the control strategy is validated through simulation results and in comparison with the conventional switching mode controller response.


10:4011:00, Paper ThA2.3  
Dissipated Energy Minimization for an ElectroMechanical Elevator of a DC Microgrid 
PHAM, Thanh Hung  LCIS Lab 
PRODAN, Ionela  Grenoble Inst. of Tech. (Grenoble INP)  Esisar 
GenonCatalot, Denis  LCIS Lab 
Lefèvre, Laurent  Grenoble Inst. of Tech. (Grenoble INP) 
Keywords: Energy systems, Constrained control, Optimization
Abstract: This paper extends some previous work on the dissipated energy minimization for an elevator system of a DC microgrid using a combination between differential flatness for the reference generation and predictive control for the reference tracking. The contribution of the present work resides in the validation of constraints at all times thanks to the effective use of Bsplines parameterization properties. The proposed improvements are validated through simulation and comparison results for a particular electromechanical elevator system.


11:0011:20, Paper ThA2.4  
Distributed Coordination of EnergyStorage Capacities in Virtual Microgrids 
Brehm, Robert Wolfgang  Univ. of Southern Denmark, Mads Clausen Inst. SDU Mech 
Ramezani, Hossein  Univ. of Southern Denmark, Mads Clausen Inst. SDU Mech 
Jouffroy, Jerome  Univ. of Southern Denmark 
Keywords: Cooperative control, Energy systems, Mechatronics
Abstract: An approach for distributed coordinated scheduling of storage capacities is presented. When storage capacities are connected in the same grid, behind a common transformer substation, mutual charging and discharging can occur, which can be prevented by the herein introduced coordination method. When cooperation is incorported, storage capacities can be operated as a virtual microgrid. The cooperation between nodes is based on the formulation of a simple objective function for coordination. The cooperation objective is then combined with each node's local objective, which is the increase of self consumption, such that load is released off the the grid. A qualitative reflection on the practical use of three distributed algorithms, to solve the formulated optimisation problem is provided. The Jacobi algorithm is qualified to be preferable for largescale networks with a great number of nodes, and the GaussSeidel algorithm is preferable when less nodes cooperate. To illustrate the concept and show the effect of coordination for prevention of mutual charging/discharging of storage capacities in a VMG, two comparative casestudy scenario are presented.


11:2011:40, Paper ThA2.5  
Voltage Stabilization in a DC MicroGrid by an ISSLike Lyapunov Function Implementing Droop Control 
Iovine, Alessio  Efficacity 
Damm, Gilney  Lab. IBISC  CNRS/Evry Univ 
De Santis, Elena  Univ. of L'Aquila 
Di Benedetto, M. Domenica  Univ. of L'Aquila 
GALAI DOL, Lilia  Efficacity 
Pepe, Pierdomenico  Univ. of L' Aquila 
Keywords: Electrical power systems, Lyapunov methods, Stability of nonlinear systems
Abstract: The interconnection of renewable energy sources with storage systems through a Direct Current (DC) MicroGrid is one of the best ways to deal with the need for reliable power networks. An approach based on the "System of Systems" philosophy using distributed control methodologies is developed here with the purpose to share the duty to ensure grid voltage stability among a number of devices ensuring fast response. The closedloop stability proof is based on an InputtoStateStable (ISS) like Lyapunov function.


11:4012:00, Paper ThA2.6  
Microgrids Aggregation Management Providing Ancillary Services 
La Bella, Alessio  Pol. Di Milano 
Farina, Marcello  Pol. Di Milano 
Sandroni, Carlo  RSE 
Scattolini, Riccardo  Pol. Di Milano 
Keywords: Electrical power systems, Optimization algorithms, Distributed control
Abstract: The electrical grid is facing a significant shift from a centralized generation system to a more distributed setting where each portion of the grid is managed by a local control system and it can work both as a producer and as a consumer. This brings many advantages, but a cooperation mechanism between the different players needs to be established in order to ensure the overall network proper working conditions. This paper presents an optimization framework for the management of an aggregation of microgrids. The main objective is to show that, if properly coordinated, microgrids can give a significant support in terms of ancillary services provision. Finally, a distributed algorithm is described, designed in such a way each microgrid preserves its internal information and the control of its resources.


ThA3 
Aphrodite + Hermes 
Embedded Optimization Algorithms for Predictive Control and Estimation I 
Invited Session 
Chair: Patrinos, Panagiotis  KU Leuven 
CoChair: Kvasnica, Michal  Slovak Univ. of Tech. in Bratislava 
Organizer: Quirynen, Rien  Mitsubishi Electric Res. Lab. (MERL) 
Organizer: Patrinos, Panagiotis  KU Leuven 
Organizer: Diehl, Moritz  AlbertLudwigsUniv. Freiburg 

10:0010:20, Paper ThA3.1  
RealTime Proximal Gradient Method for Linear MPC (I) 
Van Parys, Ruben  KU Leuven 
Pipeleers, Goele  KU Leuven, LRD 
Keywords: Predictive control for linear systems, Optimal control, Constrained control
Abstract: This paper presents a realtime implementation of the proximal gradient method (PGM) in a model predictive control (MPC) setting. In each control update only one iteration of the PGM is performed, while a next update is warmstarted using the solution of the previous one. When applied to linear timeinvariant (LTI) systems with simple input constraints, the resulting control law becomes extremely simple and offers possibilities to obtain fast control rates even on resourceconstrained hardware. The paper provides a proof of closedloop stability of the realtime PGM applied to LTI systems. A numerical simulation example validates the resulting closedloop performance.


10:2010:40, Paper ThA3.2  
Double Moving Horizon Estimation: Linearization by a Nonlinear Transformation (I) 
Abdollahpouri, Mohammad  Chalmers Univ. of Tech 
Haring, Mark  NTNU 
Johansen, Tor Arne  Norweigian Univ. of Sci. & Tech 
Takács, Gergely  Slovak Univ. of Tech. in Bratislava, Faculty Ofmechani 
RohalIlkiv, Boris  STU in Bratislava 
Keywords: Observers for nonlinear systems, Optimal control, Linear timevarying systems
Abstract: Moving horizon estimation (MHE) is a constrained nonconvex optimization problem in principle, which needs to be solved online. One approach to avoid dealing with several local minima is to linearize the nonlinear dynamics. This type of convex approximation usually utilizes the estimated state as a linearization trajectory, providing no guarantees of stability and optimality in general. In this paper, we study the cascade of a linear and linearized observer, which is called double MHE. The first stage makes use of a model transformation, that in the nominal case is globally equivalent to the nonlinear dynamics. Since this approach does not consider the input and output disturbances optimally, the second stage uses the first stage estimates as an external signal for linearizing the nonlinear dynamics to improve the quality of estimation. The overall configuration can be transformed into two quadratic programs. This approach not only avoids solving a nonconvex optimization problem, but also reduces the computational complexity significantly compared to the one needed for solving a nonconvex problem. This estimation method has been validated in a simulation study, where our approach converged to the global minimum without the need to explicitly solve a nonconvex optimization problem.


10:4011:00, Paper ThA3.3  
Block Structured Preconditioning within an ActiveSet Method for RealTime Optimal Control (I) 
Quirynen, Rien  Mitsubishi Electric Res. Lab. (MERL) 
Knyazev, Andrew  Mitsubishi Electric Res. Labs (MERL) 
Di Cairano, Stefano  Mitsubishi Electric Res. Lab 
Keywords: Optimization algorithms, Predictive control for linear systems
Abstract: Model predictive control (MPC) requires solving a blockstructured optimal control problem at each sampling instant. We propose an iterative preconditioned solver with computational cost that scales linearly with the number of intervals and quadratically with the number of state and control variables, and can be efficiently implemented on embedded hardware for realtime optimal control. Blockstructured factorizations and lowrank updates are combined with blockdiagonal preconditioning within a primal activeset strategy (PRESAS). Multiple numerical tests using our preliminary C implementation demonstrate competitiveness with the stateoftheart, as illustrated on an ARM CortexA53 processor.


11:0011:20, Paper ThA3.4  
Plug and Play Distributed Model Predictive Control with Dynamic Coupling: A Randomized PrimalDual Proximal Algorithm (I) 
Latafat, Puya  KU Leuven 
Bemporad, Alberto  IMT Inst. for Advanced Studies Lucca 
Patrinos, Panagiotis  KU Leuven 
Keywords: Distributed cooperative control over networks, Predictive control for linear systems, Optimization
Abstract: This paper proposes an algorithm for distributed model predictive control that is based on a primaldual proximal algorithm developed recently by two of the authors. The proposed scheme does not require strong convexity, involves one round of communication at every iteration and is fully distributed. In fact, both the iterations and the stepsizes are computed using only local information. This allows a plug and play implementation where addition or removal of a subsystem only affects the neighboring nodes without the need for global coordination. The proposed scheme enjoys a linear convergence rate. In addition, we provide a randomized variant of the algorithm in which at every iteration subsystems wake up randomly independent of one another. Numerical simulations are performed for the frequency control problem in a power network, demonstrating the attractive performance of the new scheme.


11:2011:40, Paper ThA3.5  
Distributed Control Algorithm for Vehicle Coordination at Traffic Intersections (I) 
Shi, Jiahe  ShanghaiTech Univ 
Zheng, Yi  ShanghaiTech Univ 
Jiang, Yuning  ShanghaiTech Univ 
Zanon, Mario  IMT Inst. for Advanced Studies Lucca 
Hult, Robert  Chalmers Univ. of Tech 
Houska, Boris  Univ. of Heidelberg 
Keywords: Traffic control, Predictive control for nonlinear systems, Distributed control
Abstract: This paper proposes a distributed closedloop control algorithm for optimal coordination of autonomous vehicles at traffic intersections. The main contribution of the paper is a distributed plugandplay closedloop optimal control scheme with rearend collision avoidance constraints enforced on each lane, which maintains recursive feasibility under the assumption that communication between neighboring vehicles is possible. The method is closely related to model predictive control, but at each sampling time new vehicles are allowed to enter the modeled region around the intersection while other vehicles are leaving. In contrast to human drivers, autonomous vehicles can collaboratively form the deceleration strategy before the intersection. Our numerical results indicate that, under certain assumptions, it is optimal for vehicles not allowed to directly pass the intersection to slow down much before the intersection area and then accelerate at the right time so that they can travel through the intersection faster.


11:4012:00, Paper ThA3.6  
Explicit MPC Based on Approximate Dynamic Programming (I) 
Bakarac, Peter  Slovak Univ. of Tech. in Bratislava 
Holaza, Juraj  Slovak Univ. of Tech. in Bratislava 
Kaluz, Martin  Slovak Univ. of Tech. in Bratislava 
Klauco, Martin  Slovak Univ. of Tech. in Bratislava 
Löfberg, Johan  Linköping Univ 
Kvasnica, Michal  Slovak Univ. of Tech. in Bratislava 
Keywords: Predictive control for linear systems
Abstract: In this paper we show how to synthesize simple explicit MPC controllers based on approximate dynamic programming. Here, a given MPC optimization problem over a finite horizon is solved iteratively as a series of problems of size one. The optimal cost function of each subproblem is approximated by a quadratic function that serves as a costtogo function for the subsequent iteration. The approximation is designed in such a way that closedloop stability and recursive feasibility is maintained. Specifically, we show how to employ sumofsquares relaxations to enforce that the approximate costtogo function is bounded from below and from above for all points of its domain. By resorting to quadratic approximations, the complexity of the resulting explicit MPC controller is considerably reduced both in terms of memory as well as the online computations. The procedure is applied to control an inverted pendulum and experimental data are presented to demonstrate viability of such an approach.


ThA4 
Poseidon 
Delay Systems I 
Regular Session 
Chair: Prieur, Christophe  CNRS 
CoChair: Barreau, Matthieu  Laas / Cnrs 

10:0010:20, Paper ThA4.1  
On HyperExponential OutputFeedback Stabilization of a Double Integrator by Using Artificial Delay 
Efimov, Denis  Inria 
Fridman, E. M.  TelAviv Univ 
Perruquetti, Wilfrid  Ec. Centrale De Lille 
Richard, JeanPierre  Ec. Centrale De Lille 
Keywords: Delay systems, Stability of nonlinear systems
Abstract: The problem of outputfeedback stabilization of a double integrator is revisited with the objective of achieving the rates of convergence faster than exponential. It is assumed that only position is available for measurements, and the designed feedback is based on the output and its delayed values without an estimation of velocity. It is shown that by selecting the closedloop system to be homogeneous with negative or positive degree it is possible to accelerate the rate of convergence in the system at the price of a small steadystate error. Efficiency of the proposed control is demonstrated in simulations.


10:2010:40, Paper ThA4.2  
Input Shaping for Infinite Dimensional Systems with Application on Oil Well Drilling 
Pilbauer, Dan  Univ. Grenoble Alpes, CNRS, Grenoble INP, GIPSALab 
BreschPietri, Delphine  MINES ParisTech 
Di Meglio, Florent  MINES ParisTech 
Prieur, Christophe  CNRS 
Vyhlidal, Tomas  Czech Tech. Univ. in Prague 
Keywords: Delay systems, Distributed parameter systems, Flexible structures
Abstract: We present an application of the input shaping technique on infinite dimensional systems such as systems described by PDEs. The presented method is a feedforward scheme which allows to target multiple modes of a flexible system and also provide robustness with respect to parameters variations. The method is based on recently developed input shaper with multiple degrees of freedom that are needed to meet all constraints for the given task. We show that, even though the system consists of infinitely many modes, it is required to target only dominant ones. In addition, the method is illustrated on a case example of oil well drilling.


10:4011:00, Paper ThA4.3  
Distributed SampledData Control of KuramotoSivashinsky Equation under the Point Measurements 
Kang, Wen  Tel Aviv Univ 
Fridman, E. M.  TelAviv Univ 
Keywords: Delay systems, Distributed parameter systems
Abstract: We consider sampleddata distributed control of nonlinear PDE system governed by KuramotoSivashinsky equation under point measurements and distributed in space shape functions. It is assumed that the sampling intervals in time and in space are bounded. We derive sufficient conditions ensuring local exponential stability of the closedloop system in terms of Linear Matrix Inequalities (LMIs) by using LyapunovKrasovskii method. Moreover, we give a bound on the domain of attraction. As it happened in the case of heat equation, the timedelay approach to sampleddata control and the descriptor method appeared to be efficient tools for the stability analysis of the sampleddata KuramotoSivashinsky equation. The efficiency of the results is demonstrated by a numerical example.


11:0011:20, Paper ThA4.4  
Static State and Output Feedback Synthesis for TimeDelay Systems 
Barreau, Matthieu  Laas / Cnrs 
Gouaisbaut, Frederic  Laas Cnrs 
Seuret, Alexandre  LAASCNRS 
Keywords: Delay systems, Output feedback, LMI's/BMI's/SOS's
Abstract: In this paper, the design of a static feedback gain for a linear system subject to an input delay is studied. This synthesis is based on a stability analysis conducted using LyapunovKrasovskii theorem and BesselLegendre inequalities. It is expressed in terms of linear matrix inequalities. Some bilinear nonconvex matrix inequalities are obtained to go from analysis to synthesis. They are then difficult to solve and an iterative procedure is given which takes advantage of the elimination lemma. Naturally, slack variables are introduced and values coming from an optimization process are proposed to reduce the conservatism. The two main corollaries discuss the static state and output feedback synthesis. Finally, a comparison is proposed and shows that this formulation introduces small conservatism.


11:2011:40, Paper ThA4.5  
Computing the Distance to Instability for Delay Systems with Uncertainties in the System Matrices and in the Delay Terms 
Borgioli, Francesco  KU Leuven 
Michiels, Wim  KU Leuven 
Keywords: Delay systems, Robust control, Stability of linear systems
Abstract: In this paper we propose an algorithm to compute the distance to instability of a linear system of delay differential equations (DDEs) containing uncertainties in the delay terms as well as in matrices coefﬁcients. For what regards the system matrices, any structure on the perturbation can be considered in order to allow only speciﬁc parameters to change; moreover, realvalued matrix perturbations are taken into account. The algorithm relies on the computation of the pseudospectral abscissa of the system and performs a bisectionNewton’s method to ﬁnd the minimum size of the perturbation that generates instability. A few illustrative examples, including a model for a rotating cutting machine, ﬁnally show the correctness and the efﬁciency of the method.


11:4012:00, Paper ThA4.6  
Stability Analysis of Systems with DelayDependent Coefficients Subject to Some Particular Delay Structure 
JIN, Chi  L2s, CentraleSupelec, Univ. ParisSud 
Gu, Keqin  Southern Illinois Univ. at Edwardsville 
Boussaada, Islam  IPSA & Lab. Des Signaux Et Systèmes 
Niculescu, SilviuIulian  Umr Cnrs 8506, CnrsSupelec 
Keywords: Delay systems, Stability of linear systems
Abstract: Stability of systems with commensurate delays and delaydependent coefficients is studied along the line of the taudecomposition approach. This particular delay structure allows the use of some sophisticated result from the matrix theory to generalize the stability analysis method developed for systems with a single delay. Criterion for determining cross directions of imaginary roots are presented, leading to a systematic stability analysis with the aid of the graphs of some functions.


ThA5 
Athenaeum 1 
Observers I 
Regular Session 
Chair: Raimondo, Davide Martino  Univ. of Pavia 
CoChair: Chong, Michelle Siu Tze  KTH Royal Inst. of Tech 

10:0010:20, Paper ThA5.1  
Bounding the L2 Sensitivity for Positive Linear Observers 
McGlinchey, Aisling  Maynooth Univ 
Mason, Oliver  Maynooth Univ 
Keywords: Observers for linear systems, Linear systems, Constrained control
Abstract: We consider the design of differentially private Luenberger observers for positive linear systems. In particular, we derive a bound for the l_2 sensitivity of Luenberger observers, which is used to quantify the noise required to achieve relaxed differential privacy via the Gaussian mechanism. An approach to minimise this bound for positive observers is described and several bounds relevant to this problem are derived.


10:2010:40, Paper ThA5.2  
Estimating the Wigner Distribution of Linear TimeInvariant Dynamical Systems 
Chong, Michelle Siu Tze  Lund Univ 
Sandsten, Maria  Lund Univ 
Rantzer, Anders  Lund Univ 
Keywords: Observers for linear systems, Signal processing
Abstract: An estimation algorithm for the Wigner distribution (timefrequency representation) of the unmeasured states of a linear timeinvariant system is presented. Given that the inputs and outputs are measured, the algorithm involves designing a Luenbergerlike observer for each frequency of interest. Under noisefree conditions, we show that the estimates converge to the true Wigner distribution under a detectability assumption on the timefrequency representation. The estimation algorithm provides estimates which converge to a neighbourhood of the true Wigner distribution where its norm is dependent on the norm of the measurement noise. We also illustrate the efficacy of the estimation algorithm on an academic example and a model of neuron populations.


10:4011:00, Paper ThA5.3  
Chattering Free High Order Sliding Mode Observer for Estimation of Liquid Water Fraction in a Proton Exchange Membrane Fuel Cell 
Luna, Julio  Chalmers Univ. of Tech 
Costa, Ramon  Univ. Pol. De Catalunya (UPC) 
Keywords: Observers for nonlinear systems, Energy systems, Sliding mode control
Abstract: In this work, a methodology for the estimation of the liquid fraction of a proton exchange membrane fuel cell (PEMFC) is developed. Specifically, a modelbased chatteringfree high order sliding mode (CHOSM) observer is designed for the estimation of two dynamic states: the PEMFC temperature and the liquid water saturation. The observation strategy is discussed in a simulation environment using the common ARTEMIS driving cycle (CADC) as a case study.


11:0011:20, Paper ThA5.4  
Design of HighGain ReducedOrder Observers for Nonlinear SampledData StrictFeedback Systems with Model Uncertainty 
Katayama, Hitoshi  Shizuoka Univ 
Keywords: Observers for nonlinear systems, Lyapunov methods, Stability of nonlinear systems
Abstract: Design of highgain reducedorder observers is considered for the nonlinear sampleddata strictfeedback system with model uncertainty. First highgain reducedorder observers are designed by introducing a small positive parameter to the reducedorder observers designed for the sampleddata strictfeedback system without model uncertainty. Then it is shown that there exists a small positive parameter such that the designed observers are semiglobal and practical in sampling period for the exact model of the sampleddata strictfeedback system with model uncertainty. A numerical example is given to show the efficiency of the designed observers.


11:4012:00, Paper ThA5.6  
Estimation of the Basin of Attraction of a Practical HighGain Observer 
Menini, Laura  Univ. Di Roma 'Tor Vergata' 
Possieri, Corrado  Pol. Di Torino 
Tornambe, Antonio  Univ. Di Roma Tor Vergata 
Keywords: Observers for nonlinear systems, Computational methods
Abstract: In this paper, a technique is given to estimate the basin of attraction of a “practical” high–gain observer. Such a goal is pursued by determining a set such that the restriction of the observability map of the system to such a set is a diffeomorphism. Differently from other techniques available in the literature, which are based on the numeric computation of the trajectories of a system corresponding to the observability map, the proposed procedure provides exact certificates of invertibility of the observability map. The application of the proposed method to the design of a nonlinear observer is discussed and validated through a numerical simulation.


ThA6 
Athenaeum 2 
Stability of Nonlinear Systems II 
Regular Session 
Chair: Mylvaganam, Thulasi  Imperial Coll. London 
CoChair: Gromov, Dmitry  St. Petersburg State Univ 

10:0010:20, Paper ThA6.1  
Dynamic Zero Finding for Algebraic Equations 
Mylvaganam, Thulasi  Imperial Coll. London 
Ortega, Romeo  Supelec 
Machado Martínez, Juan Eduardo  Lab. Des Signaux Et Systèmes (L2S)  CentraleSupélec 
Astolfi, Alessandro  Imperial Coll. London 
Keywords: Stability of nonlinear systems, Lyapunov methods, Nonlinear system theory
Abstract: In a variety of contexts, for example the solution of differential games and the control of power systems, the design of feedback control laws requires the solution of nonlinear algebraic equations: obtaining such solutions is often not trivial. Motivated by such situations we consider systems of nonlinear algebraic equations and propose a method for obtaining their solutions. In particular, a dynamical system is introduced and (locally) stabilizing control laws which ensure that elements of the state converge to a solution of the algebraic equations are given. Illustrative numerical examples are provided. In addition it is shown that the proposed method is applicable to determine the equilibria of electrical networks with constant power loads.


