| |
Last updated on July 6, 2021. This conference program is tentative and subject to change
Technical Program for Thursday July 1, 2021
|
ThSP1 |
|
Semi-Plenary 3: Danica Kragic Jensfelt |
Semi-Plenary Session |
Chair: Keviczky, Tamas | Delft University of Technology |
|
09:30-10:30, Paper ThSP1.1 | |
Representation Learning for Interaction Tasks |
|
Kragic, Danica | KTH |
Keywords: Machine learning
Abstract: All day long, our fingers touch, grasp and move objects in various media such as air, water, oil. We do this almost effortlessly - it feels like we do not spend time planning and reflecting over what our hands and fingers do or how the continuous integration of various sensory modalities such as vision, touch, proprioception, hearing help us to outperform any other biological system in the variety of the interaction tasks that we can execute. Largely overlooked, and perhaps most fascinating is the ease with which we perform these interactions resulting in a belief that these are also easy to accomplish in artificial systems such as robots. When interacting with objects, the robot needs to consider geometric, topological, and physical properties of objects. This can be done either explicitly, by modeling and representing these properties, or implicitly, by learning them from data. The main scientific objective of this project is to create new informative and compact representations of deformable objects that incorporate both analytical and learning-based approaches and encode geometric, topological, and physical information about the robot, the object, and the environment. We will do this in the context of challenging multimodal, bimanual object interaction tasks. The focus in our work is on physical interaction with deformable objects using multimodal feedback, generative models and address stability in contact rich tasks.
|
|
ThSP2 |
|
Semi-Plenary 4: Jan Willem Van Wingerden |
Semi-Plenary Session |
Chair: Parisini, Thomas | Imperial College & Univ. of Trieste |
|
09:30-10:30, Paper ThSP2.1 | |
Closed-Loop Dynamic Wind Farm Control |
|
van Wingerden, Jan-Willem | Delft University of Technology |
Keywords: Energy systems
Abstract: Wind energy is expected to be the largest European source of energy by 2030 and therefore largely responsible for enabling Europe to achieve its goal of having at least 27% of its electrical energy generated by renewable sources. In existing and new wind farms the wind turbines still operate on an individual level, therefore each wind turbine optimizes its own power production, resulting in sub-optimal power output at wind farm level. Previously, many researchers showed that yaw control, which is an already existing control degree of freedom, can help minimize the interaction between different turbines for existing wind farms under quasi-steady conditions. However, for realistic inflow conditions a challenging time-varying control problem has to be solved. For this emerging field a dynamic control solution is still lacking. In this presentation, I will present recent results and challenges for the field of dynamic wind farm control.
|
|
ThA1 |
|
Powertrain Optimization |
Invited Session |
Chair: Salazar, Mauro | Eindhoven University of Technology |
Co-Chair: hofman, theo | TU/e |
Organizer: Salazar, Mauro | Eindhoven University of Technology |
Organizer: Willems, Frank | Eindhoven University of Technology |
Organizer: hofman, theo | TU/e |
|
11:00-11:20, Paper ThA1.1 | |
Design of a Hierarchical Energy Management Strategy for a Range-Extender Electric Delivery Truck (I) |
|
Villani, Manfredi | The Ohio State University |
Shiledar, Ankur | The Ohio State University |
Ahmed, Qadeer | The Ohio State University |
Rizzoni, Giorgio | Ohio State University |
Keywords: Optimization, Hybrid systems, Automotive
Abstract: In this work we propose an on-line implementable hierarchical strategy merging heuristic and locally optimal features for the energy management of a range-extender electric vehicle. This class of vehicles can operate as pure electric as well as hybrid, with a range-extender internal combustion engine working as a generator. Standard energy management strategies, such as Charge Depleting/Charge Sustaining and Equivalent Consumption Minimization Strategy (ECMS), show several limitations when applied to range-extender electric powertrains. These limitations can include excessive engine start/stops, potential cold starts, distance from optimal fuel consumption, inefficient usage of the grid energy. Therefore, in this work the energy management problem has been split over two levels: at the higher level, a heuristic rule-based controller is used to decide when to operate the vehicle as pure electric; at the lower level, the ECMS is implemented to locally manage the powertrain energy flows whenever the vehicle is in hybrid mode. The lower level controller is implemented in three different variants in the effort of finding a trade-off between fuel consumption, emissions, drivability and efficient battery recharge. The performance of the hierarchical energy management strategy is numerically evaluated for a range-extender Class 6 delivery truck and compared to pure ECMS. The results show the ability of the proposed approach to ensure near-optimal fuel economy with an acceptable number of start/stops of the range-extender.