10:2010:40, Paper ThA6.2  
Global ClosedLoop Equilibrium Properties of a Geometric PDAV Controller Via a CoordinateFree Linearization 
Ramp, Michalis  National Tech. Univ. of Athens (NTUA) 
Papadopoulos, Evangelos  National Tech. Univ. of Athens 
Keywords: Stability of nonlinear systems, Modeling, Algebraic/geometric methods
Abstract: Tracking a desired Pointing Direction and simultaneously obtaining a reference Angular Velocity (PDAV) around the pointing direction constitutes a very involved and complicated motion encountered in a variaty of robotic, industrial and military applications. In this paper through the utilization of global analysis and simulation techniques, the smooth closedloop vector fields induced by the geometric PDAV controller from [1], are visualized to gain a deeper understanding of its global stabilization properties. First through the calculation of a coordinatefree form of the closedloop linearized dynamics, the local stability of each equilibrium of the system is analyzed. The results acquired by means of eigenstructure analysis, are used in predicting the frequency of complex precession/nutation oscillations that arise during PDAV trajectory tracking; an important tool in actuator selection. Finally, by utilizing variational integration schemes, the flow converging to the desired equilibrium and the flow "close" to the stable manifold of the saddle equilibrium of the closedloop system is visualized and analyzed. Results offer intimate knowledge of the closedloop vector fields bestowing to the control engineer the ability to anticipate and/or have a rough estimate of the evolution of the solutions.


10:4011:00, Paper ThA6.3  
Gaussian Process Based Passivation of a Class of Nonlinear Systems with Unknown Dynamics 
Beckers, Thomas  Tech. Univ. of Munich 
Hirche, Sandra  Inst. for InformationOriented Control 
Keywords: Stability of nonlinear systems, Machine learning, Stochastic systems
Abstract: The paper addresses the problem of passivation of a class of nonlinear systems where the dynamics are unknown. For this purpose, we use the highly flexible, datadriven Gaussian process regression for the identification of the unknown dynamics for feedforward compensation. The closed loop system of the nonlinear system, the Gaussian process model and a feedback control law is guaranteed to be semipassive with a specific probability. The predicted variance of the Gaussian process regression is used to bound the model error which additionally allows to specify the state space region where the closedloop system behaves passive. Finally, the theoretical results are illustrated by a simulation.


11:0011:20, Paper ThA6.4  
A Certificate of Global Asymptotic Stability for Planar Polynomial Systems 
Menini, Laura  Univ. Di Roma 'Tor Vergata' 
Possieri, Corrado  Pol. Di Torino 
Tornambe, Antonio  Univ. Di Roma Tor Vergata 
Keywords: Stability of nonlinear systems
Abstract: The goal of this paper is to give an algorithmic procedure for obtaining a certificate of global asymptotic stability for planar polynomial systems having the origin as equilibrium point. The procedure is not Lyapunov based and uses methods from Algebraic Geometry to study the sign of polynomial functions.


11:2011:40, Paper ThA6.5  
An Integral LineOfSight Guidance Law with a SpeedDependent Lookahead Distance 
Wiig, Martin Syre  Norwegian Univ. of Science and Tech 
Pettersen, Kristin Y.  Norwegian Univ. of Science and Tech 
Ruud, ElseLine Malene  FFI 
Krogstad, Thomas R.  Norwegian Defense Res. Establishment 
Keywords: Maritime, Stability of nonlinear systems, Nonlinear system theory
Abstract: This paper presents an algorithm that makes an underactuated marine vehicle follow a straight line path while in the presence of a constant ocean current. When following the path, the vehicle maintains a desired surge speed which is measured relative to the water, and which may be constant or timevarying. The algorithm is an integral lineofsight guidance law where the lookahead distance is designed to depend linearly on the desired relative surge speed of the vehicle. This dependency makes it possible to keep the maneuvering demands of the vehicle limited, even when the vehicle surge speed is large. It is shown that if the desired relative surge speed is constant along the path, the resulting error dynamics has a uniformly semiglobally exponentially stable equilibrium at the origin, thus achieving the path following and velocity control objectives. Furthermore, in the case of a general, timevarying desired speed trajectory, it is shown that the solutions of the system remain bounded. The results are supported by simulations, as well as experiments with an unmanned surface vehicle.


11:4012:00, Paper ThA6.6  
Projected Dynamics of Constrained Hamiltonian Systems 
Gromov, Dmitry  St. Petersburg State Univ 
Castanos, Fernando  Cinvestav Del IPN 
Fradkov, Alexander L.  Acad. of Sciences of Russia 
Keywords: Algebraic/geometric methods, Stability of nonlinear systems, Differential algebraic systems
Abstract: А novel formulation for the description of implicit portHamiltonian control systems is proposed and its potential use for design of the control laws stabilizing a given submanifold described as a zero level set of an admissible energy function is shown. Using the developed formulation, a number of results on the stabilization of portHamiltonian systems are presented. The obtained results are formulated in a way that allows for direct application.


ThA7 
Athenaeum 3 
Automotive and Traffic Control 
Regular Session 
Chair: Ferrara, Antonella  Univ. of Pavia 
CoChair: Velenis, Efstathios  Cranfield Univ 

10:0010:20, Paper ThA7.1  
A Feasible MPCBased Negotiation Algorithm for Automated Intersection Crossing 
Kneissl, Maximilian  DENSO AUTOMOTIVE Deutschland GmbH 
Molin, Adam  KTH Royal Inst. of Tech 
Esen, Hasan  DENSO AUTOMOTIVE Deutschland GmbH 
Hirche, Sandra  Inst. for InformationOriented Control 
Keywords: Traffic control, Automotive, Distributed control
Abstract: We propose an intersection crossing algorithm for autonomous vehicles with vehicle to infrastructure (V2I) communication capability. All vehicles attempting to cross the intersection share their expected time of entering a critical zone based on decentralized model predictive control (MPC) results. These time suggestions are collected at a central intersection management (IM) unit, which is responsible for coordinating the vehicles. A timebased negotiation process between vehicles and IM is conducted to find a safe solution. An advantage of the approach is that modelbased vehicle data is kept private, while the computational burden of the intersection coordination is distributed between the central IM and the vehicles. We prove the existence of a feasible solution and illustrate the introduced negotiation algorithm by simulation of an intersection crossing scenario with disturbances. The results show that vehicles remain in a safe distance without sharing private data.


10:2010:40, Paper ThA7.2  
A RiskConstrained and Energy Efficient Stochastic Approach for Autonomous Overtaking 
Ramezani, Zahra  Johannes Kepler Univ. Linz 
Gagliardi, Davide  Johannes Kepler Univ. Linz 
Del Re, Luigi  Johannes Kepler Univ. Linz 
Keywords: Automotive, Optimal control, Stochastic control
Abstract: In this paper a control approach for safe and energy efficient autonomous overtaking is presented. The proposed method combines a stochastic approach based on the use of safety indicators as the Time headway (TH) and Time to collision (TTC) as boundary conditions and a cost index on fuel consumption for the trajectory of an Ego vehicle in a traffic environment. The approach has been evaluated in two typical overtaking scenarios: 1) braking of the leading vehicle and 2) entry of a third vehicle in the gap between the leading and the ego vehicle. The efficiency of TH and TTC based stochastic approach have been compared for the same maximal risk threshold. Results show that both risk function are suitable to perform overtaking manoeuvres both with and without an additive fuel cost. Moreover the TTC–based stochastic approach shows a reduction of about 5% in fuel consumption with respect to the THbased approach for both considered scenarios.


10:4011:00, Paper ThA7.3  
TwoLayer Hierarchical Control for LargeScale Urban Traffic Networks 
Kouvelas, Anastasios  EPFL 
Triantafyllos, Dimitris  TSS  Transport Simulation Systems, S.L 
Geroliminis, Nikolas  Ec. Pol. Fédérale De Lausanne (EPFL), Urban Transport 
Keywords: Traffic control, Transportation systems, Intelligent systems
Abstract: Many efforts have been carried out to optimize the traffic signal settings in cities. Nevertheless, stateoftheart and practice strategies cannot deal efficiently with oversaturated conditions (i.e. queue spillbacks and partial gridlocks), as they are either based on applicationspecific heuristics or they fail to replicate accurately the propagation of congestion. An alternative approach for realtime networkwide control is the perimeter flow control (or gating). This can be viewed as an upperlevel control layer, and be combined with other strategies (e.g. local or coordinated regulators) in a hierarchical control framework. In the current work, a recently developed perimeter control regulator is utilized for the upperlevel layer. Another lowerlevel control layer utilizes the maxpressure regulator, which constitutes a local feedback control law, applied in coupled intersections, in a distributed systemsofsystems (SoS) concept. Different approaches are discussed about the design of the hierarchical structure of SoS and a traffic microsimulation tool is used to assess the impact of each approach to the overall traffic conditions. Preliminary results show that integrating a networklevel approach within a local adaptive framework can significantly improve the system performance when spillback phenomena occur (a common feature of city centres with short links).


11:0011:20, Paper ThA7.4  
Energy and TimeOptimal Connected Autonomous Vehicle Interaction: Cruising and Overtaking 
Stryszowski, Marcin  Cranfield Univ 
Longo, Stefano  Cranfield Univ 
Velenis, Efstathios  Cranfield Univ 
Bin Raja Ahsan Shah, Raja Mazuir  Arrival Limited 
Keywords: Traffic control, Transportation systems, Optimal control
Abstract: This paper considers the energy optimality aspect of Connected Autonomous Vehicle (CAV) interactions. The aim is to study optimal velocity profiles from the perspective of balance between cost of energy and cost of opportunity in platooning or overtaking scenario. As both costs contribute to the overall cost, the objective is to find optimal velocity profiles. In the proposed analysis a vehicle encounters another agent cruising at lower speed, and while the agents can platoon, the problem is set up to study an overtaking maneuver. Two scenarios are considered. A noncooperative one, where the overtakee ignores the overtaken, and a cooperative one where both agents are subject to a control input. Results suggest clues for development of the decisionmaking framework for optimal CAV operation, but also present the scale of improvement in traffic energy efficiency which the connectivity will enable. Such approach results in vehicles in traffic cruising at various velocities but can also offer average energy savings of 20.3% while performing an overtaking maneuver.


11:2011:40, Paper ThA7.5  
Towards a Robust Traffic Admission Control in Homogeneous Urban Vehicular Networks under QoS Constraints 
Csikós, Alfréd   
Charalambous, Themistoklis  Aalto Univ 
Kulcsar, Balazs  Chalmers Univ. of Tech 
Keywords: Traffic control, Transportation systems, Robust control
Abstract: In this paper, we consider the problem of controlling the input flow to a homogeneous urban vehicular network such that certain Quality of Service (QoS) constraints are preserved. In such a network, we model the system with two types of queues: external and internal. External queues represent vehicles waiting to enter the urban vehicular network under control, and the internal queue is used to describe the network’s aggregated behavior based on the Network Fundamental Diagram (NFD). While most of the works assume perfect knowledge of the NFD describing the urban vehicular network, in reality we can only approximate it. On these grounds, by taking a model of the NFD with uncertainties, we propose a robust control design approach in order to gate input flow to a protected urban vehicular network such that travel time Quality of Service (QoS) constraints are preserved within the network. The proposed controller is compared via simulations with controllers assuming perfect knowledge of the NFD and it is shown that it can provide a larger stability region.


11:4012:00, Paper ThA7.6  
Traffic Flow Inspired Analysis and Boundary Control for a Class of 2x2 Hyperbolic Systems 
Karafyllis, Iasson  National Tech. Univ. of Athens 
BekiarisLiberis, Nikolaos  Tech. Univ. of Crete 
Papageorgiou, Markos  Tech. Univ. of Crete 
Keywords: Distributed parameter systems, Nonlinear system theory, Traffic control
Abstract: The paper presents results for a class of 2x2 systems of nonlinear hyperbolic PDEs on a 1D bounded domain, inspired by secondorder traffic flow models. The model consists of two firstorder hyperbolic PDEs with a dynamic boundary condition that involves the time derivative of the velocity. The developed model has features that are important from a traffictheoretic point of view: is completely anisotropic and information travels forward exactly at the same speed as traffic. It is shown that, for all physically meaningful initial conditions, the model admits a globally defined, unique, classical solution that remains positive and bounded for all times. Furthermore, a nonlinear, explicit boundary feedback law is developed, which achieves global stabilization of arbitrary equilibria. The stabilizing feedback law depends only on the inlet velocity and consequently, the measurement requirements for the implementation of the proposed boundary feedback law are minimal. The efficiency of the proposed boundary feedback law is demonstrated by means of a numerical example.


ThA8 
Athenaeum 4 
Learning and Adaptivity in Predictive and Optimal Control: Theory,
Applications, and Future Perspectives I 
Invited Session 
Chair: Farina, Marcello  Pol. Di Milano 
CoChair: Fagiano, Lorenzo  Pol. Di Milano 
Organizer: Farina, Marcello  Pol. Di Milano 
Organizer: Fagiano, Lorenzo  Pol. Di Milano 

10:0010:20, Paper ThA8.1  
HumansInTheLoop: A GameTheoretic Perspective on Adaptive Building Energy Systems (I) 
Eichler, Annika  ETH Zurich 
Darivianakis, Georgios  ETH 
Lygeros, John  ETH Zurich 
Keywords: Game theoretical methods, Adaptive control, Predictive control for linear systems
Abstract: Efficient building energy management has attracted a great interest in diverse research areas due to significant potential energy savings. A remaining challenge is how to combine the efforts of engineers on improving the energy management system with approaches developed by social scientists to integrate the occupants actively in the energy management system. This paper proposes the formulation of a game between the building energy management system and occupants to agree on room temperature comfort bounds. Under mild assumptions on the cost functions of the occupants, we show that a generalized Nash equilibrium exists and it can be shown to equal the social optimum. The alternating direction method of multipliers is used to solve the resulting consensus optimization problem in a distributed way, with the building management system as coordinator. An advantage of the proposed method is that the building energy management system does not need to rely on an explicit model of the occupant behavior but due to the game theoretic approach, indirectly receives an adapted model at each iteration. An extensive numerical study demonstrates the efficacy of the proposed approach.


10:2010:40, Paper ThA8.2  
Impact of Occupancy Modeling and Horizon Length on HVAC Controller Efficiency (I) 
Garaza, Christian Rabbi  Univ. of the Philippines Diliman 
Hespanhol, Pedro  Univ. of California Berkeley 
Mintz, Yonatan  UC Berkeley 
Pedrasa, Jhoanna Rhodette  Univ. of the Philippines Diliman 
Aswani, Anil  Univ. of California, Berkeley 
Keywords: Energy systems, Predictive control for linear systems, Modeling
Abstract: Controlling heating, ventilation, and air conditioning (HVAC) to improve its energy efficiency or implement ancillary services (e.g., demand response or frequency regulation) requires compensating for occupants and the computers/equipment they use, which significantly impact the thermal dynamics of buildings. Several studies have explored the use of either binary models or online learning for occupancy within model predictive control (MPC) frameworks, but the use of more finegrained predictive models for occupancy has been less wellstudied. This paper uses real data from an HVAC testbed to develop predictive models of occupancy and heating load, evaluate the predictive accuracy of the occupancy model, and then quantify its efficacy in achieving energy reductions in conjunction with a learningbased MPC (LBMPC) controller. We conclude by conducting a simulation study to determine the optimal horizon for an LBMPC controller. In particular, we find there is a tradeoff with increasing the horizon length between decreasing prediction accuracy of occupancy heating load and the ability of the MPC to better anticipate thermal dynamics due to occupancy and weather.


10:4011:00, Paper ThA8.3  
Learning MultiStep Prediction Models for Receding Horizon Control (I) 
Terzi, Enrico  Pol. Di Milano 
Fagiano, Lorenzo  Pol. Di Milano 
Farina, Marcello  Pol. Di Milano 
Scattolini, Riccardo  Pol. Di Milano 
Keywords: Identification for control, Identification
Abstract: In this paper, the derivation of multistepahead prediction models from sampled inputoutput data of a linear system is considered. Specifically, a dedicated prediction model is built for each future time step of interest. Each model is linearly parametrized in a suitable regressor vector, composed of past output values and past and future input values. In addition to a nominal model, the set of all models consistent with data and prior information is derived as well, making the approach suitable for robust control design within a Model Predictive Control framework. The resulting parameter identification problem is solved through a sequence of convex programs. Convergence of the identified error bounds to their theoretical minimum is demonstrated, under suitable assumptions on the measured data, and features like worstcase accuracy computation are illustrated in a numerical example.


11:0011:20, Paper ThA8.4  
Cautious NMPC with Gaussian Process Dynamics for Miniature Race Cars (I) 
Hewing, Lukas  ETH Zürich 
Liniger, Alexander  ETH Zürich 
Zeilinger, Melanie N.  ETH Zurich 
Keywords: Predictive control for nonlinear systems, Machine learning, Stochastic control
Abstract: This paper presents an adaptive highperformance control method for autonomous miniature race cars. Racing dynamics are notoriously hard to model from first principles, which is addressed by means of a cautious nonlinear model predictive control (NMPC) approach that learns to improve its dynamics model from data and safely increases racing performance. The approach makes use of a Gaussian Process (GP) and takes residual model uncertainty into account through a chance constrained formulation. We present a sparse GP approximation with dynamically adjusting inducing inputs, enabling a realtime implementable controller. The formulation is demonstrated in simulations, which show significant improvement with respect to both laptime and constraint satisfaction compared to an NMPC without model learning.


11:2011:40, Paper ThA8.5  
Constrained and Stabilizing Stacked Adaptive Dynamic Programming and a Comparison with Model Predictive Control (I) 
Beckenbach, Lukas  Chemnitz Univ. of Tech 
Osinenko, Pavel  Tech. Univ. Chemnitz 
Göhrt, Thomas  Tech. Univ. Chemnitz 
Streif, Stefan  Tech. Univ. Chemnitz 
Keywords: Adaptive control, Predictive control for nonlinear systems, Stability of nonlinear systems
Abstract: Model predictive control (MPC) is in many applications the de facto approach to optimal control. It typically provides an optimal input (sequence) for a finitehorizon of given running costs. Another approach, called dynamic programming (DP), is based on the HamiltonJacobiBellman formalism and usually seeks optimal inputs over an infinite horizon of running costs. Unlike MPC, DP is much less computationally tractable and typically requires state space discretization which leads to the socalled curse of dimensionality. Adaptive dynamic programming (ADP), an approach based on reinforcement learning, seeks to address the difficulties of DP by introducing approximation models for the optimal cost function and control policies. In a variant of ADP called stacked ADP (sADP), control policies are optimized over a finite stack of value function approximants, thus making it somewhat similar to MPC. First, similarities and differences between a variant of ADP and MPC are discussed. Second, MPC stability results are transferred to ADP and state and input constraints are considered. The work is concluded by a case study.


11:4012:00, Paper ThA8.6  
Nonlinear Reference Tracking with Model Predictive Control: An Intuitive Approach 
Köhler, Johannes  Univ. of Stuttgart 
Muller, Matthias A.  Univ. of Stuttgart 
Allgower, Frank  Univ. of Stuttgart 
Keywords: Predictive control for nonlinear systems, Constrained control, Output regulation
Abstract: In this paper, we study the system theoretic properties of a reference tracking Model Predictive Control (MPC) scheme for general reference signals and nonlinear discretetime systems subject to input and state constraints. Contrary to other existing theoretical results for reference tracking MPC, we do not modify the optimization problem with artificial references or terminal ingredients. Instead, we consider a simple, intuitive, implementation and derive theoretical guarantees under a stabilizability property of the system and a reachability condition on the reference trajectory. We provide sufficient conditions for exponential reference tracking and analyze the region of attraction.


ThA9 
Hera 
Predictive Control IV 
Regular Session 
Chair: Bitmead, Robert  Univ. of California San Diego 
CoChair: Schulze Darup, Moritz  Univ. of Paderborn 

10:0010:20, Paper ThA9.1  
Tailored MPC for Mobile Robots with Very Short Prediction Horizons 
Schulze Darup, Moritz  Univ. of Paderborn 
Worthmann, Karl  Tech. Univ. Ilmenau 
Keywords: Predictive control for nonlinear systems, Constrained control, Algebraic/geometric methods
Abstract: We propose a novel predictive control algorithm tailored for the kinematic model of a mobile robot. The scheme exploits two properties of the robot: finite time controllability and the fact that any given path can be traversed with arbitrarily chosen time parametrization if input constraints are ignored. Combining these two properties allows us to design a terminal cost that guarantees recursive feasibility and set point convergence of the MPC closedloop  independently of the chosen prediction horizon. We illustrate the efficiency of the proposed controller with some numerical experiments.


10:2010:40, Paper ThA9.2  
Analytical Tuning of a TSMPC Dedicated to Nonlinear MIMO Systems 
Turki, Marwa  Irseem / Esigelec 
Benkhoud, Khaled  Univ. of Sciences of Monastir 
Langlois, Nicolas  Irseem / Esigelec 
Yassine, Adnan  NORMANDIE Univ. UNILEHAVRE LMAH 
Keywords: Predictive control for nonlinear systems, Optimal control, Robotics
Abstract: In this paper, an analytical tuning method for TakagiSugenomodelbased predictive control (TSMPC) is presented. As an advantage, it can be applied to nonlinear multiinputmultioutput (MIMO) controllable processes with constraints and guarantees closedloop stability while limiting required computational load. Its application to a simulated quadtiltwing (QTW) unmanned aerial vehicle (UAV) emphasizes its effectiveness when it is compared to a metaheuristic tuning approach.


10:4011:00, Paper ThA9.3  
Performance of Model Predictive Control of POMDPs 
Sehr, Martin Arno  UC San Diego 
Bitmead, Robert  Univ. of California San Diego 
Keywords: Predictive control for nonlinear systems, Predictive control for linear systems, Stochastic control
Abstract: We revisit closedloop performance guarantees for Model Predictive Control in the deterministic and stochastic cases, which extend to novel performance results applicable to receding horizon control of Partially Observable Markov Decision Processes. The general intractability of stochastic optimal control relaxes for this particular instance of stochastic systems, provided reasonable problem dimensions are taken. This motivates extending available performance guarantees to this particular class of systems, which may also be used to approximate general nonlinear dynamics via gridding of state, observation, and control spaces. We demonstrate applicability of the novel closedloop performance results on a particular example in healthcare decision making, which relies explicitly on the duality of the control decisions.


11:0011:20, Paper ThA9.4  
An Algorithm for MixedInteger Optimal Control of Solar Thermal Climate Systems with MPCCapable Runtime 
Bürger, Adrian  Karlsruhe Univ. of Applied Sciences 
Zeile, Clemens  OVGU Magdeburg 
AltmannDieses, Angelika  Karlsruhe Univ. of Applied Sciences 
Sager, Sebastian  OVGU Magdeburg 
Diehl, Moritz  AlbertLudwigsUniv. Freiburg 
Keywords: Predictive control for nonlinear systems, Switched systems, Energy systems
Abstract: This work presents an algorithm for solution of MixedInteger Optimal Control Problems (MIOCPs) for Solar Thermal Climate Systems (STCSs) with MPCcapable runtime. We implement the socalled Combinatorial Integral Approximation (CIA) algorithm for a model of an STCS of a building that incorporates an adsorption cooling machine and apply the algorithm within a numerical case study to solve a MixedInteger NonLinear Program (MINLP) resulting from an MIOCP for the system. We compare the results of the CIA algorithm to those of a general MINLP solver and show that our algorithm achieves comparable solution quality at a runtime that is up to 1000 times smaller.


11:2011:40, Paper ThA9.5  
ScenarioBased Optimal Control for Gaussian Process State Space Models 
Umlauft, Jonas  Tech. Univ. of Munich 
Beckers, Thomas  Tech. Univ. of Munich 
Hirche, Sandra  Inst. for InformationOriented Control 
Keywords: Stochastic control, Predictive control for nonlinear systems, Optimal control
Abstract: Datadriven approaches from machine learning provide powerful tools to identify dynamical systems with limited prior knowledge of the model structure. More particular, the Gaussian process state space model, a Bayesian nonparametric approach, is increasingly utilized in control. Its probabilistic nature is interpreted differently in the control literature, but so far, it is not considered as a distribution over dynamical system which allows a scenariobased control design. This paper introduces how scenarios are sampled from a Gaussian process and utilizes them in a differential dynamic programming approach to solve an optimal control problem. For the linearquadratic case, we derive probabilistic performance guarantees using results from robust convex optimization. The proposed methods are evaluated numerically for the nonlinear and linear case.


11:4012:00, Paper ThA9.6  
A Regularized Newton Solver for Linear Model Predictive Control 
Malyshev, Alexander  Univ. of Bergen 
Quirynen, Rien  Mitsubishi Electric Res. Lab. (MERL) 
Knyazev, Andrew  Mitsubishi Electric Res. Labs (MERL) 
Di Cairano, Stefano  Mitsubishi Electric Res. Lab 
Keywords: Optimization algorithms, Predictive control for linear systems
Abstract: We investigate direct numerical solvers in linear model predictive control, where the prediction model is given by linear systems subject to linear inequality constraints on the state and the input, and the performance index is convex and quadratic. The inequality constraints are treated by the primaldual interiorpoint method. We propose a novel direct solver based on the augmented Lagrangian regularization of a reduced Hessian. The new solver has the same arithmetic complexity as the factorized Riccati recursion. The direct solver can be implemented in terms of BLAS3 matrix operations.


ThA10 
The Grill Room 
Cooperative Systems I 
Regular Session 
Chair: Frasca, Paolo  Univ. of Twente 
CoChair: Aranda, Miguel  SIGMA Clermont, Inst. Pascal, Univ. Clermont Auvergne 

10:0010:20, Paper ThA10.1  
Cooperative Adaptive Cruise Control Over Unreliable Networks: An ObserverBased Approach to Increase Robustness to Packet Loss 
Acciani, Francesco  Univ. of Twente 
Frasca, Paolo  Univ. of Twente 
Stoorvogel, Anton A.  Univ. of Twente 
semsarkazerooni, Elham  TNO 
Heijenk, Geert  Univ. of Twente 
Keywords: Cooperative control, Agents and autonomous systems, Cooperative autonomous systems
Abstract: Cooperative Adaptive Cruise Control is nowadays a promising tool to increase highway throughput, safety and comfort for vehicles. Enabled by wireless communication, it allows a platoon of vehicles to achieve better performance than Adaptive Cruise Control, but since wireless is employed, problems related to unreliability arise. In this paper, we design a digital controller to achieve platoon stability, enhanced by an observer to increase robustness with regard to packet losses. A preliminary set of simulation results is presented, which confirms the interest of using an observer in combination with a local and cooperative digital controller.


10:2010:40, Paper ThA10.2  
Multirobot Target Enclosing with Freely Selected Observation Distances 
Aranda, Miguel  SIGMA Clermont, Inst. Pascal, Univ. Clermont Auvergne 
mezouar, youcef  Inst. Pascal 
Keywords: Cooperative control, Agents and autonomous systems, Robotics
Abstract: Tasks such as surveillance, escorting or robotic object interaction benefit from having complete and detailed perception of a target entity, provided by multiple mobile sensors. The ensuing challenge of suitably coordinating the sensor motions for such collective observation is addressed in this paper. In particular, we consider a moving target in 3D space, and focus on achieving prescribed relative viewing angles, which we encapsulate by a default desired enclosing pattern, with respect to this target. The paper contributes a novel control method that makes a team of mobile agents converge to the desired configuration of viewpoints with respect to a point that tracks the target. Relative agent position regulation and target tracking are integrated via a formationbased controller relying on global information that incorporates an optimal pattern rotation. We also introduce flexibility in the team geometry, allowing each agent to select freely, without knowledge of the others, its desired distance to the target. This can allow to, e.g., optimize perception quality and avoid collisions. Notably, we show that even with these distributed adjustments, the team motions remain steady, which contributes to obtaining stable perception and efficient control performance. In addition, each robot can operate on its independent local reference frame. Simulation tests illustrate the presented methodology.


10:4011:00, Paper ThA10.3  
Verification of Cooperative Maneuvers in FlightGear Using MPC and Backwards Reachable Sets 
Persson, Linnea  KTH Royal Inst. of Tech 
Wahlberg, Bo  KTH Royal Inst. of Tech 
Keywords: Cooperative control, Autonomous systems, UAV's
Abstract: In this paper we develop a simulation setup for testing and analyzing cooperative maneuvers and corresponding control algorithms. We also find feasible initial sets using backwards reachable set computations for the cooperative control problem, which we then test using the simulation setup. The particular example considered is a cooperative rendezvous between a fixedwing unmanned aerial vehicle and a unmanned ground vehicle. The opensource software FlightGear and JSBSim are used for the vehicle dynamics, enabling testing of algorithms in a realistic environment. The aircraft models include nonlinear, statedependent dynamics, making it possible to capture complex behaviors like stall and spin. Moreover, environmental effects such as wind gusts and turbulence are directly integrated into the simulations. From the simulations we get a comprehensive understanding of the controller performance and feasibility when tested in a realtime scenario. Results from several landing simulations are presented, and demonstrate that the MPC solution for the cooperative rendezvous problem is a promising method also for use in complex, safetycritical systems.


11:0011:20, Paper ThA10.4  
Distributed Model Predictive Control of Multiple Vehicles Transporting a Flexible Payload 
Mulagaleti, Sampath Kumar  IMT School of Advanced Studies Lucca 
Van Parys, Ruben  KU Leuven 
Pipeleers, Goele  KU Leuven, LRD 
Keywords: Cooperative control, Distributed cooperative control over networks, Optimal control
Abstract: This paper presents a control strategy for multiple vehicles that cooperatively transport a flexible payload. To this end, an algorithm is developed which generates optimal trajectories for the vehicles to follow. Solving an optimization problem composes the core of the algorithm. The problem is first decomposed over the vehicles using the Alternating Direction Method of Multipliers (ADMM) algorithm. This results in each vehicle solving a subproblem to generate its own optimal trajectory. The algorithm instructs that the optimization problem be solved repeatedly in a receding horizon fashion, making it fit into a distributed model predictive control (DMPC) framework. One ADMM iteration is performed per DMPC iteration, reducing the interagent communication rate. Numerical validation of the developed control scheme is performed and the results are presented.


11:2011:40, Paper ThA10.5  
Consensus Speed of Static Pinning Consensus Control of MultiAgent Systems 
Sakaguchi, Akinori  Osaka Univ 
Ushio, Toshimitsu  Osaka Univ 
Keywords: Cooperative control, Largescale systems, Concensus control and estimation
Abstract: This paper is concerned with the pinning consensus problem of multiagent systems with single integrator agents in general directed networks. In this problem, the consensus speed is critical for the analysis of the convergence rate of the system. We show that the upper bound of the consensus speed of static pinning control is the eigenvalue with the smallest real part of the matrix reduced by deleting rows and columns corresponding to indexes of pinned agents from the Laplacian matrix. We also investigate the limit of eigenvalues of the controlled multiagent system as the control gains go to infinity. In examples, we discuss the two cases where the network topologies are a directed smallscale and a directed scalefree graph.


11:4012:00, Paper ThA10.6  
A GroupBased Navigation Algorithm for Multiple AntiShip Missile Systems 
Fabregas, Marc Sans  Cranfield Univ 
Lee, ChangHun  Cranfield Univ 
Tsourdos, Antonios  Cranfield Univ 
Keywords: Cooperative autonomous systems, Cooperative control, Sensor and signal fusion
Abstract: In this paper, feasibility of cooperative navigation for antiship missile systems is investigated with the purpose of reducing cost and complexity of each missile system in a cooperative mission. We consider a homing engagement scenario of intercepting a highvalue target using multiple missile systems equipped with low cost GPSbased navigation systems without onboard seekers. In order to ensure precise interception performance, the ultimate goal of proposed concept is to improve poor navigation performance of each missile system using the concept of cooperative navigation with additional measurements provided by simple and cheap sensors. In the proposed concept, the local GPS information is shared through communication. We then determine the estimated relative geometry such as relative range and relative angle between the elements in the group based on the shared GPS information. By comparing the estimated values with the additional measurements of relative range and angle, the navigation solution of each missile system is improved. In this study, the proposed concept is realized by means of the multirate Extended Kalman Filter (EKF). The performance and the feasibility of proposed method are validated through comparing Fisher Information Matrix (FIM) and numerical interception simulations.


ThB1 
Demetra 
Agent Networks 
Regular Session 
Chair: Anderson, Brian D.O.  Australian National Univ 
CoChair: Schenato, Luca  Univ. of Padova 

14:3014:50, Paper ThB1.1  
A Generalized DiscreteTime Altafini Model 
Wang, Lili  Yale Univ 
Liu, Ji  Stony Brook Univ 
Morse, A. Stephen  Yale Univ 
Anderson, Brian D.O.  Australian National Univ 
Fullmer, Daniel  Yale Univ 
Keywords: Agents networks, Concensus control and estimation, Largescale systems
Abstract: A discretetime modulus consensus model is considered in which the interactions between the members of a networked family of n agents is described by a timedependent gain graph whose vertices correspond to agents and whose arcs are assigned complex numbers from a prescribed cyclic group. Limiting behavior of the model's state is studied using a graphical approach. It is shown that a certain type of clustering of agents' ``opinions'' or states will be reached exponentially fast for almost all initial conditions if and only if the sequence of gain graphs is ``repeatedly jointly structurally balanced'' corresponding to the type of clustering being reached, where the number of clusters is at most the order of the prescribed cyclic group. It is also shown that the agents' states will all converge to zero asymptotically if the sequence of gain graphs is repeatedly jointly strongly connected and structurally unbalanced. In the special case when the cyclic group is of order two, the model simplifies to the socalled Altafini model whose gain graph is simply a signed graph.


14:5015:10, Paper ThB1.2  
Realization of Homogeneous MultiAgent Networks 
Szabo, Zoltan  Mta Sztaki 
Bokor, Jozsef  Hungarian Acad. of Sciences 
Hara, Shinji  The Univ. of Tokyo 
Keywords: Agents networks, Control over networks, Modeling
Abstract: A class of largescale systems with decentralized information structures, such as multiagent systems, can be represented by a linear system with a generalized frequency variable. In these models agents are modelled through a strictly proper SISO state space model while the supervisory structure, representing the information exchange among the agents, is represented via a linear statespace model. The starting point of the paper is that the agent h(s) and the overall system mathcal{G}(s) are known through their Markov parameters. Based on these data a condition is given that characterizes compatibility, i.e., the existence of a transfer function G(s) that describes the network and leads to the relation mathcal{G}(s)=G(frac{1}{h(s)}). If compatibility holds, the paper also presents an algorithm to compute the Markov parameters of the unknown transfer function G(s). Then, a minimal state space representation of this transfer function can be computed through the HoKalman algorithm.


15:1015:30, Paper ThB1.3  
A Suite of Distributed Methodologies to Solve the Sparse Analytic Hierarchy Process Problem 
Menci, Marta  Univ. Campus Biomedico 
Oliva, Gabriele  Univ. Degli Studi Roma Tre 
Papi, Marco  Univ. Campus Bio Medico 
Setola, Roberto  Univ. Campus BioMedico of Rome 
Scala, Antonio  CNR 
Keywords: Agents networks, Distributed cooperative control over networks, Distributed estimation over sensor nets
Abstract: In this paper we aim at finding effective distributed algorithms to solve the Sparse Analytic Hierarchy Process (SAHP) problem, where a set of networked agents (e.g., wireless sensors, mobile robots or IoT devices) need to be ranked based on their utility/importance. However, instead of knowing their absolute importance, the agents know their relative utility/importance with respect to their neighbors. Moreover, such a relative information is perturbed due to errors, subjective biases or incorrect information. Recently, the Sparse Eigenvector Method proved its effectiveness in tackling this problem. However, such a method has several drawbacks, such as demanding computation/communication requirements and lack of control on the magnitude of the computed estimate. With the aim to mitigate such issues, in this paper we inspect the possibility to resort to a suite of different methodologies, each inspired to well known algorithms in the literature, i.e., MetropolisHastings Markov chains, HeatBath Markov chains and formation control. The proposed methodologies are less demanding in terms of memory and communication capabilities; however, each approach has its own strength points and drawbacks. The aim of this paper is thus to provide a numerical comparison of their performances over networks with different characteristics.


15:3015:50, Paper ThB1.4  
Designing Communication Topologies for Optimal Synchronization Trajectories of Homogeneous Linear MultiAgent Systems 
Hermann, Jonathan  Tech. Univ. Darmstadt 
Bernhard, Sebastian  Tech. Univ. Darmstadt 
Konigorski, U.  Tech. Univ. Darmstadt 
Adamy, Juergen  Tech. Univ. Darmstadt 
Keywords: Agents networks
Abstract: In this paper the synchronization of homogeneous linear multiagent systems is considered. In this process, all agents are required to converge to a common trajectory called synchronization trajectory. Without synchronization every single agent would have followed its own autonomous trajectory determined by its initial state. Hence, we interpret the initial states of the agents as their respective preferences. We define a cost function that penalizes the compromise each agent has to make when the synchronization trajectory differs from the agent's preferred trajectory. Using classical synchronization controllers, the synchronization trajectory essentially depends on the communication topology and the initial states of the agents. Therefore, we pose optimization problems concerned with finding the communication topology yielding minimal cost, i.e. the synchronization trajectory that constitutes an optimal compromise for all agents. In this respect, we minimize the cost for a given initial state, for the worstcase as well as for the average over all admissible initial states of the agents. The introduced minimization problems are reformulated as semidefinite programs so that they can be efficiently solved. A numerical example illustrates the results.


15:5016:10, Paper ThB1.5  
MultiAgent Distributed Optimization Algorithms for PartitionBased Linear Programming (LP) Problems 
Carli, Ruggero  Univ. Di Padova 
Yildirim, Kasim Sinan  Delft Univ. of Tech 
Schenato, Luca  Univ. of Padova 
Keywords: Optimization algorithms, Agents networks, Largescale systems
Abstract: The paper addresses the problem of multiagent distributed solutions for a class of linear programming (LP) problems which include box constraints on the decision variables and inequality constraints. The major difference with existing literature on distributed solution of LP problems is that each agent is expected to compute only a single or few entries of the global minimizer vector, often referred as a partitionbased optimization. This class of LP problems is relevant in different applications such as optimal power transfer in remotely powered batteryless wireless sensor networks, minimum energy LED luminaries control in smart offices, and optimal temperature control in start buildings. Via a suitable approximation of the original LP problem, we propose three different primaldual distributed algorithms based on dual gradient ascent , on the methods of multipliers and on the Alternating Direction Methods of Multipliers. We discuss the computational and communication requirements of these methods and we provide numerical comparisons.


16:1016:30, Paper ThB1.6  
Modeling Wireless Power Transfer in a Network of Smart Devices: A Compartmental System Approach 
Fontan, Angela  Linköping Univ 
Altafini, Claudio  Univ. of Linkoping 
Keywords: Network analysis and control, Agents networks, Stability of linear systems
Abstract: Wireless power transfer technology provides a possible sustainable and costeffective way to prolong indefinitely the lifetime of networks of smart devices needed in future InternetofThings, while equipping them with batteries of limited capacity. In this paper we show that the theory of compartmental systems, positive interconnected systems exchanging 'mass' (here power) and ruled by mass conservation laws, provides a suitable framework to describe wireless power transfer networks. In particular we show that sustainability of the network of smart devices corresponds to each of them being alimented, directly or indirectly, by nodes having an external source of power, condition known as inflow connectivity in the compartmental systems literature. The framework allows to compute the topology which is optimal in terms of maximizing the overall efficiency of the power transfer.


ThB2 
Ares 
Control Over Networks 
Regular Session 
Chair: Allgower, Frank  Univ. of Stuttgart 
CoChair: Schenato, Luca  Univ. of Padova 

14:3014:50, Paper ThB2.1  
Performance Oriented Triggering Mechanisms with Guaranteed Traffic Characterization for Linear DiscreteTime Systems 
Linsenmayer, Steffen  Univ. of Stuttgart 
Allgower, Frank  Univ. of Stuttgart 
Keywords: Control over communication, Linear systems, Optimal control
Abstract: This paper is concerned with transmission reduction in linear discretetime Networked Control Systems with computational capabilities at the sensor only. A triggering mechanism is derived to reduce the number of transmissions while being oriented at the control performance, measured by an infinite horizon linear quadratic cost functional. It is shown how trigger functions can be designed and parametrized such that an a priori demanded performance bound will not be violated. Furthermore the general approach is used to derive a triggering mechanism with guaranteed characterization of the resulting network traffic. Therefore a model for network traffic, known from traffic shaping in communication networks, is presented and a corresponding trigger function is designed. All resulting mechanisms are evaluated in a numerical example.


14:5015:10, Paper ThB2.2  
HeavyTails in Kalman Filtering with Packet Losses: Confidence Bounds vs Second Moment Stability 
Dey, Subhrakanti  Uppsala Univ 
Schenato, Luca  Univ. of Padova 
Keywords: Control over communication, Stochastic filtering
Abstract: In this paper, we study the existence of a steady state distribution and its tail behaviour for the estimation error arising from Kalman filtering for unstable scalar systems. Although a large body of literature has studied the problem of Kalman filtering with packet losses in terms of analysis of the second moment, no study has addressed the actual distribution of the estimation error. By drawing results from Renewal Theory, in this work we show that under the assumption that packet loss probability is smaller than unity, and the system is on average contractive, a stationary distribution always exists and is heavytailed, i.e. its absolute moments beyond a certain order do not exist. We also show that under additional technical assumptions, the steady state distribution of the Kalman prediction error has an asymptotic powerlaw tail, i.e. with an exponent that can be explicitly computed. We further explore how to optimally select the sampling period assuming exponential decay of packet loss probability with respect to it, in order to minimize the expected value of second moment or the confidence bounds showing that in general a larger sampling period will need to be chosen in the latter case as a result of the heavy tail behaviour.


15:1015:30, Paper ThB2.3  
Stability Analysis on the Networked MultiAgent System 
Lee, HaeIn  Cranfield Univ 
Shin, HyoSang  Cranfield Univ 
Tsourdos, Antonios  Cranfield Univ 
Keywords: Control over networks, Agents networks
Abstract: In this paper, a new stability analysis method on the networked multiagent system is proposed. The main idea is to model the networked control system as a multioutputmultiinput transfer function and to apply the Nyquist criterion for the closedloop stability. This paper provides two application examples: formation control of first and secondorder agents. The stability conditions are derived with the proposed method, showing the effect of network connectivity with multihop communication. The numerical simulations are conducted to verify the analysis. This work is applicable for general agent dynamics, controllers, and communication characteristics.


15:3015:50, Paper ThB2.4  
Stabilization of MIMO Systems Over Additive Correlated Noise Channels Subject to Multiple SNRConstraints 
González, Rodrigo A.  Univ. Técnica Federico Santa María 
Vargas, Francisco J.  Univ. Técnica Federico Santa María 
Chen, Jie  City Univ. of Hong Kong 
Keywords: Control over networks, Linear systems
Abstract: This paper studies the mean square stabilization of MIMO discretetime linear timeinvariant systems over a MIMO additive correlated channel. We assume that such channel consists of multiple correlated SISO channels subject to independent input signaltonoise ratio (SNR) constraints. We derive explicit conditions for which mean square stabilization can be achieved under such constraints for unstable minimum phase plants, and characterize the controller that achieves such SNR. We also present the set of admissible SNR constraints for mean square stability for a particular set of plants. Our results show that noise correlation can reduce the SNR requirements for stability compared to independent additive white noise channels. In addition, a numerical simulation is provided to illustrate the theoretical results.