|
|
11:20-11:40, Paper ThA1.2 | |
A GPU Implementation of a Look-Ahead Optimal Controller for Eco-Driving Based on Dynamic Programming (I) |
|
Zhu, Zhaoxuan | Ohio State University |
Gupta, Shobhit | The Ohio State University |
Pivaro, Nicola | The Ohio State University |
Rajakumar Deshpande, Shreshta | The Ohio State University |
Canova, Marcello | Ohio State University |
Keywords: Automotive, Optimal control, Computational methods
Abstract: Predictive energy management of Connected and Automated Vehicles (CAVs), in particular those with multiple power sources, has the potential to significantly improve energy savings in real-world driving conditions. In particular, the eco-driving problem seeks to design optimal speed and power usage profiles based upon available information from connectivity and advanced mapping features to minimize the fuel consumption between two designated locations. In this work, the eco-driving problem is formulated as a three-state receding horizon optimal control problem and solved via Dynamic Programming (DP). The optimal solution, in terms of vehicle speed and battery State of Charge (SoC) trajectories, allows a connected and automated hybrid electric vehicle to intelligently pass the signalized intersections and minimize fuel consumption over a prescribed route. To enable real-time implementation, a parallel architecture of DP is proposed for an NVIDIA GPU with CUDA programming. Simulation results indicate that the proposed optimal controller delivers more than 15% fuel economy benefits compared to a baseline control strategy, and that the solver time can be reduced by more than 90% by the parallel implementation when compared to a serial implementation.
|
|
11:40-12:00, Paper ThA1.3 | |
Time-Optimal Control of Electric Race Cars under Thermal Constraints (I) |
|
Locatello, Alessandro | Eindhoven University of Technology |
Konda, Mouleeswar | Eindhoven University of Technology |
Borsboom, Olaf Johan Theo | Eindhoven University of Technology |
hofman, theo | TU/e |
Salazar, Mauro | Eindhoven University of Technology |
Keywords: Optimal control, Automotive, Electrical machine control
Abstract: This paper presents a quasi-convex optimization framework to compute the minimum-lap-time control strategies of electric race cars, accurately accounting for the thermal limitations of the electric motor (EM). To this end, we leverage a previously developed thermally-unconstrained framework and extend it as follows: First, we identify a thermal network model of an interior permanent magnet EM comprising its shaft, rotor, magnets, stator, windings and end-windings, including their individual loss-models. Second, we devise a convex battery model capturing the impact of the state of energy on the battery losses. Third, in order to cope with the nonlinearities stemming from the transcription of the problem from time-domain to a position-dependent representation, we leverage an iterative algorithm based on second-order conic programming to efficiently compute the solution. Finally, we showcase our framework on the Le Mans racetrack. A comparison with high-fidelity simulations in Motor-CAD reveals that our proposed model can accurately capture the temperature dynamics of the EM, revealing the end-windings and the magnets to be the limiting components in a cold-start and a long-run operation scenario, respectively. Furthermore, our numerical results underline the considerable impact of the EM thermal dynamics on lap time, while suggesting that using a continuously variable transmission could significantly improve lap time with respect to a fixed-gear transmission.