15:5016:10, Paper ThB2.5  
EventTriggered Control for DiscreteTime Nonlinear Systems Using StateDependent Riccati Equation 
Kishida, Masako  National Inst. of Informatics 
Keywords: Control over networks, Nonlinear system theory, Optimal control
Abstract: Motivated by the need for the resourceefficient control in networked control systems, this paper considers eventtriggered control approaches to discretetime nonlinear controlaffine systems. In particular, we use the discretetime statedependent Riccati equation along with eventtrigger conditions to obtain control laws that satisfy 1) a stability condition and 2) a performance cost condition, respectively, while reducing the frequency of control input updates. A numerical example of the inverted pendulum is included to illustrate the proposed approach.


16:1016:30, Paper ThB2.6  
InformationConstrained Optimal Control of Distributed Systems with Power Constraints 
Causevic, Vedad  Tech. Univ. of Munich 
Ugo Abara, Precious  Tech. Univ. of Munich 
Hirche, Sandra  Inst. for InformationOriented Control 
Keywords: Control over networks, Optimal control, Constrained control
Abstract: In this paper we address the problem of informationconstrained optimal control for an interconnected system subject to onestep communication delays and power constraints. The goal is to minimize a finitehorizon quadratic cost by optimally choosing the control inputs for the subsystems, accounting for power constraints in the overall system and different information available at the decision makers. To this purpose, due to the quadratic nature of the power constraints, the LQG problem is reformulated as a linear problem in the covariance of stateinput aggregated vector. The zeroduality gap allows us to equivalently consider the dual problem, and decompose it into several subproblems according to the information structure present in the system. Finally, the optimal control inputs are found in a form that allows for offline computation of the control gains.


ThB3 
Aphrodite + Hermes 
Embedded Optimization Algorithms for Predictive Control and Estimation II 
Invited Session 
Chair: Patrinos, Panagiotis  KU Leuven 
CoChair: Zanelli, Andrea  Univ. of Freiburg 
Organizer: Quirynen, Rien  Mitsubishi Electric Res. Lab. (MERL) 
Organizer: Patrinos, Panagiotis  KU Leuven 
Organizer: Diehl, Moritz  AlbertLudwigsUniv. Freiburg 

14:3014:50, Paper ThB3.1  
Experimental Validation of Distributed Optimal Vehicle Coordination (I) 
Zanon, Mario  IMT Inst. for Advanced Studies Lucca 
Hult, Robert  Chalmers Univ. of Tech 
Gros, Sébastien  Chalmers Univ. of Tech 
Falcone, Paolo  Chalmers Univ. of Tech 
Keywords: Automotive, Predictive control for nonlinear systems, Optimization algorithms
Abstract: In this paper we solve the problem of coordinating autonomous vehicles approaching an intersection in experiments. We cast the problem in the distributed optimisation framework and use the algorithm proposed in [10], [14] to solve it in real time. We compare two variants of the algorithm in simulations and test our algorithm in experiments using real cars on a test track. The experimental results demonstrate the applicability and realtime feasibility of the algorithm and show that the underlying assumptions are justified.


14:5015:10, Paper ThB3.2  
A RealTime Iteration Scheme with QuasiNewton Jacobian Updates for Nonlinear Model Predictive Control (I) 
Hespanhol, Pedro  Univ. of California Berkeley 
Quirynen, Rien  Mitsubishi Electric Res. Lab. (MERL) 
Keywords: Predictive control for nonlinear systems, Optimization algorithms
Abstract: Nonlinear model predictive control (NMPC) requires the solution of a dynamic optimization problem at each sampling instant under strict timing constraints, involving nonlinear dynamics that can often be stiff or implicitly defined. The realtime iteration (RTI) scheme has been shown to allow realworld embedded applications of NMPC. The present paper proposes an extension of the standard RTI algorithm with a blockstructured quasiNewton method to obtain lowrank Jacobian updates that preserve the block structure of the optimal control problem. In addition, a particular structureexploiting implementation is presented for implicit integration schemes such that no Jacobian evaluation is needed neither any matrix factorization. Based on a proof of concept implementation in C code, the computational performance of the algorithm is illustrated for multiple NMPC case studies.


15:1015:30, Paper ThB3.3  
Embedded Nonlinear Model Predictive Control for Obstacle Avoidance Using PANOC (I) 
Sathya, Ajay Suresha  KU Leuven 
Sopasakis, Pantelis  KU Leuven 
Van Parys, Ruben  KU Leuven 
Themelis, Andreas  IMT Inst. for Advanced Studies Lucca 
Pipeleers, Goele  KU Leuven, LRD 
Patrinos, Panagiotis  KU Leuven 
Keywords: Predictive control for nonlinear systems, Optimization algorithms, Autonomous robots
Abstract: We employ the proximal averaged Newtontype method for optimal control (PANOC) to solve obstacle avoidance problems in real time. We introduce a novel modeling framework for obstacle avoidance which allows us to easily account for generic, possibly nonconvex, obstacles involving polytopes, ellipsoids, semialgebraic sets and generic sets described by a set of nonlinear inequalities. PANOC is particularly wellsuited for embedded applications as it involves simple steps, its implementation comes with a low memory footprint and its fast convergence meets the tight runtime requirements of fast dynamical systems one encounters in modern mechatronics and robotics. The proposed obstacle avoidance scheme is tested on a labscale autonomous vehicle.


15:3015:50, Paper ThB3.4  
RealTime Nonlinear Model Predictive Control of a Motion Simulator Based on a 8DOF Serial Robot (I) 
Katliar, Mikhail  Max Planck Inst. for Biological Cybernetics 
Drop, Frank  Max Planck Inst. for Biological Cybernetics 
Teufel, Harald  Max Planck Inst. for Biological Cybernetics 
Diehl, Moritz  AlbertLudwigsUniv. Freiburg 
Bülthoff, Heinrich H.  Max Planck Inst. for Biological Cybernetics 
Keywords: Predictive control for nonlinear systems, Optimal control, Robotics
Abstract: In this paper we present the implementation of a model predictive controller (MPC) for realtime control of a motion simulator based on a serial robot with 8 degrees of freedom. The goal of the controller is to accurately reproduce six reference signals simultaneously (the accelerations and angular velocities in the body frame of reference) taken from a simulated or real vehicle, by moving the human participant sitting inside the cabin located at the end effector. The controller computes the optimal combined motion of all axes while keeping the axis positions, velocities and accelerations within their limits. The motion of the axes is computed every 12ms based on a prediction horizon consisting of 60 steps, spaced 48ms apart, thus looking ahead 2.88s. To evaluate tracking performance, we measured the acceleration and angular velocity in the cabin using an Inertial Measurement Unit (IMU) for synthetic (doublets and triangledoublets) and realistic (recorded car and helicopter maneuvers) reference signals. We found that fastchanging acceleration inputs excite the natural frequencies of the system, leading to severe mechanical oscillations. These oscillations can be modelled by a secondorder LTI system and mitigated by including this model in the controller. The use of proper algorithms and software allows the computations to be done in realtime.


15:5016:10, Paper ThB3.5  
Embedded MixedInteger Quadratic Optimization Using the OSQP Solver (I) 
Stellato, Bartolomeo  MIT 
Naik, Vihangkumar Vinaykumar  IMT School for Advanced Studies Lucca 
Bemporad, Alberto  IMT Inst. for Advanced Studies Lucca 
Goulart, Paul J.  Univ. of Oxford 
Boyd, Stephen P.  Stanford Univ 
Keywords: Optimization, Computational methods, Hybrid systems
Abstract: We present a novel branchandbound solver for mixedinteger quadratic programs (MIQPs) that efficiently exploits the firstorder OSQP solver for the quadratic program (QP) subproblems. Our algorithm is very robust, requires no dynamic memory allocation and is divisionfree once an initial factorization is computed. Thus, it suitable for embedded applications with low computing power. Moreover, it does not require any assumption on the problem data such as strict convexity of the objective function. We exploit factorization caching and warmstarting to reduce the computational cost of QP relaxations during branchandbound and over repeated solutions of parametric MIQPs such as those arising in embedded control, portfolio optimization, and machine learning. Numerical examples show that our method, using a simple highlevel Python implementation interfaced with the OSQP solver, is competitive with established commercial solvers.


16:1016:30, Paper ThB3.6  
Nonlinear Model Predictive Control of a HumanSized Quadrotor (I) 
Zanelli, Andrea  Univ. of Freiburg 
Horn, Gregory  KU Leuven 
Frison, Gianluca  Tech. Univ. of Denmark 
Diehl, Moritz  AlbertLudwigsUniv. Freiburg 
Keywords: Predictive control for nonlinear systems, Optimal control, Aerospace
Abstract: This paper discusses the design, implementation and deployment of an attitude controller for a quadrotor based on nonlinear model predictive control on a lowpower embedded system equipped with a Cortex A9 CPU running at 800 MHz. Due to the limited computational power of the available hardware, a modified interiorpoint solver for the socalled partially tightened RealTime Iteration is used. The algorithm splits the prediction horizon in two sections. A Riccatilike recursion is exploited that relies on a single linearization of the complementarity conditions per samplingtime for the terminal section. In this way, it is possible to achieve a speedup of a factor 3 with respect to a standard realtime iteration formulation for the application under consideration. Simulation results that show the improvement in performance obtained by using NMPC over standard control techniques are discussed and experimental results using the proposed implementation are presented.


ThB4 
Poseidon 
Delay Systems II 
Regular Session 
Chair: BekiarisLiberis, Nikolaos  Tech. Univ. of Crete 
CoChair: Efimov, Denis  INRIA 

14:3014:50, Paper ThB4.1  
Control of Systems with Arbitrary Bounded Input Delay Using Implicit Lyapunov Function Technique 
Zimenko, Konstantin  ITMO Univ 
Efimov, Denis  Inria 
Polyakov, Andrey  INRIA Lille NordEurope 
Kremlev, Artem  ITMO Univ 
Keywords: Lyapunov methods, Delay systems, Stability of nonlinear systems
Abstract: The paper presents control algorithms for systems with input delay. There are two main results based on using Implicit Lyapunov Function (ILF) technique: 1) an LMIbased approach is presented to evaluate the domain of attraction of a finitetime stable control in the case of the arbitrary bounded delayed control input; 2) a uniting control is designed with commutation between two laws providing a global boundedness of all trajectories of systems with any input delay, and convergence to the origin for a sufficiently small one. The results are also preserved for the timevarying delay case. The theoretical results are supported by numerical examples.


14:5015:10, Paper ThB4.2  
Compensation of Transport Actuator Dynamics with Input Dependent Moving Controlled Boundary 
BekiarisLiberis, Nikolaos  Tech. Univ. of Crete 
Krstic, Miroslav  Univ. of California at San Diego 
Keywords: Distributed parameter systems, Delay systems, Stability of nonlinear systems
Abstract: We introduce and solve the stabilization problem of a transport PDE/nonlinear ODE cascade, in which the PDE state evolves on a domain whose length depends on the boundary values of the PDE state itself. In particular, we develop a predictorfeedback control design, which compensates such transport PDE dynamics. We prove local asymptotic stability of the closedloop system in the C1 norm of the PDE state employing a Lyapunovlike argument and introducing a backstepping transformation. We also highlight the relation of the PDEODE cascade to a nonlinear system with input delay that depends on past input values and present the predictorfeedback control design for this representation as well.


15:1015:30, Paper ThB4.3  
Control of Nonlinear Systems with Actuator Dynamics Governed by Quasilinear FirstOrder Hyperbolic PDEs 
BekiarisLiberis, Nikolaos  Tech. Univ. of Crete 
Krstic, Miroslav  Univ. of California at San Diego 
Keywords: Distributed parameter systems, Delay systems, Stability of nonlinear systems
Abstract: We present a methodology for stabilization of general nonlinear systems with actuator dynamics governed by quasilinear transport PDEs. Since for such PDEODE cascades the speed of propagation depends on the PDE state itself (which implies that the prediction horizon cannot be a priori known analytically), the key design challenge is the determination of the predictor state. We resolve this challenge and introduce a PDE predictorfeedback control law that compensates the transport actuator dynamics. Due to the potential formation of shock waves in the solutions of quasilinear, firstorder hyperbolic PDEs (which is related to the fundamental restriction for systems with timevarying delays that the delay rate is bounded by unity), we limit ourselves to a certain feasibility region around the origin and we show that the PDE predictorfeedback law achieves asymptotic stability of the closedloop system, providing an estimate of its region of attraction. Our analysis combines Lyapunovlike arguments and ISS estimates. Since it may be intriguing as to what is the exact relation of the cascade to a system with input delay, we highlight the fact that the considered PDEODE cascade gives rise to a system with input delay, with a delay that depends on past input values (defined implicitly via a nonlinear equation).


15:3015:50, Paper ThB4.4  
Design of Robust Structurally Constrained Controllers for MIMO Plants with TimeDelays 
Dileep, Deesh  KU Leuven 
Michiels, Wim  KU Leuven 
Hetel, Laurentiu  CNRS 
Richard, JeanPierre  Ec. Centrale De Lille 
Keywords: Delay systems, Decentralized control, H2/Hinfinity methods
Abstract: The structurally constrained controller design problem for linear time invariant neutral and retarded timedelay systems (TDS) is considered in this paper. The closed loop system of the plant and structurally constrained controller is modelled by a system of delay differential algebraic equations (DDAEs). A robust controller design approach using the existing spectrum based stabilisation and the Hinfinity norm optimisation of DDAEs has been proposed. A MATLAB based tool has been made available to realise this approach. This tool allows the designer to select the subcontroller inputoutput interactions and fix their orders. The results obtained while stabilising and optimising two TDS using structurally constrained (decentralised and overlapping) controllers have been presented in this paper.


15:5016:10, Paper ThB4.5  
A GramianBased Observer with Uniform Convergence Rate for Delayed Measurements 
RuedaEscobedo, Juan Gustavo  Inst. De Ingeniería, UNAM 
Ushirobira, Rosane  Inria 
Efimov, Denis  Inria 
Moreno, Jaime  Univ. Nacional Autonoma De MexicoUNAM 
Keywords: Observers for linear systems, Delay systems, Linear systems
Abstract: The problem of designing an observer for a linear timeinvariant system with delayed measurements of the state is revisited in this paper. The delay is assumed to be timevarying with finite unknown lower and upper bounds. A Gramianbased observer is proposed with a fixedtime convergence rate to a ball. The efficiency of the obtained solution is illustrated by a numerical comparison with linear observers (one of which is tuned under the assumption that a nominal value of delay is available).


16:1016:30, Paper ThB4.6  
Robust Linear Quadratic Regulator for Uncertain Linear DiscreteTime Systems with Delay in the States: An Augmented System Approach 
Bortolin, Daiane Cristina  Univ. De São Paulo 
Odorico, Elizandra Karla  Univ. De São Paulo 
Terra, Marco Henrique  Univ. of Sao Paulo at Sao Carlos 
Keywords: Delay systems, Uncertain systems, Robust control
Abstract: In this paper we deal with the regulation problem for a class of uncertain discretetime systems with known constant delays in the states. Uncertainties are assumed normbounded and affect all parametric matrices of the system. Applying the lifting method, the delayed system is transformed into an augmented delayfree system. Then, the control law is obtained from combination of penalty functions and robust regularized leastsquares problem, when there exist uncertainties in the data. The solution provided is given in terms of augmented Riccati equations presented in a framework given by an array of matrices.


ThB5 
Athenaeum 1 
Observers II 
Regular Session 
Chair: Possieri, Corrado  Pol. Di Torino 
CoChair: Gehan, Olivier  ENSICAEN 

14:3014:50, Paper ThB5.1  
SetBased State Estimation of Nonlinear Systems Using Constrained Zonotopes and Interval Arithmetic 
Rego, Brenner Santana  Federal Univ. of Minas Gerais 
Raimondo, Davide Martino  Univ. of Pavia 
Raffo, Guilherme Vianna  Federal Univ. of Minas Gerais 
Keywords: Observers for nonlinear systems, Nonlinear system theory, Filtering
Abstract: This paper proposes a novel setvalued state estimation algorithm for nonlinear discretetime systems with unknownbutbounded disturbances. The problem is often addressed through conservative linearization, leading to severe overestimation. By combining important properties from interval arithmetic and the recently proposed constrained zonotopes, a highly tunable and accurate state estimation algorithm is developed, capable of providing tight bounds for the setvalued state estimation problem. A numerical experiment is presented to demonstrate the performance of the proposed strategy.


14:5015:10, Paper ThB5.2  
A Chain Observer for a Class of Nonlinear Systems with Long Multiple Delays in Output Measurements 
targui, boubekeur  Ensicaen Caen, France 
Hernandez Gonzalez, Omar  Univ. De Caen Basse Normandie 
AstorgaZaragoza, Carlos  TecnolÓgico Nacional De MÉxico  Cenidet 
Pouliquen, Mathieu  Univ. of Caen 
Gehan, Olivier  ENSICAEN 
Keywords: Observers for nonlinear systems, Nonlinear system theory
Abstract: The main contribution of this paper is to present a chain observer for a class of nonlinear systems when the output measurements are affected by a long known and multiple time delay. The observer is firstly presented when the output measurements are affected by a long multiple and constant time delay, then the observer is extended to the case of a long known and multiple bounded timevarying delay. The proposed observer is composed of a chain of observers with similar algorithm. Each observer estimates the state over a time horizon, while the first observer of the chain estimates the current state. The structure of each observer of the chain is based on the presence of a dynamical term which permits the compensation of the delay. The observer gains are computed by a set of parameterized Linear Matrix Inequalities (LMIs) which depends on the delay. A LyapunovKrasovskii functional is used to demonstrate the asymptotical convergence to zero of the observation error. The performance of the proposed observer is evaluated through a numerical example. The observer exhibits good estimation of the states of the system, even in the presence of significant delayed measurements.


15:1015:30, Paper ThB5.3  
An Interval Observer Approach for the Online Temperature Estimation in Solid Oxide Fuel Cell Stacks 
Rauh, Andreas  Univ. of Rostock 
Kersten, Julia  Univ. of Rostock 
Aschemann, Harald  Univ. of Rostock 
Keywords: Observers for nonlinear systems, Randomized algorithms, Modeling
Abstract: Interval observers that are based on the structural property of cooperativity allow for the computation of guaranteed lower and upper bounds for all state trajectories of dynamic systems by defining two sets of bounding systems for the dynamic behavior. In such a way, they remove the disadvantage of predictorcorrector interval estimation schemes which often suffer from the fact that the dynamic system model has to be evaluated over finitely large domains of state variables and parameters (often interval boxes or zonotopes). This evaluation typically results in overestimation that includes unphysical parts of the statespace in the computed results. To counteract this socalled wrapping effect, computationally expensive, problemspecific algorithms need to be implemented. However, this drawback can be removed for systems that have certain monotonicity and stability properties. In this paper, it is shown that appropriately defined system models for the thermal behavior of hightemperature fuel cells belong to this class of systems. The corresponding interval observer design is presented methodologically and demonstrated with the help of measured data from a test rig available at the Chair of Mechatronics at the University of Rostock.


15:3015:50, Paper ThB5.4  
Homography Observer Design on Special Lineal Group SL(3) with Application to Optical Flow Estimation 
Manerikar, Ninad  Univ. of Nice, Sophia Antipolis 
HUA, MinhDuc  I3s UcaCnrs Umr7271 
Hamel, Tarek  Univ. De Nice Sophia Antipolis 
Keywords: Observers for nonlinear systems, Sensor and signal fusion, Robotics
Abstract: In this paper we present a novel approach for homography observer design on the Special Linear group SL(3). This proposed approach aims at computing online, the optical flow estimation extracted from continuous homography obtained by capturing video sequences from aerial vehicles equipped with a monocular camera. The novelty of the paper lies in the linearization approach undertaken in order to linearize the nonlinear observer on SL(3). Experimental results have been presented to show the performance and robustness of the proposed approach.


15:5016:10, Paper ThB5.5  
A New Robust Observer Approach for Unknown Input and State Estimation 
POPESCU, Andrei  Univ. Grenoble Alps 
Besancon, Gildas  Ense3  Grenoble INP 
Voda, Alina  Joseph Fourier Univ. of Grenoble 
Keywords: Observers for linear systems
Abstract: This paper proposes a new approach for a robust estimation of unknown inputs as well as state variables in a dynamical system subject to noise or uncertainties. Following a recently proposed controlbased methodology for observer design, the idea here is to take advantage of robust control methods, and in particular of Hinf techniques. This provides a new method which is applied for an example of tunneling current estimation in an STMlike device. Simulation results are finally provided together with a comparison to the formerly available Hinf filtering, but also to the robust slidingmode technique.


16:1016:30, Paper ThB5.6  
Position and Speed Observer for PMSM with Unknown Stator Resistance 
Bazylev, Dmitry  ITMO Univ 
Pyrkin, Anton  ITMO Univ 
Bobtsov, Alexey  ITMO Univ 
Keywords: Observers for nonlinear systems, Stability of nonlinear systems, Nonlinear system identification
Abstract: In this paper a new nonlinear parameterization of the PMSM model is proposed for the case of an uncertain stator resistance. The assumption of known inductance only is applied. After parameterization the regression model of six parameters is obtained from which it becomes possible to reconstruct the resistance and two necessary parameters involved in the position and speed observers design. The dynamic regressor extension and mixing (DREM) estimator is used to provide good performance and fast estimation of a large regression model which is preferable than the standard gradient approach. Simulation results illustrating proposed approach are given.