|
|
12:00-12:20, Paper ThA1.4 | |
Time-Optimal Energy Management of the Formula 1 Power Unit with Active Battery Path Constraints (I) |
|
Duhr, Pol | ETH Zürich |
Schaller, Maximilian | ETH Zürich |
Arzilli, Luca | Ferrari SpA |
Cerofolini, Alberto | Ferrari SpA |
Onder, Christopher | ETH Zürich |
Keywords: Automotive, Optimal control, Optimization
Abstract: Modern Formula 1 racing cars are high-performance hybrid-electric vehicles whose battery acts as an energy storage. When the powertrain is operated close to the lower or upper state-of-charge bound of the battery, its finite size limits the electric boosting and recuperation capacity, respectively. Given the detrimental effect on the achievable lap time, such scenarios call for a careful optimization of the energy management strategies. Based on a convex model of the car's powertrain, we first study the impact of battery path constraints on the optimal control policy analytically, using a non-smooth version of Pontryagin's minimum principle. We then corroborate the derivations with the numerical solution obtained from a convex optimization framework and discuss the time-optimal energy management strategy when the lower bound on the battery state-of-charge is active. Finally, we leverage the non-causal results to improve an existing online controller in the case of an overtake maneuver. Our simulations yield a lap time gain of about 370 ms over three laps.
|
|
12:20-12:40, Paper ThA1.5 | |
Black-Box Model-Based Active Damping of Driveline Oscillations (I) |
|
Corno, Matteo | Politecnico Di Milano |
Dattilo, Stefano | Politecnico Di Milano |
Savaresi, Sergio M. | Politecnico Di Milano |
Keywords: Automotive, Mechatronics, Identification for control
Abstract: This paper presents a driveline active damping strategy for electric vehicles. We first propose a black-box identification of the main oscillation modes of a transmission of a 4 Wheel Driven electric vehicle. We identify the models in a variety of conditions. We then design the active damping control using H infinity considerations on the identified models. Extensive experimental validation shows that the active damping reduces of 24% the longitudinal jerk during sharp acceleration maneuvers without negatively affecting the longitudinal acceleration. Furthermore, we test the controller under a variety of condition to assess its robustness with respect to transmission load variations, friction changes and velocity.
|
|
12:40-13:00, Paper ThA1.6 | |
Integrated Design of a CVT-Equipped Electric Powertrain Via Analytical Target Cascading (I) |
|
Fahdzyana, Chyannie Amarillio | Eindhoven University of Technology |
Salazar, Mauro | Eindhoven University of Technology |
Donkers, M.C.F. (Tijs) | Technische Universiteit Eindhoven |
hofman, theo | TU/e |
Keywords: Optimization, Automotive, Optimization algorithms
Abstract: Electric vehicles are gaining momentum as a valid alternative to conventional engine-based cars. In order to meet the high expectation of the market, they must strive for a similar, if not better, performance and driving range. To this end, their powertrain must be carefully designed and account for the interconnections among the various components in an integrated fashion. In this paper, we present a co-design framework for electric powertrains, whereby we jointly optimize the size of the electric machine (EM) and the geometry of a continuously variable transmission (CVT) together with its ratio trajectory, with the goal of minimizing the energy consumption of the vehicle. Specifically, we first frame the minimum-energy co-design problem in an integrated manner, accounting for the CVT geometry and dynamics, and the EM size. Given the problem complexity, we decompose it into an EM-design and a CVT-design subproblem, whereby we jointly optimize the CVT-ratio trajectory, and leverage analytical target cascading (ATC) to effectively solve the design problem. Finally, we showcase our framework on the New European Driving Cycle (NEDC), highlighting the importance of designing powertrains in an integrated fashion: Compared to the case whereby only the EM, the CVT, or the control are optimized, our joint EM-CVT design can improve the energy consumption of the vehicle by up to 22%.