ThB6 
Athenaeum 2 
Stabilization 
Regular Session 
Chair: Lazar, Mircea  Eindhoven Univ. of Tech 
CoChair: Schiffer, Johannes  Univ. of Leeds 

14:3014:50, Paper ThB6.1  
Feedback Stabilization of Positive Nonlinear Systems with Applications to Biological Systems 
Steentjes, Tom Robert Vince  Eindhoven Univ. of Tech 
Doban, Alina Ionela  Tech. Univ. of Eindhoven 
Lazar, Mircea  Eindhoven Univ. of Tech 
Keywords: Lyapunov methods, Constrained control, Biological systems
Abstract: Various feedback stabilizers based on Sontag's ``universal'' formula for stabilizing control laws are presented, incorporating restrictions inspired by models from systems biology. The main contribution is an extension of Sontag's ``universal'' formula for positive nonlinear control systems. More specifically, an auxiliary function is introduced in the feedback interconnection, such that invariance of the positive orthant is retained for the system in closed loop with the ``universal'' stabilizer. We further state a ``universal'' eventbased stabilizer with bounded controls for positive systems. Two examples inspired by systems biology are presented for illustration and demonstration of the effectiveness of the proposed stabilizers. One example considers the problem of regulating the cortisol level, where the methodology is shown to provide clinically realistic control inputs, suitable for treatment in real life.


14:5015:10, Paper ThB6.2  
A Piecewise Affine Control Lyapunov Function for Robust Control 
NGUYEN, Ngoc Anh  Univ. Paris Saclay 
Olaru, Sorin  CentraleSupélec 
Keywords: Lyapunov methods, Constrained control, Optimization
Abstract: This paper presents the construction of a convex piecewise affine control Lyapunov function for constrained linear discretetime systems, affected by bounded additive disturbances. Exploiting the properties of this control Lyapunov function, the closedloop dynamics are shown to converge to a given fulldimensional robust positively invariant set. Moreover, the proposed method leads to a simple robust control algorithm which only requires solving a linear programming problem at each sampling instant. Finally, the controller design is illustrated via a numerical example.


15:1015:30, Paper ThB6.3  
FiniteTime and FixedTime Stabilization for Integrator Chain of Arbitrary Order 
Zimenko, Konstantin  ITMO Univ 
Polyakov, Andrey  INRIA Lille NordEurope 
Efimov, Denis  Inria 
Perruquetti, Wilfrid  Ec. Centrale De Lille 
Keywords: Lyapunov methods, Stability of nonlinear systems
Abstract: In the present paper, homogeneous control laws are designed for finitetime and fixedtime stabilization of integrator chains of arbitrary order. Provided analysis is based on Lyapunov function method and homogeneity concept. Fixedtime convergence is achieved by use of hybrid control algorithm with homogeneity degree changing. Performance of the resulting finitetime and fixedtime feedbacks is illustrated by numerical simulations.


15:3015:50, Paper ThB6.4  
Constructive Lyapunov Stabilization with Approximate Optimality for a Class of Nonlinear Systems 
Xu, Zhenhui  Sophia Univ 
Shen, Tielong  Sophia Univ. of Tokyo 
Keywords: Lyapunov methods, Nonlinear system theory, Optimal control
Abstract: This paper presents a recursive constructing approach to Lyapunov function with optimality for a class of nonlinear systems. The targeted systems are formulated as cascaded system with triangular structure. For this class of nonlinear systems, stabilization problem has been a typical issue and solved by recursively constructing Lyapunov function using socalled backstepping process. However, as is well known, this constructive design of feedback tabilizing control law is usually lacking time response performance due to the attention of controller design focuses stability only. The presented design approach in this paper puts an optimality into the recursive design process by targeting an approximate solution of Hamiltonian equality. It has been shown that at each stage of the recursive design a Lyapunov function that guarantees optimality can be obtained approximately by policy iteration. Finally, numerical examples are shown to demonstrate the design process.


15:5016:10, Paper ThB6.5  
A New Criterion for Boundedness of Solutions for a Class of Periodic Systems 
Efimov, Denis  Inria 
Schiffer, Johannes  Univ. of Leeds 
Keywords: Lyapunov methods, Nonlinear system theory
Abstract: A wide range of practical systems exhibits dynamics, which are periodic with respect to several state variables and which possess multiple invariant solutions. Yet, when analyzing stability of such systems, many classical techniques often fall short in that they only permit to establish local stability properties. Motivated by this, we present a new sufficient criterion for global stability of such a class of nonlinear systems. The proposed approach is characterized by two main properties. First, it develops the conventional cell structure framework to the case of multiple periodic states. Second, it extends the standard Lyapunov theory by relaxing the usual definiteness requirements of the employed Lyapunov functions to signindefinite functions.


16:1016:30, Paper ThB6.6  
On the Almost Global Stability of Invariant Sets 
Karabacak, Özkan  Aalborg Univ 
Wisniewski, Rafael  Section for Automation and Control, Aalborg Univ 
Leth, John  Aalborg Univ 
Keywords: Lyapunov methods, Stability of nonlinear systems, Nonlinear system theory
Abstract: For a given invariant set of a dynamical system, it is known that the existence of a Lyapunovtype density function, called Lyapunov density or Rantzer’s density function, may imply the convergence of almost all solutions to the invariant set, in other words, the almost global stability (also called almost everywhere stability) of the invariant set. For discretetime systems, related results in literature assume that the state space is compact and the invariant set has a local basin of attraction. We show that these assumptions are redundant. Using the duality between FrobeniusPerron and Koopman operators, we provide a Lyapunov density theorem for discretetime systems without assuming the compactness of the state space or any local attraction property of the invariant set. As a corollary to this new discretetime Lyapunov density theorem, we provide a continuoustime Lyapunov density theorem which can be used as an alternative to Rantzer’s original theorem, especially where the solutions are known to exist globally.


ThB7 
Athenaeum 3 
Transportation Systems 
Regular Session 
Chair: Geroliminis, Nikolas  Ec. Pol. Fédérale De Lausanne (EPFL), Urban Transport Systems Lab 
CoChair: Bechlioulis, Charalampos  National Tech. Univ. of Athens 

14:3014:50, Paper ThB7.1  
A Structured Linear Quadratic Controller for Transportation Problems 
Heyden, Martin  Lund Univ 
Pates, Richard  Lund Univ 
Rantzer, Anders  Lund Univ 
Keywords: Network analysis and control, Distributed control, Transportation systems
Abstract: We study a linear quadratic control problem for transportation optimization on a directed line graph. We show that the solution to the Riccati equation associated with this problem is highly structured. The feedback law is almost upper triangular, and the synthesis of the feedback law is given by a recursion, making it scalable. The structure of the feedback law also allows for an efficient realization of the controller using a local communication scheme.


14:5015:10, Paper ThB7.2  
Hybrid Model Predictive Control of Bus Transport Systems 
Sirmatel, Isik Ilber  EPFL 
Geroliminis, Nikolas  Ec. Pol. Fédérale De Lausanne (EPFL), Urban Transport 
Keywords: Transportation systems, Hybrid systems, Predictive control for nonlinear systems
Abstract: Schedule instability adversely affects the performance of bus transport systems. Bus bunching and similar irregularities decrease quality of bus services and increase travel times. In this regard, bus system management schemes are of high importance. Motivated by the potential impact of developing advanced bus control schemes on transportation practice, in this paper a hybrid model predictive control scheme with actuation via bus speeds is developed, which can regularize headways and improve bus service quality. The controller extends upon earlier work by considering detailed dynamics of the interactions between buses and stops via passenger flows. Performance of the controller is compared with a no control case and a PI controller via simulation experiments, conducted with a recently proposed mixed logical dynamical bus loop model involving both continuous (e.g., bus positions) and binary (e.g., the state of a bus regarding whether it is holding at a certain stop or not) states. Results showcase the capability of the proposed controller in regularizing headways, avoiding bus bunching and decreasing passenger travel times


15:1015:30, Paper ThB7.3  
Personalised Optimal Speed Advice to Cyclists Approaching an Intersection with Uncertain Green Time 
Dabiri, Azita  Delft Univ. of Tech 
Hegyi, Andreas  Delft Univ. of Tech 
Keywords: Transportation systems, Markov processes, Stochastic control
Abstract: When travelling in urban areas with signalised intersections, cyclists are currently unable to optimize their speed profile according to their preferences. By integrating information from traffic signal phase and timing and cyclists’ preferences, in this paper an algorithm is developed that generates personalised optimal speed advice for cyclists. Results from the simulation experiment indicate that in spite of stochastic nature of traffic light phase and timing, with the developed algorithm based on stochastic dynamic programming, an speed profile can be advised that meets the cyclist’s preferences in terms of travel time, energy consumption, or the number of stops.


15:3015:50, Paper ThB7.4  
Optimal Motion Planning for Automated Vehicles with Scheduled Arrivals at Intersections 
Müller, Eduardo  Federal Univ. of Santa Catarina 
Wahlberg, Bo  KTH Royal Inst. of Tech 
Carlson, Rodrigo Castelan  Federal Univ. of Santa Catarina 
Keywords: Transportation systems, Optimal control, Traffic control
Abstract: We design and compare three different optimal control strategies for the motion planning of automated vehicles approaching an intersection with scheduled arrivals. The objective is to minimize a combination of energy consumption and deviation from the schedule. The strategies differ in allowed deviations. When taking only vehicles inside the control region into account, the strategy that achieves the lowest energy consumption is the less strict one, albeit at the expense of higher travel times. When traffic conditions beyond the control region are considered, no strategy is able to achieve lower energy consumption or vehicle delay than the strategy that is the most strict in keeping with the schedule. Results suggests that in high traffic situations, from a global energy consumption standpoint, it is best to have vehicles crossing the intersection as soon as possible.


15:5016:10, Paper ThB7.5  
Density and Flow Reconstruction in Urban Traffic Networks Using Heterogeneous Data Sources 
Ladino, Andres  Univ. Grenoble Rhône Alpes, GIPSALab, CNRS 
CanudasdeWit, Carlos  CNRSGIPSALabGrenoble 
Kibangou, Alain Y.  Grenoble INP/Univ. Joseph Fourier/CNRS 
fourati, Hassen  Univ. Joseph Fourrier, GIPSALAB 
Rodriguez Vega, Martin  Univ. Grenoble Alpes, CNRS, Grenoble INP, GIPSALab 
Keywords: Transportation systems, Sensor and mesh networks, Optimization
Abstract: In this paper, we consider the problem of joint reconstruction of flow and density in an urban traffic network using heterogeneous sources of information. The traffic network is modelled within the framework of macroscopic traffic models, where we adopt acf{LWR} conservation equation characterized by a piecewise linear fundamental diagram. The estimation problem considers two key principles. First, the error minimization between the measured and reconstructed flows and densities, and second the equilibrium state of the network which establishes flow propagation within the network. Both principles are integrated together with the traffic model constraints established by the supply/demand paradigm. Finally, the problem is cast as a constrained quadratic optimization with equality constraints in order to shrink the feasible region of estimated variables. Some simulation scenarios based on synthetic data for a Manhattan grid network are provided in order to validate the performance of the proposed algorithm.


16:1016:30, Paper ThB7.6  
NeuroAdaptive Traffic Congestion Control for Urban Road Networks 
Bechlioulis, Charalampos  National Tech. Univ. of Athens 
Kyriakopoulos, Kostas J.  National Tech. Univ. of Athens 
Keywords: Traffic control, Adaptive control, Neural networks
Abstract: The rapid increase of private vehicles combined with the limited capabilities of the urban road infrastructure has made congestion one of the main problems of major cities worldwide, having a severe impact on both the economy and the environment. In this work, we shall attempt to solve the traffic management problem by examining in a unified manner the traffic network, the route guidance of the vehicles and the regulation of the traffic lights, as the basic elements of a single controlled system. In particular, we propose a decentralized adaptive control system, comprised of three main modules: i) the network congestion estimator, ii) the reference travel time estimator, and iii) the rate controller, that is capable of efficiently regulating the travel time along the traffic network while avoiding congestion at the junctions. The design of decentralized control algorithms and their implementation as traffic management applications for portable computing devices (e.g., 3^{mathrm{rd}} and 4^{mathrm{th}} generation mobile phones, tablets, computers embedded in "smart" vehicles) is expected to improve drastically the traffic condition of urban road networks. Meanwhile, in future traffic networks, where the navigation of the vehicles will be conducted by autopilots in the absence of humandrivers, the use of such a distributed autonomous management system will be essential.


ThB8 
Athenaeum 4 
Learning and Adaptivity in Predictive and Optimal Control: Theory,
Applications, and Future Perspectives II 
Invited Session 
Chair: Farina, Marcello  Pol. Di Milano 
CoChair: Fagiano, Lorenzo  Pol. Di Milano 
Organizer: Farina, Marcello  Pol. Di Milano 
Organizer: Fagiano, Lorenzo  Pol. Di Milano 

14:3014:50, Paper ThB8.1  
Scalable Synthesis of Safety Certificates from Data with Application to LearningBased Control (I) 
Wabersich, Kim Peter  ETH Zurich 
Zeilinger, Melanie N.  ETH Zurich 
Keywords: Robust control, Machine learning, Safety critical systems
Abstract: The control of complex systems faces a tradeoff between high performance and safety guarantees, which in particular restricts the application of learningbased methods to safetycritical systems. A recently proposed framework to address this issue is the use of a safety controller, which guarantees to keep the system within a safe region of the state space. This paper introduces efficient techniques for the synthesis of a safe set and control law, which offer improved scalability properties by relying on approximations based on convex optimization problems. The first proposed method requires only an approximate linear system model and Lipschitz continuity of the unknown nonlinear dynamics. The second method extends the results by showing how a Gaussian process prior on the unknown system dynamics can be used in order to reduce conservatism of the resulting safe set. We demonstrate the results with numerical examples, including an autonomous convoy of vehicles.


14:5015:10, Paper ThB8.2  
Adaptive Model Predictive Control for Constrained Time Varying Systems (I) 
Tanaskovic, Marko  Singidunum Univ 
Fagiano, Lorenzo  Pol. Di Milano 
Gligorovski, Vojislav  Univ. of Belgrade 
Keywords: Predictive control for linear systems, Adaptive systems, Identification for control
Abstract: An approach to design feedback controllers for discretetime, uncertain, linear timevarying systems subject to constraints is proposed. Building on previous contributions in the framework of timeinvariant systems, in each sampling period a twostep procedure is carried out. In the first step, a set of linear models that are consistent with past inputoutput data and prior assumptions is built and refined. This set is guaranteed to contain also the true system dynamics if the considered working assumptions are valid. The timevarying nature of the plant is captured by assuming known bounds on the rate of change of the model parameters in time. In the second step, a robust finitehorizon optimal control problem is formulated and solved. The resulting optimal control sequence guarantees that the outputs of all possible plants inside the model set satisfy the operational constraints. The approach is showcased in numerical simulations on a threetank system.


15:1015:30, Paper ThB8.3  
DataDriven Control Design in the Loewner Framework: Dealing with Stability and Noise (I) 
Kergus, Pauline  ONERA 
Formentin, Simone  Pol. Di Milano 
PoussotVassal, Charles  Onera 
Demourant, Fabrice  ONERACERT 
Keywords: Linear systems, Model/Controller reduction
Abstract: The LDDC (Loewner Data Driven Control) algorithm is a datadriven controller design method based on frequencydomain inputoutput data. The identification of the plant is skipped and the controller is designed directly from the measurements using the Loewner approach, known for model approximation and reduction. However, in the LDDC method, the identified controller is not guaranteed to be stable and the effect of noise on the identified controller is unknown. In this article, we ensure the stability of the controller and propose a solution to deal with noisy data. The method is validated on a numerical example.


15:3015:50, Paper ThB8.4  
Robust Predictive Control with DataBased MultiStep Prediction Models (I) 
Terzi, Enrico  Pol. Di Milano 
Farina, Marcello  Pol. Di Milano 
Fagiano, Lorenzo  Pol. Di Milano 
Scattolini, Riccardo  Pol. Di Milano 
Keywords: Robust control, Optimal control
Abstract: In this paper a novel method for the design of MPC controllers based on multisteps ahead external representation system models is proposed. These models are assumed to be identified from available data, e.g., using the setmembership approach, that computes a guaranteed estimate of the model uncertainty bounds. Exploiting this information, the proposed algorithm guarantees input and output constraint satisfaction, recursive feasibility, and robust convergence properties. A simulation case study is shown to demonstrate the effectiveness of the proposed approach.


15:5016:10, Paper ThB8.5  
Historian Data Based Predictive Control of a Water Distribution Network (I) 
Salvador, Jose R.  Univ. De Sevilla 
Muñoz de la Peña, David  Univ. of Sevilla 
Ramirez, Daniel R.  Univ. of Sevilla 
Alamo, Teodoro  Univ. De Sevilla 
Keywords: Statistical learning, Predictive control for nonlinear systems, Largescale systems
Abstract: In this paper, a novel historian data based predictive control strategy is presented and used to control a wáter distribution network simulated using the EPANET software. The control actions are computed based on past historian data. The historian stores closed loop operation data of the process with different controllers used in the past. The predictive controller computes the current control actions as a weighted sum of past control actions so that a performance cost over a prediction horizon is minimized. This predictive strategy does not need an explicit model of the process and it is well suited to control applications of large and complex processes such as water distribution networks. To limit the computational burden, only a subset of the past control actions in the historian are considered in current control computations. This subset is comprised of closed loop trajectories starting from a initial state close to the current state of the process. Furthermore, other parameters different from the initial state can be considered when choosing the historian subset (e.g., the set point values).


16:1016:30, Paper ThB8.6  
DataDriven Polynomial MPC and Application to Blood Glucose Regulation in a Diabetic Patient (I) 
Novara, Carlo  Pol. Di Torino 
Rabbone, Ivana  Ospedale Sant’Anna Torino 
Tinti, Davide  Ospedale Sant’Anna Torino 
Keywords: Identification for control, Predictive control for nonlinear systems, Biomedical systems
Abstract: The majority of control design approaches assume that an accurate firstprinciple model of the system to control is available. However, in many realworld applications, deriving an accurate model is extremely difficult, since the system dynamics may be not well known and/or too complex. In this paper, a polynomial model predictive control (PMPC) approach for nonlinear systems is presented, relying on the identification from data of a polynomial prediction model. The main advantages of this approach over the standard methods are that it does not require a detailed knowledge of the plant to control and it is computationally efficient. A realdata application is presented, concerned with regulation of blood glucose concentration in a type 1 diabetic patient. This application shows that the PMPC approach can be effective in the biomedical field, where accurate firstprinciple model can seldom be found.


ThB9 
Hera 
Optimal Control I 
Regular Session 
Chair: Dimarogonas, Dimos V.  KTH Royal Inst. of Tech 
CoChair: Menner, Marcel  ETH Zurich 

14:3014:50, Paper ThB9.1  
Optimal Control of LeftInvariant MultiAgent Systems with Asymmetric Formation Constraints 
Colombo, Leonardo, J.  KTH Royal Inst. of Tech 
Dimarogonas, Dimos V.  KTH Royal Inst. of Tech 
Keywords: Variational methods, Optimal control, Agents and autonomous systems
Abstract: In this work we study an optimal control problem for a multiagent system modeled by an undirected formation graph with nodes describing the kinematics of each agent, given by a left invariant control system on a Lie group. The agents should avoid collision between them in the workspace. Such a task is done by introducing some potential functions into the cost functional for the optimal control problem, corresponding to fictitious forces, induced by the formation constraint among agents, that break the symmetry of the individual agents and the cost functions, and rendering the optimal control problem partially invariant by a Lie group of symmetries. Reduced necessary conditions for the existence of normal extremals are obtained using techniques of variational calculus on manifolds. As an application we study an optimal control problem for multiple unicycles.


14:5015:10, Paper ThB9.2  
Optimal Control of a RigidWing Rotary Kite System for Airborne Wind Energy 
De Schutter, Jochem  Univ. of Freiburg 
Leuthold, Rachel  Univ. of Freiburg 
Diehl, Moritz  AlbertLudwigsUniv. Freiburg 
Keywords: Optimal control, Energy systems, Modeling
Abstract: Multiplekite airborne wind energy (AWE) systems are typically characterized by unstable and highly nonlinear dynamics which often translates to intricate controller design and challenging coordination problems. Rotary kite AWE systems (RAWES) have been alternatively proposed for smallscale applications, under the assumption that they can reduce the complexity of the control problem. This paper confirms that a small, rigidwing RAWES in pumping mode can be controlled effectively in a large operational range, using only pitch control as onboard actuation. Optimal control is applied to compute RAWES pumping trajectories in different operating regions, for a design geometry that is optimized for a rated wind speed under structural constraints. The reduced control complexity comes at the cost of a low harvesting factor, close to that of conventional wind turbines.


15:1015:30, Paper ThB9.3  
Merging Vehicles at Junctions Using MixedInteger Model Predictive Control 
Bali, Csaba  Univ. of Bristol 
Richards, Arthur  Univ. of Bristol 
Keywords: Optimal control, Autonomous systems, Transportation systems
Abstract: A method is proposed for vehicle merging scenarios in junctions with relative cost prioritization. The method is based on Model Predictive Control, employing Mixed Integer Quadratic Program optimization. The scheme provides optimal control properties while maintaining safety and recursive feasibility. The latter properties are ensured through positive control invariance of simple time headway constraints. For examples with two vehicles, tunable prioritization and gap acceptance are verified and presented on a decision graph. Priorities are then demonstrated to be respected in an example with four vehicles.


15:3015:50, Paper ThB9.4  
Distributed Scenario Model Predictive Control for Driver Aided Intersection Crossing 
Katriniok, Alexander  Ford Res. & Innovation Center 
Kojchev, Stefan  Eindhoven Univ. of Tech 
Lefeber, Erjen  Eindhoven Univ. of Tech 
Nijmeijer, Hendrik  Eindhoven Univ. of Tech 
Keywords: Optimal control, Control over communication, Transportation systems
Abstract: The automation of road intersections has significant potential to improve traffic throughput and efficiency. While the related control problem is usually addressed assuming fully automated vehicles, we focus on the problem of issuing appropriate speed advices to the driver in order to optimize traffic flow in intersections without any traffic lights or signs. Therefore, a distributed scenariobased model predictive control regime is proposed which accounts for uncertainties in the driver reaction to speed advices issued by the control system. In the scenario approach, we draw independently and identically distributed samples from a bounded uncertainty set and optimize over scenarios which reflect a potential driver reaction. Based on the number of samples, we can give guarantees on avoiding collisions under acting uncertainties. Simulation results demonstrate that the scenario approach is capable of avoiding collisions when the driver reacts uncertain while the nominal approach is not.


15:5016:10, Paper ThB9.5  
Film Growth Minimization in a LiIon Cell: A Pseudo Two Dimensional ModelBased Optimal Charging Approach 
Pozzi, Andrea  Univ. of Pavia 
Torchio, Marcello  Univ. Degli Studi Di Pavia 
Raimondo, Davide Martino  Univ. of Pavia 
Keywords: Optimal control, Energy systems, Emerging control applications
Abstract: Safety, fast charging and aging are among the most important issues when dealing with high power battery packs in electric and hybrid vehicles. Today’s charging strategies are designed to guarantee safe operation in a conservative way, but are far from being optimal in terms of aging reduction and time charging minimization. For this reason, the interest of the research is focused on developing modelbased battery management systems. Comparing standard charging strategies with healthaware optimizationbased ones is difficult since they usually provide different charging times: are we willing to charge the battery in more time if this comes with an aging reduction? When a customer is faced with such a question, the answer is not so trivial. For this reason, in this paper we provide healthaware Pseudo Two Dimensional modelbased strategies with the same charging time of standard CCCV protocols. The results show that significant aging improvements can be obtained, even by constraining the charging time to be the same as the CCCV protocol.


16:1016:30, Paper ThB9.6  
A User Comfort Model and Index Policy for Personalizing Discrete Controller Decisions 
Menner, Marcel  ETH Zurich 
Zeilinger, Melanie N.  ETH Zurich 
Keywords: Optimal control, Intelligent systems, Machine learning
Abstract: User feedback allows for tailoring system operation to ensure individual user satisfaction. A major challenge in personalized decisionmaking is the systematic construction of a user model during operation while maintaining control performance. This paper presents both an indexbased control policy to smartly collect and process user feedback and a user comfort model in the form of a Markov decision process with a priori unknown userspecific state transition probabilities. The control policy utilizes explicit user feedback to optimize a reward measure reflecting user comfort and addresses the explorationexploitation tradeoff in a multiarmed bandit framework. The proposed approach combines restless bandits and upper confidence bound algorithms. It introduces an exploration term into the restless bandit formulation, utilizes user feedback to identify the user model, and is shown to be indexable. We demonstrate its capabilities with a simulation for learning a user’s tradeoff between comfort and energy usage.


ThB10 
The Grill Room 
Cooperative Systems II 
Regular Session 
Chair: Bajcinca, Naim  Univ. of Kaiserslautern 
CoChair: Jones, Colin N  EPFL, Lausanne 

14:3014:50, Paper ThB10.1  
Exploiting the Superposition Property of Wireless Communication for Average Consensus Problems in MultiAgent Systems 
Molinari, Fabio  TU Berlin 
Stanczak, Slawomir  Tech. Univ. Berlin 
Raisch, Joerg  Tech. Univ. Berlin 
Keywords: Distributed cooperative control over networks, Control over communication, Concensus control and estimation
Abstract: This paper studies system stability and performance of multiagent systems in the context of consensus problems over wireless multipleaccess channels (MAC). We propose a consensus algorithm that exploits the broadcast property of the wireless channel. Therefore, the algorithm is expected to exhibit fast convergence and high efficiency in the usage of scarce wireless resources. Although the designed algorithm shows robustness against variations in the channel and consensus is always reached, the agreement value will be depending on these variations.


14:5015:10, Paper ThB10.2  
A CoordinatorDriven Communication Reduction Scheme for Distributed Optimization Using the Projected Gradient Method 
Stathopoulos, Georgios  EPFL 
Jones, Colin N  EPFL, Lausanne 
Keywords: Distributed cooperative control over networks, Optimization algorithms, Agents networks
Abstract: We propose a way to estimate the value function of a convex proximal minimization problem. The scheme constructs a convex set within which the optimizer resides and iteratively refines the set every time that the value function is sampled, namely every time that the proximal minimization problem is solved exactly. The motivation stems from multiagent distributed optimization problems, where each agent is described by a proximal minimization problem unknown to the global coordinator. We prove convergence results related to the solution of such distributed optimization problems in the special case where the projected gradient method is used and demonstrate that the developed scheme significantly reduces communication requirements when applied to a microgrid setting.


15:1015:30, Paper ThB10.3  
EventTriggered Adaptive OFDMA Protocol for MultiVehicle Clusters 
Guma, Shaban  TU Berlin 
Bajcinca, Naim  Univ. of Kaiserslautern 
Keywords: Optimal control of communication networks, Control over networks, Distributed cooperative control over networks
Abstract: We present an eventtriggered based OFDMA resource allocation algorithm in conjunction with the solution to a distributed optimization problem. The optimal OFDMA subcarriers assignment is supported by the sensitivity function of the objective function to exchange information between the nodes of the underlying network. The resulting sensitivitybased eventtriggered policy invokes a crosslayer communication protocol by introducing sensitivitydependent weights in the resource sharing optimization problem at the MAC layer. For practical evaluation purposes, we consider steering of wireless multivehicles, where each vehicle on its own is controlled by a cluster of collaborative onboard distributed processors communicating over a single wireless link.


15:3015:50, Paper ThB10.4  
A GameTheoretic Framework for Distributed Voltage Regulation Over HVDC Grids 
Rodríguez del Nozal, Álvaro  Univ. Loyola Andalucía 
Orihuela, Luis  Univ. Loyola Andalucía 
Millán, Pablo  Univ. Loyola Andalucía 
Keywords: Distributed cooperative control over networks, Optimal control of communication networks, Communication networks
Abstract: In this paper we propose a noncooperative gametheoretical framework to design the strategic behaviour of buses involved in a HVDC grid that are connected to renewable resources. This framework takes under consideration voltage stability issues as well as economic factors when the current carried by the power lines is constrained. In order to reach an agreement in the decisions of the buses, an iterative distributed Nash equilibrium seeking algorithm is implemented. Existence and stability conditions for the equilibrium are introduced for the unconstrained case and simulations show the robustness of the algorithm for the constrained scenario.


15:5016:10, Paper ThB10.5  
Cooperative Defense Strategy for Active Aircraft Protection Considering Launch Time of Defense Missile 
Song, Huiseong  Seoul National Univ 
Lee, Seokwon  Seoul National Univ 
Cho, Namhoon  Seoul National Univ 
Kim, Youdan  Seoul National Univ 
Keywords: Cooperative control, UAV's, Aerospace
Abstract: An aircraft can equip itself with a defense missile for active protection against an attack missile. Although a lot of strategies for cooperation between the aircraft and the defense missile have been studied, little attention has been paid to the early phase of the engagement. In this study, a cooperative defense strategy considering both launch and boost phase of the defense missile is proposed based on geometric approach. The strategy is derived to provide kinematic advantages to the defense team before it enters the endgame phase. Launch time of the defense missile is computed from the collision geometry. Numerical simulation demonstrates that the proposed defense strategy is useful for the defense team having no advantages on velocity and maneuverability over the adversary.


16:1016:30, Paper ThB10.6  
Distributed CommunicationFree Control of Multiple Robots for Circumnavigation of a Speedy Unpredictably Maneuvering Target 
Matveev, Alexey S.  St.Petersburg Univ 
Ovchinnikov, Kirill  Saint Petersburg State Univ 
Keywords: Cooperative autonomous systems, Constrained control, Sliding mode control
Abstract: Multiple fully actuated mobile robots travel in a plane with upperlimited speeds and are driven by accelerations limited in magnitude. The robots should arrive at a prespecified distance from an unpredictable speedy target, then maintain this distance and achieve an even selfdistribution over the respective moving circle, along with a given angular velocity of rotation about the target. The robots do not communicate with anybody and are anonymous to one another; preassignment of different roles to various robots is impossible. Every robot measures only the relative position of the target and companion robots (within a finite range of ``visibility'' in the latter case) and has access to the angular velocity of its own pure rotation; access to its own linear velocity may also be needed in some cases. Necessary conditions for the solvability of the mission are disclosed and a distributed control strategy is proposed. Its global convergence and collision avoidance property are rigorously proved under slight and partly unavoidable enhancement of the just mentioned necessary conditions. The performance of the control law is illustrated by computer simulations.


ThC1 
Demetra 
Autonomous Robots 
Regular Session 
Chair: Bokor, Jozsef  Hungarian Acad. of Sciences 
CoChair: Sassano, Mario  Univ. of Rome, Tor Vergata 

16:5017:10, Paper ThC1.1  
Prescribed Time Scale Robot Navigation in Dynamic Environments 
Vrohidis, Constantinos  National Tech. Univ. of Athens 
Vlantis, Panagiotis  National Tech. Univ. of Athens 
Bechlioulis, Charalampos  National Tech. Univ. of Athens 
Kyriakopoulos, Kostas J.  National Tech. Univ. of Athens 
Keywords: Autonomous robots, Robotics
Abstract: In this work, we consider the problem of prescribed time scale robot navigation in dynamic environments. Initially, we treat the problem for a special class of configuration spaces, namely sphere worlds, proposing a timevarying control scheme that drives the robot from (almost) all initial configurations to an arbitrary neighborhood of any desired configuration within a predetermined time span, and at the same time prevents any collisions with static and moving obstacles as well as with the workspace boundary along the way. The introduction of a novel vector field allows us to establish the safety of the system and simultaneously apply the Prescribed Performance Control technique to guarantee any predefined transient behavior. Subsequently, we leverage wellestablished transformations to apply the proposed scheme to the far more practical class of generalized sphere worlds. Finally, we validate the theoretical findings via a nontrivial numerical simulation.


17:1017:30, Paper ThC1.2  
A Simple Optimal Planer Path Following Algorithm for Unmanned Aerial Vehicles 
Yang, Jun  Southeast Univ 
Liu, Cunjia  Loughborough Univ 
Zuo, Zongyu  Beihang Univ 
Chen, WenHua  Loughborough Univ 
Keywords: Optimization, Aerospace, UAV's
Abstract: In this paper, we present a simple optimal path following algorithm for a generic small fixedwing unmanned aerial vehicle by virtue of a predictive control approach. Different from most of exiting path following algorithms, the proposed algorithm is designed in an optimal manner where the control action is generated based on a welldefined cost function. In addition, the presented approach is designed without resorting to any complex geometric coordinate transformation. Thereby the resultant optimal control law is straightforward for practical implementation. The effectiveness of the present method is validated by three cases of simulation studies.


17:3017:50, Paper ThC1.3  
A Generalized Frame Adaptive MPC for the LowLevel Control of UAVs 
Fresk, Emil  Luleå Univ. of Tech 
Nikolakopoulos, George  Luleå Univ. of Tech. Sweden 
Keywords: Adaptive control, UAV's, Linear parametervarying systems
Abstract: The aim of this article is to establish an adaptive Model Predictive Control (MPC) scheme for the angular rate and thrust control of a multirotor Unmanned Aerial Vehicle (UAV), where the model adaptiveness comes from estimating the movement of Center of Gravity (COG) combined with the thrust constant of the motors, making the system robust to disturbances and fast to adapt to changing parameters while also taking control signal bounds under consideration to guarantee that no motor stalls while flying. The linear requirements of the MPC are adhered to by transforming the estimation and control problem into a control signal squared domain, making the system linear. The efficacy of the proposed estimation and control scheme is shown in simulations where it’s pushed to its limits.


17:5018:10, Paper ThC1.4  
Patrolling and Collision Avoidance Beyond Classical Navigation Functions 
Possieri, Corrado  Pol. Di Torino 
Sassano, Mario  Univ. of Rome, Tor Vergata 
Keywords: Autonomous robots, Robotics
Abstract: In this paper, we consider the problem of navigating a single unicyclelike robot, while avoiding obstacles in a known environment, and, at the same time, of steering the agent itself to monitor and patrol an assigned path. To this end, we propose a novel framework that combines tools and algorithms borrowed from algebraic geometry with techniques inspired by those associated with classical navigation functions. The former aspect permits the systematic construction of Lyapunov functions that certify the convergence  with an assignable decaying rate  to the desired patrolling path in the absence of obstacles. This control action is then combined with an additional term and a supervisory logic obtained by relying on the collision avoiding abilities of the underlying navigation function. Such a mixed strategy may potentially lead beyond the current understanding and implementation of classical navigation functions. The paper is then concluded by several numerical simulations that corroborate the theoretical results.


18:1018:30, Paper ThC1.5  
Monocular VisionBased Aircraft Ground Obstacle Classification 
Bauer, Peter  Inst. for Computer Science and Control, HungarianAcademyof S 
Vanek, Balint  The Computer and Automation Res. Inst. HungarianAcademy 
Bokor, Jozsef  Hungarian Acad. of Sciences 
Keywords: Signal processing, UAV's, Aerospace
Abstract: This article presents the first steps towards extending the applicability of the author’s monocular visionbased aircraft sense and avoid method for steady ground obstacles. The goal is to decide if a ground obstacle is a collision threat or not. The focus of the development is realtime on board applicability that’s why simple calculations are proposed. After extending the calculation formulae for steady obstacles the results of a softwareintheloop simulation campaign are presented for car and tower obstacles. The results are all acceptable so further developments will target a proper avoidance strategy and real flight tests.


18:3018:50, Paper ThC1.6  
Effective Angular Constrained Trajectory Generation for ThrustPropelled Vehicles 
NGUYEN, Ngoc Thinh  LCIS (Lab. of Conception and Integration of Systems), Gren 
PRODAN, Ionela  Grenoble Inst. of Tech. (Grenoble INP)  Esisar 
Lefèvre, Laurent  Grenoble Inst. of Tech. (Grenoble INP) 
Keywords: Robust control, Optimization, UAV's
Abstract: This paper proposes a novel approach to impose constraints on angles and angular velocities for the trajectory generation of thrustpropelled underactuated vehicles. The constraints formulation requires only the translational acceleration and jerk values, without requiring the vehicle's attitude. Hence, the proposed angular constraints are satisfied even if the direction angle is designedly modified or it varies uncontrollably under unexpected events. Discussions over simulation results and comparisons with the existing methods in the literature highlight the benefits of our approach.


ThC2 
Ares 
Network Analysis and Control 
Regular Session 
Chair: Zampieri, Sandro  Univ. Di Padova 
CoChair: Mousavi, Shima Sadat  Sharif Univ. of Tech 

16:5017:10, Paper ThC2.1  
On the Relation between NonNormality and Diameter in Linear Dynamical Networks 
Baggio, Giacomo  Univ. of Padova 
Zampieri, Sandro  Univ. Di Padova 
Keywords: Network analysis and control, Control over networks, Agents networks
Abstract: Understanding how the "degree" of nonnormality of a networked system is connected with the topological structure of the underlying graph is of crucial importance in many areas of the engineering and natural sciences, most notably in the controllability analysis of largescale networks. This paper explores this relation in terms of the graph diameter. More precisely, we derive diameterdependent upper and lower bounds on network nonnormality. Further, we outline a gradientbased optimization procedure to increase the nonnormality of a network.


17:1017:30, Paper ThC2.2  
Null Space Strong Structural Controllability Via Skew Zero Forcing Sets 
Mousavi, Shima Sadat  Sharif Univ. of Tech 
Chapman, Airlie  Univ. of Washington 
Haeri, Mohammad  Sharif Univ. of Tech 
Mesbahi, Mehran  Univ. of Washington 
Keywords: Network analysis and control, Control over networks, Agents networks
Abstract: In this paper, we examine strong structural controllability of a particular subspace of linear timeinvariant networks, namely, the null space of the parameterized family of system matrices. In this direction, we establish a onetoone correspondence between the set of input nodes for null space controllability and the notion of skew zero forcing sets. Using this class of zero forcing sets, we provide conditions to guarantee that the null space of the parameterized set of state matrices sharing a common network topology is trivial. Moreover, the uncontrollability of the zero mode of directed and undirected networks from a single node is discussed. In addition, methods for growing a network while preserving strong structural controllability of its null space from a set of control nodes is presented. Finally, we provide an application of the developed results for the bipartite consensus dynamics.


17:3017:50, Paper ThC2.3  
Zero Controllability in DiscreteTime Structured Systems 
van der Woude, Jacob  Delft Univ. of Tech 
Keywords: Network analysis and control, Control over networks, Linear systems
Abstract: In this paper we consider complex dynamical networks modeled by means of state space systems running in discrete time. We assume that the dependency structure of the variables within the (nonlinear) network equations is known and use directed graphs to represent this structure. The dependency structure also appears in the equations of a linearization of the network. In order for such a linearization to be a good approximation of the original network, its state should stay as close as possible to the point of linearization. In this paper, we investigate how the latter can be achieved by an appropriate selection of states as driver nodes, so that through these driver nodes the whole state of the network can be steered to the point of linearization. We present conditions in graph terms for this to be possible and deriver an algorithm for the associate driver node selection. By means of a simple reasoning, we show that finding such a selection of smallest size comes down to solving a minimal cover problem, which is known to be an NPhard problem.


17:5018:10, Paper ThC2.4  
Controllability Degree of Directed Line Networks: Nodal Energy and Asymptotic Bounds 
Zhao, Shiyu  Univ. of Sheffield 
Pasqualetti, Fabio  Univ. of California, Riverside 
Keywords: Network analysis and control, Control over networks
Abstract: This paper studies the controllability degree of dynamical networks with continuoustime dynamics. To quantify the controllability degree of a network, we introduce a new notion termed nodal energy, which is the control energy required to change the state of a single node while keeping the final states of the other nodes unchanged. Since it is extremely challenging to analyze the nodal energy of general networks, this paper focuses on a special class of directed line networks with single control nodes. This choice allows us to derive the explicit expression of the nodal energy of different network nodes, and hence, reveal novel controllability properties of complex networks. Our analysis shows that (i) differently from their discretetime counterpart, directed line networks with continuoustime dynamics are always difficult to control, as the control energy grows linearly or even exponentially with the network cardinality, (ii) the numerical investigation of the controllability degree of line networks is unreliable, because the condition number of the controllability Gramian grows exponentially fast as the network cardinality increases, and (iii) nodal energies may be inversely related to the graphical distance from the control node because, depending on the network weights, distant nodes may require far less nodal energy than immediate neighbors.


18:1018:30, Paper ThC2.5  
Reachability and Controllability of Delayed Switched Boolean Control Networks 
Yerudkar, Amol  Univ. of Sannio 
Del Vecchio, Carmen  Univ. of Sannio 
Singh, Navdeep  Veermata Jijabai Tech. Inst 
Glielmo, Luigi  Univ. of Sannio 
Keywords: Network analysis and control, Genetic regulatory systems, Switched systems
Abstract: In this paper we investigate the reachability and controllability of delayed switched Boolean control networks (DSBCNs). By resorting to the algebraic state space representation method built using semitensor product (STP) of matrices, we provide several necessary and sufficient conditions for these properties to hold which are based on inputstate incidence matrix carrying entire network dynamics information. Also, to realize the reachability of DSBCNs in shortest time, an algorithm is presented which finds switching and control sequences forcing initial state trajectory to destination state. At last, an example is given to illustrate the main results.


18:3018:50, Paper ThC2.6  
Controllability Analysis of Threshold Graphs and Cographs 
Mousavi, Shima Sadat  Sharif Univ. of Tech 
Haeri, Mohammad  Sharif Univ. of Tech 
Mesbahi, Mehran  Univ. of Washington 
Keywords: Network analysis and control, Control over networks, Concensus control and estimation
Abstract: In this paper, we investigate the controllability of a linear timeinvariant network following a Laplacian dynamics defined on a threshold graph. In this direction, by considering the modal matrix associated with the Laplacian matrix for this class of graphs and based on the PopovBelevitchHautus criteria, a procedure for the selection of control nodes is proposed. The procedure involves partitioning the nodes of the graph into cells with the same degree; one node from each cell is then selected. We show that the remaining nodes can be chosen as the control nodes, rendering the network controllable. Finally, we consider a wider class of graphs, namely cographs, and examine their controllability properties.


ThC3 
Aphrodite 
Discrete Event Systems and Supervisory Control 
Regular Session 
Chair: Gong, Weibo  Univ. of Massachusetts at Amherst 
CoChair: Sarkar, Arnab  Indian Inst. of Tech. Guwahati 

16:5017:10, Paper ThC3.1  
On Concept Abstraction Algorithms 
Gong, Weibo  Univ. of Massachusetts at Amherst 
Keywords: Discrete event systems, Modeling, Intelligent systems
Abstract: We address two fundamental issues in the computational theory of mind (CTM): the extreme accuracy of visual signal perception and the quick recall of relevant information from a vast memory. The key component of our approach is an algorithm that converts the spatial signal into a temporal representation. The model is motivated by human vision, but it is valid for concept abstraction and analogical thinking in general.


17:1017:30, Paper ThC3.2  
Maximizing the Electrical Efficiency of a Solid Oxide Fuel Cell System 
Dolenc, Bostjan  Jozef Stefan Inst 
Vrancic, Damir  Jožef Stefan Inst 
Vrecko, Darko  Department of Systems and Control, Jozef Stefan Inst 
Juričić, Đani  Jožef Stefan Inst 
Keywords: Supervisory control, Energy systems, Process control
Abstract: The solid oxide fuel cells (SOFCs) represent a promising technology for sustainable power generation from hydrogen rich fuels with high efficiency of energy conversion. However, only a limited number of papers address the problem of online maximisation of the efficiency of SOFC operation. Optimal operating conditions are normally chosen either based on experience or by using elaborated models, which are not easy to obtain. Moreover, the process changes over time due to degradation, hence to modelbased performance optimisation requires online updating of the model. To avoid these problems, a modelfree approach is proposed. It is realised in the form of a twotier control structure where the lowlevel controllers take over control of the auxiliary units around the fuel cells, while the supervisory controller (SC) controller optimises the operation point of the system. The lowlevel controllers are conventional feedforward feedback controllers, while optimisation on the higher level is solved by using the extremum seeking approach.The proposed control system is demonstrated on a simulated 10 kW SOFC system showing reliable convergence, relative rise of efficiency up to 2% and easy design and maintenance.


17:3017:50, Paper ThC3.3  
Tracking Control for Petri Nets with Forbidden States 
Fritz, Raphael  Tech. Univ. Kaiserslautern 
Zhang, Ping  Univ. of Kaiserslautern 
Keywords: Petri nets, Discrete event systems, Supervisory control
Abstract: In this paper, an approach for tracking control for Petri nets with forbidden states is presented. The control aim is that the Petri net is forced by a firing sequence from an initial marking into a destination marking while avoiding the forbidden states described by Generalized Mutual Exclusion Constraints (GMEC). A new type of constraints called NOTGMEC is introduced to allow a compact representation of specific tokenplace combinations. The firing sequence is determined by transforming the constraints into a suitable form and executing a twostep algorithm.


17:5018:10, Paper ThC3.4  
EventBased Observer and MPC with Disturbance Attenuation Using ERM Learning 
Yoo, Jaehyun  KTH Royal Inst. of Tech 
Nekouei, Ehsan  KTH Royal Inst. of Tech 
Johansson, Karl Henrik  Royal Inst. of Tech 
Keywords: Autonomous systems, Discrete event systems, Identification for control
Abstract: This paper presents a learningbased approach for disturbance attenuation for a nonlinear dynamical system with eventbased observer and model predictive control. Using the empirical risk minimization method, we can obtain a learning error bound which is the function of the number of samples, learning parameters, and the model complexity. It enables us to analyze the closedloop stability in terms of the learning property and derive an upper bound on the the state estimation error which depends on the risk of the learning algorithm. Simulation results underline the learning’s capability, the control performance and the eventtriggering efficiency in comparison to the conventional eventtriggered control scheme.


18:1018:30, Paper ThC3.5  
Supervisory Controller for a LNTSCR Diesel Exhaust AfterTreatment System 
Velmurugan, Dhinesh  Volvo Car Corp 
McKelvey, Tomas  Chalmers Univ. of Tech 
Lundberg, Daniel  Volvo Car Corp 
Keywords: Supervisory control, Automotive, Optimization
Abstract: Statistical analysis of route history and online traffic information system can provide realtime lookahead information regarding the route ahead which could be used for powertrain optimisation. A diesel engine NOx Exhaust AfterTreatment System (EATS) for a passenger car application comprised of an engine closecoupled Lean NOx Trap (LNT) and an underfloor Selective Catalytic Reduction (SCR) is studied. Conventionally, the LNTSCR operation is coordinated using a rule based controller that primarily utilises the SCR catalyst bed temperature. This paper presents a supervisory control structure that uses look ahead information to improve the performance of the EATS coordinator. Therefore, the supervisory control based EATS coordinator is parameterised with respect to the look ahead data. The parameterised controller calculates setpoints for the NOx EATS based on Emission Equivalent Fuel Consumption (EEFC). The Supervisory control performance using the EEFC strategy is analysed for the Worldwide harmonized Light vehicles Test Cycle (WLTC). A simulation environment that has been benchmarked with data from the production system was used to carry out the evaluation and compare with the baseline controller. The paper explores a method to utilise a supervisory control structure for the EATS coordinator in an Engine Control Unit. Subsystem synergies that could be harnessed using the supervisory control approach are demonstrated for the EATS. The future work will focus on extending the approach to more subsystems and characterising the look ahead information.


18:3018:50, Paper ThC3.6  
Exact Task Completion Time Aware RealTime Scheduling Based on Supervisory Control Theory of Timed DES 
Devaraj, Rajesh  Indian Inst. of Tech. Guwahati 
Sarkar, Arnab  Indian Inst. of Tech. Guwahati 
Biswas, Santosh  IIT GuwahatiINDIA 
Keywords: Discrete event systems, Automata, Supervisory control
Abstract: Realtime scheduling strategies for safetycritical applications are primarily focused towards ensuring correctness, both functional and temporal. However, in order to guarantee deadlines for critical tasks, such strategies tend to assume very conservative Worst Case Execution Time (WCET) estimates for tasks, thus resulting in poor resource utilization. One way of improving this situation is through the evolution of scheduling mechanisms that can reclaim resources which are provisioned for critical functionalities at design time, but remain unused at runtime. Such reclaimed resources may prove to be very useful in executing lower criticality best effort tasks. In this work, we present a Supervisory Control Theory of Timed Discrete Event Systems (SCT of TDES) based formal mechanism for synthesizing nonpreemptive schedulers which can recognize exact completion times of tasks at runtime and allow the processor to be immediately relinquished instead of waiting up to WCETs of tasks. This approach allows reclamation of a higher amount of resources at runtime in comparison to models which consider worstcase execution time only. Conducted experiments have shown promising results and indicate to the practical efficacy of our approach.


ThC4 
Poseidon 
Output Regulation 
Invited Session 
Chair: Scarciotti, Giordano  Imperial Coll. London 
CoChair: Wang, Lei  Nanyang Tech. Univ 
Organizer: Scarciotti, Giordano  Imperial Coll. London 

16:5017:10, Paper ThC4.1  
Robust Output Regulation for a Class of Nonlinear Systems Not Detectable by Regulated Output (I) 
Wang, Lei  Nanyang Tech. Univ 
Wen, Changyun  Nanyang Tech. Univ 
Marconi, Lorenzo  Univ. Di Bologna 
Su, Hongye  Zhejiang Univ 
Keywords: Output regulation, Lyapunov methods, Stability of nonlinear systems
Abstract: This paper considers the robust output regulation problem for a class of nonlinear systems, not detectable by the regulated output. Differently from the previous results on the subject, it is supposed that the considered class of systems is in the normal form not defined on the conventional regulated output. We propose a novel internal model structure that, joined to a highgain stabilizing action, solves the output regulation problem robustly with respect to possible uncertainties affecting the systems.


17:1017:30, Paper ThC4.2  
Output Regulation of Linear Stochastic Systems: The FullInformation Case (I) 
Scarciotti, Giordano  Imperial Coll. London 
Keywords: Output regulation, Stochastic control, Stochastic systems
Abstract: The full information output regulation problem for linear stochastic systems is addressed. A general class of linear systems is considered, namely systems in which the state, control variable and exogenous variable may appear simultaneously in the drift term and in the diffusion term of the differential equation. Similarly, we consider a stochastic signal generator, thus allowing tracking and/or rejecting Brownian motions in addition to deterministic trajectories. In the paper we first characterize the steadystate response of the interconnection of the system with the signal generator and then we solve the full information output regulation problem. The results of the paper are illustrated by means of two examples. Finally a short discussion of the error feedback regulator problem concludes the paper.


17:3017:50, Paper ThC4.3  
Output Robust Controller Design for InputSaturated Robotic Boat with Disturbance Cancellation 
Borisov, Oleg  ITMO Univ 
Gromov, Vladislav  ITMO Univ 
Vlasov, Sergey  ITMO Univ 
Somov, Sergey  ITMO Univ 
Pyrkin, Anton  ITMO Univ 
Keywords: Robust control, Robotics, Output regulation
Abstract: In this paper a problem of saturated control for a robotic boat with unknown parameters and unmeasurable velocity and acceleration is addressed. The controller design is based on the output robust control approach consecutive compensator. It was augmented with an internal model, which allows to eliminate a static error and implement the antiwindup scheme to reduce overshoot of the output variable. As result, the regulator generating the bounded control signal and avoiding windup for the boat was obtained. The proof of closedloop system stability is presented. The efficiency of the proposed algorithm was illustrated by series of experiments using the setup with robotic boat. The experimental results and comparison between three types of controllers (regular consecutive compensator, integral modification and one equipped with antiwindup) are presented.


17:5018:10, Paper ThC4.4  
EmulationBased Semiglobal Output Regulation of Minimum Phase Nonlinear Systems with Sampled Measurements (I) 
Astolfi, Daniele  Univ. De Lorraine 
Casadei, Giacomo  Univ. Grenoble Alpes 
Postoyan, Romain  CNRS 
Keywords: Output regulation, Control over communication, Hybrid systems
Abstract: We investigate the semiglobal output regulation of minimumphase singleinput singleoutput nonlinear systems with sampled measurements. We proceed by emulation. We start by considering a continuoustime regulator, which solves the problem in the absence of sampling. Then, we consider sampled measurements and we model the overall system as a hybrid system. We show that the original continuouscase properties are preserved when the measurements are sampled provided that the maximum allowable transmission interval satisfies a given explicit bound.


18:1018:30, Paper ThC4.5  
Output Regulation of MultiInput Systems under Packet Dropout with Application to Trajectory Tracking of Cooperative Robots 
Ling, Rongyao  Zhejiang Univ. of Tech 
CLAVEAU, Fabien  IRCCyN (UMR CNRS 6597)  Ec. Des Mines De Nantes 
Feng, Yu  Zhejiang Univ. of Tech 
Chevrel, Philippe  IRCCyN / Ec. Des Mines De Nantes 
Keywords: Control over communication, Stochastic control, Cooperative autonomous systems
Abstract: This paper addresses the output regulation problem for discretetime linear systems over lossy actuating channels. In order to solve this problem, a codesign approach of controller, encoder and decoder is adopted, i.e., an encoder matrix and a decoder matrix are introduced at each end of the controlleractuator channels for taking full advantage of the communication resource. The solvability conditions are firstly derived to design the feedback gain such that the closedloop system is meansquare (MS) stable. Then the regulation equations are deduced for designing the feedforward gain to drive the controlled output following the desired trajectory. Finally, a simulation is performed to implement the design method on cooperative robots, thereby showing the effectiveness of the proposed results.


18:3018:50, Paper ThC4.6  
Output Regulation for Redundant Plants Via Orthogonal Moments (I) 
Sassano, Mario  Univ. of Rome, Tor Vergata 
Galeani, Sergio  Univ. Di Roma Tor Vergata 
Keywords: Output regulation, Optimization, Linear systems
Abstract: In this paper we address the output regulation problem for redundant plants, namely systems possessing more control inputs than regulated outputs, with focus on the specific opportunity and difficulty deriving from the additional inputs. The former consists in the fact that, having a wealth of input configurations achieving the same steadystate behavior, it is possible to optimize additional performance criteria while preserving the primary task of output regulation. The latter stems from the fact that the naive approach of replicating the required internal model of the exosystem on each input channel leads to loss of observability/detectability of the cascaded interconnection of the internal model and the plant, thus preventing the achievement of overall closedloop stability. The main result of this paper consists in the design of an inner auxiliary control loop that allows to break the above mentioned conflict between advantages and drawbacks of redundant plants. The result is then revisited by exploiting the orthogonal moments of the plant at the exosystem's frequencies. Differently from classic moments, which, for an asymptotically stable plant, describe the relation between inputs and steadystate output response at given frequencies, orthogonal moments characterize the input directions that yield zero steadystate output response at given frequencies.


ThC5 
Athenaeum 1 
Observers III 
Regular Session 
Chair: Besancon, Gildas  Ense3  Grenoble INP 
CoChair: Efimov, Denis  Inria 

16:5017:10, Paper ThC5.1  
Regularization Approach for an ImmersionBased Observer Design 
Besancon, Gildas  Ense3  Grenoble INP 
Ticlea, Alexandru  Pol. Univ. of Bucharest 
Keywords: Observers for nonlinear systems
Abstract: This paper is about immersionbased observer design for nonlinear systems. Starting from the fact that direct immersion may result in stability issues, which may in turn affect the observer performance, a method is proposed to overcome such problems, by somehow ’stabilizing’ the transformation (or ’regularizing’ it). A couple of examples are presented to illustrate the effectiveness of the method, including the challenging case of state and parameter estimation in a speed sensorless induction machine.


17:1017:30, Paper ThC5.2  
Output Injection Filtering Redesign in HighGain Observers 
Astolfi, Daniele  Univ. De Lorraine 
JUNGERS, Marc  CNRS Univ. De Lorraine 
Zaccarian, Luca   
Keywords: Observers for nonlinear systems
Abstract: We propose a new paradigm to redesign highgain observers in order to improve performances in the presence of measurement noise. In particular, instead of driving the observer by means of a standard output injection term, we filter it with a dynamical system having good filtering properties. In this first preliminary result we also select the filter in order to address numerical challenges.


17:3017:50, Paper ThC5.3  
Comparison of the TimeDelay Margin of a Distributed and Centralized Observer 
Silm, Haik  Centrale Lille / KU Leuven 
Ushirobira, Rosane  Inria 
Efimov, Denis  Inria 
Michiels, Wim  KU Leuven 
Richard, JeanPierre  Ec. Centrale De Lille 
Fridman, E. M.  TelAviv Univ 
Keywords: Observers for linear systems, Delay systems
Abstract: The benefit of a distributed observer concept for largescale linear plants is shown by taking timedelays into account. It is asserted that a centralized observer suffers from delays in the measurement input, while a distributed structure allows to avoid them. In contrast, the network of the distributed observer induces communication delays among observer nodes. The delay margins for both observer concepts are estimated on a numerical example and compared using an eigenvaluebased frequency domain approach and an LMI based timedomain approach.


17:5018:10, Paper ThC5.4  
A Class of Nonlinear Adaptive Observers for SIR Epidemic Model 
Bliman, PierreAlexandre J  Inria, France and Fundação Getulio Vargas, Rio De Janeiro, Brazi 
Efimov, Denis  Inria 
Ushirobira, Rosane  Inria 
Keywords: Biological systems, Observers for nonlinear systems
Abstract: Mathematical epidemic models describe the spread of an infectious disease in a host population. The SIR model, which is one of the simplest, is based on the representation of interactions between three compartments in the population: the number of susceptible, infective and recovered individuals. In this note we study the problem of state estimation for such a model, subject to seasonal variations and uncertainties in the measured incidence rate (assuming continuous measurement), and design for this purpose a class of nonlinear adaptive observers. Asymptotic stability and robustness with respect to variation rates are ensured by an appropriate choice of the gain components as a function of the state estimate, through the use of the theory of inputtooutput stability. Numerical experiments are presented to illustrate the method efficiency.


18:1018:30, Paper ThC5.5  
Interval Observer for Uncertain TimeVarying SIRSI Epidemiological Model of VectorBorne Disease 
Aronna, Maria Soledad  Fundação Getúlio Vargas 
Bliman, PierreAlexandre J  Inria, France and Fundação Getulio Vargas, Rio De Janeiro, Brazi 
Keywords: Biological systems, Observers for nonlinear systems
Abstract: The issue of state estimation is considered for an SIRSI epidemiological model describing a vectorborne disease such as dengue fever, subject to seasonal variations. Assuming continuous measurement of the incidence rate (that is the number of new infectives in the host population per unit time), a class of interval observers with estimatedependent gain is constructed, and asymptotic error bounds are provided. The synthesis method is based on the search for a common linear Lyapunov function for monotone systems that represent the evolution of the estimation errors.


18:3018:50, Paper ThC5.6  
Observation of the Nuclear Reactor Using WaveletBased Extended Kalman Filter 
PATEL, SHRENIK  HOMI BHABHA NATIONAL Inst 
Keywords: Observers for nonlinear systems, Stochastic filtering, Wavelets
Abstract: An online algorithm for the estimation of state variables of a nuclear reactor is presented in this paper. In the reactor core design, distribution of average fuel temperature, coolant temperature, delayed neutron precursors’ concentration, and thermal hydraulic behavior play an important role. The associated reactivity feedback induced by the core temperature distributions further affect the control design. Therefore, the knowledge of reactor core temperature is crucial for the effective thermal power control. The proposed algorithm is based on and preserves the merits of the standard extended Kalman filtering (EKF) technique. While the application of stationary wavelet transform effectively captures the multiscale dynamics of the system. The efficacy of the proposed algorithm is demonstrated by simulation results using point kinetics model of the nuclear reactor.


ThC6 
Athenaeum 2 
Stabilization and SampledData Control 
Regular Session 
Chair: Etienne, Lucien  IMT LilleDouai 
CoChair: Ferrante, Francesco  GipsaLab and Univ. Grenoble Alpes 

16:5017:10, Paper ThC6.1  
Generalized Feedback Homogenization and Stabilization of Linear MIMO Systems 
Zimenko, Konstantin  ITMO Univ 
Polyakov, Andrey  INRIA Lille NordEurope 
Efimov, Denis  Inria 
Perruquetti, Wilfrid  Ec. Centrale De Lille 
Keywords: Lyapunov methods, Stability of nonlinear systems
Abstract: Generalized homogenization of linear MIMO systems via linear feedback is introduced. The control algorithm for finitetime (or asymptotic) stabilization of linear MIMO systems via homogenization technique is developed. The robustness of the control algorithm with respect to system uncertainties and disturbances is studied. The theoretical results are supported by numerical examples.


17:1017:30, Paper ThC6.2  
EventTriggered Control for a Class of Cascade Systems 
Theodosis, Dionysios  Royal Inst. of Tech. (KTH) 
Dimarogonas, Dimos V.  KTH Royal Inst. of Tech 
Keywords: Lyapunov methods, Stability of nonlinear systems
Abstract: This paper addresses the eventtriggered control of cascade connected systems. In particular, we present eventtriggered mechanisms that guarantee the stabilization of cascade systems with partial state feedback without infinitely fast sampling. Our approach is based on growth conditions on the interconnection terms and does not follow the framework of inputtostate stability with respect to the subsystems and/or measurement errors.


17:3017:50, Paper ThC6.3  
On Myopic Strategies for Resource Constrained Informative Sampling 
Emadi, Hamid  Iowa State Univ 
Bhattacharya, Sourabh  Iowa State Univ 


17:5018:10, Paper ThC6.4  
Stability Analysis and Gain Synthesis for Lipschitz Non Linear Systems under Dynamic Event Triggered Sampling 
Etienne, Lucien  IMT LilleDouai 
Di Gennaro, Stefano  Univ. of L'Aquila 
Barbot, Jean Pierre  ENSEA 
Keywords: Sampled data control, LMI's/BMI's/SOS's, Stability of hybrid systems
Abstract: In this paper, we investigate the stabilization of a Lipschitz non linear plant under the assumption of dynamic event triggered sampling. Sufficient conditions, making use of a hybrid Lyapunov function and convex embedding, are given to guarantee the existence of a triggering mechanism, leading to asymptotic stability. Both gain analysis and gain synthesis are considered.


18:1018:30, Paper ThC6.5  
Hybrid Regional Stabilization of Linear Systems with Actuator Saturation and MultiRate Samplers 
Ferrante, Francesco  GipsaLab and Univ. Grenoble Alpes 
Sanfelice, Ricardo  Univ. of California, Santa Cruz 
Tarbouriech, Sophie  LAASCNRS 
Keywords: Sampled data control, LMI's/BMI's/SOS's, Stability of hybrid systems
Abstract: Regional stability analysis of linear systems with multirate samplers and actuator saturation is studied. A hybrid controller is used to perform a fusion of measurements sampled at different times. In between sampling events, the controller behaves as a copy of the plant. When a new measurement is available, the controller state undergoes a jump. The resulting system is analyzed in a hybrid system framework. Sufficient conditions in the form of matrix inequalities are given to determine estimates of the basin of attraction of the closedloop system. Finally, the effectiveness of the proposed methodology is shown in an example.


18:3018:50, Paper ThC6.6  
Regional Stability Analysis of Nonlinear SampledData Control Systems: A QuasiLPV Approach 
Kimura Palmeira, Alessandra Helena  Univ. Federal Do Rio Grande Do Sul 
Gomes Da Silva Jr., Joao Manoel  Univ. Federal Do Rio Grande Do Sul (UFRGS) 
Flores, Jeferson V.  Univ. Federal Do Rio Grande Do Sul 
Keywords: Sampled data control, Linear parametervarying systems, Stability of nonlinear systems
Abstract: This paper addresses the stability analysis of sampleddata control for a class of continuoustime nonlinear systems. The proposed approach is based on a local quasiLPV model for the nonlinear system and the use of a parameter dependent looped functional to deal with the aperiodic sampling effects. From these ingredients, LMI conditions are proposed to assess local stability. These conditions are then incorporated in convex optimization problems aiming at obtaining maximized estimates of the region of attraction of the origin or maximizing the intersampling time for which the stability is regionally ensured.


ThC7 
Athenaeum 3 
Biological Systems and Applications in Neuroscience 
Regular Session 
Chair: Sepulchre, Rodolphe J.  Univ. of Cambridge 
CoChair: Proskurnikov, Anton  Delft Univ. of Tech 

16:5017:10, Paper ThC7.1  
Gaussian MeanField Models of Linear Systems 
Seslija, Marko  Cambridge Univ 
Sepulchre, Rodolphe J.  Univ. of Cambridge 
Keywords: Modeling, Linear systems, Applications in neuroscience
Abstract: This paper addresses the issue of modeling meanfield behavior in heterogeneous populations of linear timeinvariant SISO systems. Our analysis is conducted in the frequency domain, where the heterogeneity of inputoutput mappings (transfer functions) is modeled as a complexvalued Gaussian process. The meanfield model of diffusively coupled agents is obtained as a Gaussian approximation of averaged inputoutput behavior. It is shown that the strong coupling and the large number of agents reduce the population variance.


17:1017:30, Paper ThC7.2  
Partial Phase Cohesiveness in Networks of Communitinized Kuramoto Oscillators 
Qin, Yuzhen  Univ. of Groningen 
Kawano, Yu  Univ. of Groningen 
Cao, Ming  Univ. of Groningen 
Keywords: Applications in neuroscience, Agents networks, Nonlinear system theory
Abstract: Partial synchronization of neuronal ensembles are often observed in the human brain, which is believed to facilitate communication among anatomical regions demanded by cognitive tasks. Since such neurons are commonly modeled by oscillators, to better understand their partial synchronization behavior, in this paper we study communitydriven partial phase cohesiveness in networks of communitinized Kuramoto oscillators, where each community itself consists of a population of alltoall coupled oscillators. Sufficient conditions on the algebraic connectivity of the selected communities are obtained to guarantee the appearance of their phase cohesiveness, while leaving the remaining communities incoherent. These conditions are further reduced to the form of the lower bounds on the coupling strengths for the connections linking the selected communities. We also show that the ultimate level of the phase cohesiveness that the oscillators asymptotically converge to is predictable. Finally, numeral studies are performed to validate the obtained results.


17:3017:50, Paper ThC7.3  
The Impact of Deep Brain Stimulation on a Simulated Neuron: Inhibition, Excitation, and Partial Recovery 
Andersson, Helena  Uppsala Univ 
Medvedev, Alexander V.  Uppsala Univ 
Cubo, Ruben  Uppsala Univ 
Keywords: Applications in neuroscience, Biomedical systems, Modeling
Abstract: Deep Brain Stimulation (DBS) is an established therapy to alleviate the symptoms of neurological disorders such as Parkinson's Disease and Essential Tremor. Depending on the disease, a certain area of the brain is subjected to electrical stimuli through a surgically implanted lead. Despite the clinically proven effectiveness of DBS, the underlying biological mechanism is poorly understood. Two dominating theories seek to explain how the DBS therapy exerts effect on neurons, one through inhibition and another through excitation. This simulation study aims at demonstrating that both scenarios are feasible within the brain domain exposed to pulsatile electrical stimulation and conditional on the temporal relationship between the neural input and the DBS pulse sequence. Since some neurons in the targeted population are assisted to fire in response to the cumulative dynamical action of a stimulation pulse and neural input from a neighbouring neuron, a partial function recovery of the neural network is expected. Simulations with a spatially distributed deterministic neuron model support the presented hypothesis and provide insights into the role of DBS frequency and pulse width (duty cycle) in restoring the neural processing ability. The obtained results highlight the role of the phase difference between the neural input and the DBS pulse sequence in the neuron's response to stimulation.


17:5018:10, Paper ThC7.4  
Numerical Simulations of TwoDimensional Neural Fields with Applications to Working Memory 
Lima, Pedro Miguel  Univ. of Lisbon 
Erlhagen, Wolfram  Univ. of Minho, Campus Azurém 
Keywords: Computational methods, Robotics, Applications in neuroscience
Abstract: In this paper we describe a neural field model which explains how a population of cortical neurons may encode in its firing pattern simultaneously the nature and time of sequential stimulus events. From the mathematical point of view, this is obtained my means of a twodimensional field, where one dimension represents the nature of the event (for example the color of a light signal) and the other represents the elapsed time. Some numerical experiments are reported which were carried out using a computational algorithm for twodimensional neural field equations. These numerical experiments are described and their results are discussed.


18:1018:30, Paper ThC7.5  
Entrainment in Harmonically Forced Continuous and Impulsive Goodwin's Oscillators: A Comparison Study 
Medvedev, Alexander V.  Uppsala Univ 
Proskurnikov, Anton  Delft Univ. of Tech 
Zhusubaliev, Zhanybai  SouthWest State Univ 
Keywords: Biological systems, Modeling, Chaotic systems
Abstract: The Goodwin oscillator is a simple yet illustrative model of a biochemical system with a stable limit cycle. Considered as a prototypical biological oscillator, Goodwin's model is broadly used e.g. to describe circadian rhythms, hormonal cycles, selfoscillatory metabolic pathways. These periodic or nonperiodic oscillations are selfsustained; at the same time, they are entrainable by external periodic signals, adjusting the characteristics of the autonomous oscillatory behavior. Mathematical analysis of entrainment phenomena, i.e. nonlinear phenomena imposed by periodic exogenous signals, remains an open problem. This paper presents a comparative analysis of forced dynamics arising in two versions of Goodwin's oscillator: the classical continuous oscillator and a more recent impulsive one, e.g. capturing pulsatile secretion of hormones. The main finding of this study is that while the continuous oscillator is always forced to a periodic solution by a sufficiently large exogenous signal amplitude, the impulsive one commonly exhibits a quasiperiodic or chaotic behavior thus highlighting the role of nonsmooth dynamics in entrainment.


18:3018:50, Paper ThC7.6  
On Synchronization in FitzHughNagumo Networks with Small Delays 
Plotnikov, Sergei  Inst. for Problems of Mechanical Engineering 
Fradkov, Alexander L.  Acad. of Sciences of Russia 
Keywords: Concensus control and estimation, Biological systems
Abstract: This article studies the influence of the small delays on the FitzHughNagumo network synchronization. It is widely known that high delays in signal propagation between the nodes make synchronization difficult or even impossible. The sufficient conditions of the linearized network synchronization for the case of the small delay are obtained. This problem is successfully reduced to the feasibility of the LMIs. The simulation results confirm the efficiency of the obtained conditions. We suppose that the similar conditions can be applicable even to the nonlinear FitzHughNagumo network.


ThC8 
Athenaeum 4 
Learning and Computational Intelligence Techniques 
Regular Session 
Chair: Ferrara, Antonella  Univ. of Pavia 
CoChair: Bombara, Giuseppe  Boston Univ 

16:5017:10, Paper ThC8.1  
Online Learning of Temporal Logic Formulae for Signal Classification 
Bombara, Giuseppe  Boston Univ 
Belta, Calin  Boston Univ 
Keywords: Machine learning, Fault detection and identification, Complex systems
Abstract: This paper introduces a method for online inference of temporal logic properties from data. Specifically, we tackle the online supervised learning problem. In this setting, the data is in form of a set of pairs of signals and labels and it becomes available over time. We propose an approach for efficiently processing the data incrementally. In particular, when a new instance is presented, the proposed method updates a binary tree that is linked with the inferred Signal Temporal Logic (STL) formula. This approach presents several benefits. Primarily, it allows the refinement of the current formula when more data is acquired. Moreover, the incremental construction offers insights on the tradeoff between formula complexity and classification accuracy. We present two case studies to emphasize the characteristics of the proposed algorithm: 1) a fault classification problem in an automotive system and 2) an anomaly detection problem in the maritime environment.


17:1017:30, Paper ThC8.2  
Deep Reinforcement Learning for Collision Avoidance of Robotic Manipulators 
Sangiovanni, Bianca  Univ. of Pavia 
Rendiniello, Angelo  Univ 
Incremona, Gian Paolo  Pol. Di Milano 
Ferrara, Antonella  Univ. of Pavia 
Piastra, Marco  Univ. of Pavia 
Keywords: Machine learning, Robotics, Safety critical systems
Abstract: In this paper a realtime collision avoidance approach using machine learning is presented for safe humanrobot coexistence. More specifically, the collision avoidance problem is tackled with Deep Reinforcement Learning (DRL) techniques, applied to robot manipulators with a workspace invaded by unpredictable obstacles. Since the robotic systems are defined in the continuous space, a Normalized Advantage Function (NAF) modelfree algorithm has been used. In order to assess the proposal, a robotic system, that is a COMAU SMART3S2 anthropomorphic robot manipulator, has been considered. The robotic system has been interfaced with external tools for evaluation, control, and automatic training. Simulations carried out on a virtual environment are finally reported to show the effectiveness of the proposed modelfree deep reinforcement learning algorithm.


17:3017:50, Paper ThC8.3  
System Identification and Indirect Inverse Control Using Fuzzy Cognitive Networks with Functional Weights 
Karatzinis, Georgios  Democritus Univ. of Thrace 
Boutalis, Yiannis  Democritus Univ. of Thrace 
Kottas, Thodoris  Democritus Univ. of Thrace, Greece 
Keywords: Fuzzy systems, Adaptive control, Nonlinear system identification
Abstract: A Fuzzy Cognitive Network (FCN) is an operational extension of a Fuzzy Cognitive Map (FCM) which assumes, ﬁrst, that it always converges to equilibrium points during its operation and second, it is in continuous interaction with the system it describes and may be used to control it. In this paper we show that the conditions that guarantee the convergence of the FCN may lead to a special, yet very powerful, form of the network that assumes functional interconnection weights with excellent system approximation abilities. Assuming that the plant is unknown it is initially approximated by a FCN and a procedure for adaptive estimation of its functional weights is proposed that guarantee approximation error convergence to zero. The FCN is then used for the Indirect adaptive Inverse Control of a plant. The methodology is tested on a coupled twotank system.


17:5018:10, Paper ThC8.4  
Deep LearningBased Embedded MixedInteger Model Predictive Control 
Karg, Benjamin  TU Berlin 
Lucia, Sergio  TU Berlin 
Keywords: Predictive control for linear systems, Neural networks
Abstract: We suggest that using deep learning networks to learn model predictive controllers is a powerful alternative to online optimization, especially when the underlying problems are complex, as in the case of mixedinteger quadratic programs. The use of deep learning has two important advantages compared to classical shallow networks regarding its embedded implementation. A better function approximation can be achieved with the same number of neurons and less weights are necessary due to the use of more, but smaller, layers. This reduces significantly the memory footprint of the necessary code for its embedded implementation. As with shallow networks, deep neural networks are extremely easy to implement and to deploy on embedded platforms. The potential of the approach is illustrated with simulation results of an energy management system in a smart building, including the implementation of the proposed controller using a lowcost microcontroller.


18:1018:30, Paper ThC8.5  
Constrained Nonlinear Control Allocation Based on Deep AutoEncoder Neural Networks 
huang, huang  Beijing Inst. of Control Engineering, China Acad. of Space 
wang, wei  Inst. of Automation， Chinese Acad. of Sciences 
wei, chunling  Beijing Inst. of Control Engineering 
He, Yingzi  Beijing Inst. of Control Engineering 
Keywords: Neural networks, Aerospace
Abstract: An intelligent control allocation method for modern reentry vehicles with a strong aerodynamic nonlinearity is presented in this paper. A specially designed deep autoencoder (DAE) neural network is proposed that shares similar information flow with the control allocation process. This similarity is showed by setting the decoder to approximate the control effector function, and by setting the encoder to allocate the expected control moments. The decoder is trained independently based on the aerodynamic coefficients database, and with the help of the welltrained decoder, the encoder is then trained in an unsupervised way without labeled data. This proposed control allocation method could deal with strong nonlinearity of the control effectors at a high accuracy, thanks to the powerful modeling and regression ability of deep neural networks. Numerical examples are provided by the end that explain the training and implementation details, as well as the strong learning and modeling ability of the deep neural network.


ThC9 
Hera 
Optimal Control II 
Regular Session 
Chair: Rubio, Francisco R.  Univ. De Sevilla 
CoChair: Zeilinger, Melanie N.  ETH Zurich 

16:5017:10, Paper ThC9.1  
Optimal Altitude Control of an Integrated Airborne Wind Energy System with Globalized LyapunovBased Switched Extremum Seeking 
Bafandeh, Alireza  Univ. of North Carolina at Charlotte 
Vermillion, Christopher  Univ. of North Carolina at Charlotte 
Keywords: Lyapunov methods, Optimal control, Energy systems
Abstract: Airborne wind energy (AWE) systems replace the tower and foundation of contemporary wind turbines with tethers and a lifting body. This enables AWE systems to adjust their operating altitudes to deliver the greatest amount of net energy possible. However, determining the optimal operating altitude requires knowledge of the wind speed vs. altitude (wind shear) profile, leading to a tradeoff between exploration and exploitation. In this work, we consider an emph{integrated} AWEbatterygenerator system in which it is possible to explore the domain of admissible altitudes during periods of low load demand and exploit the best altitude at other times. Specifically, we propose and evaluate four candidate hierarchical structures, based on a globalized Lyapunovbased switched extremum seeking (GLSES) control structure, for control of the integrated system. We present simulationbased results that are based on actual wind speed and load demand data.


17:1017:30, Paper ThC9.2  
Optimal Control of Solar Thermal Plants with Energy Storage 
Navas, Sergio J.  Univ. De Sevilla 
Rubio, Francisco R.  Univ. De Sevilla 
Ollero, Pedro  Univ. De Sevilla 
Lemos, Joao M.  INESCID 
Keywords: Optimal control, Predictive control for nonlinear systems, Energy systems
Abstract: This paper describes and assesses an optimal strategy to control distributed solar collector fields with thermal storage systems, especially during days with partial radiation due to the passage of clouds. The main objective of these control strategies is to maximize the thermal energy stored during different situations in which different parts of the solar field receive different degrees of solar radiation while keeping the net electrical power produced close to its set point. Simulations were carried out using three connected models, one for the solar field (taking into account all of its loops individually), that includes the passage of clouds, another one for the thermal storage system, and finally one for the power cycle. The solar field simulated is a small demonstration plant, in which it is assumed that all the loops have the same characteristics; and the nominal power range of the Rankine cycle is 8002330kW.


17:3017:50, Paper ThC9.3  
LinearQuadratic Robust Path Tracking for a Dubins Vehicle 
Jha, Bhargav  Technion Israel Inst. of Tech. Haifa 
Turetsky, Vladimir  Ort Braude Coll 
Shima, Tal  Technion 
Keywords: Optimal control, Robotics, Autonomous robots
Abstract: Previous theoretical results on a robust linear tracking are applied to a practical problem of trajectory tracking by a Dubins’ ground vehicle. The tracking control is constructed as the optimal strategy in an auxiliary linearquadratic control problem for a linearized vehicle model. Numerical and experimental results are presented and compared. Practical improvements are proposed to circumvent real world problems such as nonideal dynamics and control saturation.Robustness of tracking method in presence of disturbances is also shown experimentally.


17:5018:10, Paper ThC9.4  
Convex Formulations and Algebraic Solutions for Linear Quadratic Inverse Optimal Control Problems 
Menner, Marcel  ETH Zurich 
Zeilinger, Melanie N.  ETH Zurich 
Keywords: Optimal control
Abstract: This paper presents convex formulations for inverse optimal control problems for linear systems to infer cost function matrices of a quadratic cost from both optimal and nonoptimal closedloop gains. It introduces an optimality measure which enables a formulation of the problem as a convex semidefinite program for the general case and a linear program for several special cases. We derive an explicit algebraic expression for general objective function matrices as well as conditions under which the solution to the inverse optimal control problem is unique. The result is derived by means of a vectorization and parametrization of the algebraic Riccati equation. A simulation example highlights the robust performance in the presence of noise on the measured closedloop gain and the computational efficiency of the proposed problem formulations.


18:1018:30, Paper ThC9.5  
LQ Optimal Tracking with Unbounded Cost for Unknown Dynamics: An Adaptive Dynamic Programming Approach 
Bernhard, Sebastian  Tech. Univ. Darmstadt 
Adamy, Juergen  Tech. Univ. Darmstadt 
Keywords: Optimal control, Adaptive control, Uncertain systems
Abstract: In case of unknown system dynamics, we consider linear quadratic optimal tracking on infinite horizons with generally unbounded cost. For the first time, we deal with this problem in the framework of adaptive dynamic programming. So far, existing methods require bounded costs which essentially limits the applicability and achievable performance. Thus, we develop a new algorithm that yields a strongly overtaking optimal control which is an adequate solution. After collecting measurement data in an exploration phase, the algorithm implicitly solves the necessary and sufficient algebraic equations in [3], but without knowledge of the dynamics. Then, implementing the control results in an optimal transition to an optimal stationary trajectory. A simulation example of almost exact tracking for an overactuated system demonstrates a highly efficient saving of inputenergy in contrast to stateoftheart approaches.


18:3018:50, Paper ThC9.6  
Optimal Synthesis in the Goddard Problem on a Constrained Time Interval 
Samylovskiy, Ivan  Lomonosov Moscow State Univ 
Keywords: Optimal control, Optimization algorithms, Aerospace
Abstract: In this paper, we consider a problem on the rocket flight final height maximization under the control and endpoint constraints. The problem is considered on a constrained time interval, filling the gap between problems on free and fixed time interval. Using Pontryagin Maximum Principle, we obtain all types of optimal trajectories and then construct optimal synthesis. In addition, we investigate evolution of obtained synthesis w.r.t problem parameters.


ThC10 
Hermes 
Consensus and Cooperative Control 
Regular Session 
Chair: Valcher, Maria Elena  Univ. Di Padova 
CoChair: Charalambous, Themistoklis  Aalto Univ 

16:5017:10, Paper ThC10.1  
Consensus of Overflowing Clocks Via Repulsive Laplacian Laws 
de Carolis, Giovanni  Univ. Tor Vegata 
Galeani, Sergio  Univ. Di Roma Tor Vergata 
Sassano, Mario  Univ. of Rome, Tor Vergata 
Keywords: Concensus control and estimation, Linear systems, Hybrid systems
Abstract: The main objective of this paper consists in imposing consensus in a network of overflowing clocks with identical speeds but potentially different initial offsets. Each (overflowing) clock is modeled as a single integrator with the state confined in a bounded set such that, whenever the state reaches its maximum allowed value, it is immediately reset to zero (overflowing phenomenon), thus exhibiting both continuoustime and discretetime behaviours. In this framework, control techniques inspired by the classical Laplacian philosophy lead to a somewhat unexpected result. In fact, it is shown that both an attractive and a repulsive Laplacian law induce two periodic orbits of the closedloop system, characterized by the feature that along only one of these trajectories consensus is reached. It is then proved that the errorzeroing periodic orbit is unstable with the attractive Laplacian, hence agreement of the clocks is not achieved, while an asymptotic convergence on it is guaranteed with the repulsive Laplacian, hence consensus is reached with a repulsive control law among the clocks.


17:1017:30, Paper ThC10.2  
Broadcasting Protocols for Coordinating Nonlinear Network Systems 
Trip, Sebastian  Univ. of Groningen 
De Persis, Claudio  Univ. of Groningen 
Tesi, Pietro  Univ. of Groningen 
Keywords: Distributed cooperative control over networks, Sampled data control, Stability of hybrid systems
Abstract: We propose a new methodology to design broadcasting protocols for coordinating nonlinear network systems. Our design of the scheduling of information transmission is based on the introduction of clock variables, whose dynamics are regulated through a suitable storage function. Required clock dynamics, ensuring stability, follow then elegantly from Lyapunov like arguments. For illustrative purposes, we first consider an example of a consensus algorithm, whereafter we discuss a distributed integral controller in feedback interconnection to a network composed of output strictly incrementally passive subsystems. Finally, we show how the proposed method can be used to redesign a popular distributed controller in power grids, enabling a sampleddata implementation.


17:3017:50, Paper ThC10.3  
A Strategy to Accelerate Consensus in LeaderFollower Networks 
Parlangeli, Gianfranco  Univ. Del Salento 
Valcher, Maria Elena  Univ. Di Padova 
Keywords: Concensus control and estimation, Distributed cooperative control over networks, Cooperative autonomous systems
Abstract: In this paper we present a new strategy to accelerate convergence in consensus networks. We consider a multiagent system with a fixed communication structure: each agent updates its state value based on the states of the neighbouring agents, with the common goal of achieving consensus. In this setup, we select p leaders among the agents whose task is to accelerate the convergence speed, by exchanging information among each other and by elaborating an additional control signal, based only on the state evolution of the p leaders.We first provide a complete description of the system, then investigate under what conditions we can freely allocate the overall system dynamics, and finally address the main objective: accelerating the consensus speed.


17:5018:10, Paper ThC10.4  
When to Stop Iterating in Digraphs of Unknown Size? an Application to FiniteTime Average Consensus 
Charalambous, Themistoklis  Aalto Univ 
Hadjicostis, Christoforos  Univ. of Cyprus 
Keywords: Concensus control and estimation, Cooperative control, Cooperative autonomous systems
Abstract: In multiagent systems, existing distributed algorithms for finitetime average consensus allow the agents to calculate the exact average in finite time, but typically require the agents to continue the iterative process indefinitely. The problem is that it is impossible for one agent to be certain that all other agents have also computed the average (at least not without a priori bounds on the network size or diameter). In this paper, we enhance an existing finitetime distributed algorithm with a distributed termination mechanism that allows the nodes to agree on terminating their iterations when they have all computed the exact average. This is accomplished by exploiting the fact that the distributed algorithm allows each node to compute, in a finite number of steps, an upper bound of its eccentricity in the network. The proposed distributed termination mechanism also facilitates the computation, in a finite number of steps, of an upper bound on the diameter of the network, which can be used by several algorithms that require such global information. Illustrative examples demonstrate the validity and performance of the proposed algorithm.


18:1018:30, Paper ThC10.5  
On the InputOutput Approach towards Distributed Estimation 
Saadatniaki, Fakhteh  Tufts Univ 
Korniienko, Anton  Ec. Centrale De Lyon, Lab. Ampère 
Scorletti, Gerard  Ec. Centrale De Lyon 
KHAN, Usman A.  Tufts Univ 
Keywords: Distributed control, Concensus control and estimation, Agents and autonomous systems
Abstract: In this paper, we study distributed estimation of continuoustime, linear timeinvariant systems monitored by a network of agents communicating over a graph. We assume that no agent may possess enough measurements in its neighborhood to estimate the entire state vector on its own. In this context, we provide a networked Kalmantype estimator that combines prediction and innovation with information fusion among the neighboring agents and consider an approach based on designing static estimator gains. The main contribution of this paper is to analyze the estimation error using the notions of dissipativity and the inputoutput approach, which enable us to formulate stability and performance arguments as quasiconvex optimization problems involving linear matrix inequalities. We show that the resulting estimation error is stable and further ensures a given level of performance regarding noise rejection. Simulations illustrate the concepts described in this paper.


18:3018:50, Paper ThC10.6  
An Internal Model Based DiscreteTime Dynamic Average Consensus Estimator 
Lee, Juwon  Kwangwoon Univ 
Ryu, Kunhee  Kwangwoon Univ 
Back, Juhoon  Kwangwoon Univ 
Keywords: Concensus control and estimation, Distributed control, Decentralized control
Abstract: The dynamic average consensus problem for a group of agents is considered. Each agent is supposed to estimate the average of inputs applied to all agents repectively in a distributed manner. A new structure for the distributed average estimator which can embed the internal model of inputs is proposed. Constructive design procedures are also given for averaging of both constant inputs and timevarying inputs. The proposed estimator is validated via numerical simulations.

 