|
|
ThA2 |
|
Security & Privacy of Cyber-Physical Systems I |
Invited Session |
Chair: Murguia, Carlos | Eindhoven University of Technology |
Co-Chair: Farokhi, Farhad | The University of Melbourne |
Organizer: Murguia, Carlos | Eindhoven University of Technology |
Organizer: Ruths, Justin | University of Texas at Dallas |
Organizer: Farokhi, Farhad | The University of Melbourne |
Organizer: Shames, Iman | University of Melbourne |
|
11:00-11:20, Paper ThA2.1 | |
The Effect of Behavioral Probability Weighting in a Simultaneous Multi-Target Attacker-Defender Game (I) |
|
Abdallah, Mustafa | Purdue University |
Cason, Timothy | Purdue University |
Bagchi, Saurabh | Purdue University |
Sundaram, Shreyas | Purdue University |
Keywords: Game theoretical methods, Optimization, Behavioural systems
Abstract: We consider a security game in a setting consisting of two players (an attacker and a defender), each with a given budget to allocate towards attack and defense, respectively, of a set of nodes. Each node has a certain value to the attacker and the defender, along with a probability of being successfully compromised, which is a function of the investments in that node by both players. For such games, we characterize the optimal investment strategies by the players at the (unique) Nash Equilibrium. We then investigate the impacts of behavioral probability weighting on the investment strategies; such probability weighting, where humans overweight low probabilities and underweight high probabilities, has been identified by behavioral economists to be a common feature of human decision-making. We show via numerical experiments that behavioral decision-making by the defender causes the Nash Equilibrium investments in each node to change (where the defender overinvests in the high-value nodes and underinvests in the low-value nodes).
|
|
11:20-11:40, Paper ThA2.2 | |
Towards Privacy Preserving Consensus Control in Multi-Agent Cyber-Physical Systems Subject to Cyber Attacks (I) |
|
Taheri, Mahdi | Concordia University |
Khorasani, Khashayar | Concordia University |
Shames, Iman | University of Melbourne |
Meskin, Nader | Qatar University |
Keywords: Agents and autonomous systems, Safety critical systems, Cooperative autonomous systems
Abstract: Multi-agent systems (MAS) require sharing their information with their neighboring agents to reach a consensus in a distributed manner. In this paper, a transformation-based consensus control methodology is developed and implemented that can be utilized to reach a consensus among agents in a distributed manner without revealing their true information to their neighboring agents. The proposed methodology protects agents privacy against eavesdropper adversaries and malicious agents capable of intercepting and accessing the agents’ exchanged data. A unique isometric isomorphism is employed for each agent to map the true value of exchanged sensor measurements and state estimates. By leveraging the property of isometric isomorphism in preserving norms, it is shown that reaching a consensus among agents is equivalent to that can be accomplished by the transformed agents dynamics. Numerical case studies are provided to illustrate the effectiveness of the proposed and developed methodology in preserving agents’ privacy while ensuring that the MAS accomplish consensus requirements despite the presence of cyber attacks.
|
|
11:40-12:00, Paper ThA2.3 | |
On Resilient Design of Cooperative Systems in Presence of Cyber-Attacks |
|
Sadabadi, Mahdieh S. | University of Sheffield |
Gusrialdi, Azwirman | Tampere University |
Keywords: Cooperative control, Concensus control and estimation, Distributed cooperative control over networks
Abstract: This paper develops a resilient cooperative control system for leader-follower consensus problems subject to false data injection attacks. The attackers are assumed to inject unknown bounded exogenous signals to the actuators of followers and/or the communication networks of the leader and follower states. In order to attenuate the effects of such attacks on the consensus and stability of the system, we develop a cooperative control system augmented with a virtual network and interconnected with a leader and followers so that the leader-follower consensus is guaranteed under unknown attacks. A Lyapunov-based design framework is proposed to guarantee stability and leader-follower consensus against attacks. The effectiveness of the theoretical results is evaluated through a simulation example.
|
|
12:00-12:20, Paper ThA2.4 | |
Encrypted Dynamic Control with Unlimited Operating Time Via FIR Filters (I) |
|
Schlüter, Nils | TU Dortmund University |
Neuhaus, Matthias | TU Dortmund University |
Schulze Darup, Moritz | University of Paderborn |
Keywords: Emerging control theory, Control over networks, Linear systems
Abstract: Encrypted control enables confidential controller evaluations in cloud-based or networked control systems. From a technical point of view, an encrypted controller is a modified control algorithm that is capable of computing encrypted control actions based on encrypted system outputs. Unsurprisingly, encrypted implementations of controllers using, e.g., homomorphic cryptosystems entail new design challenges. For instance, in order to avoid overflow or high computational loads, only a finite number of operations should be carried out on encrypted data. Clearly, this guideline is hard to satisfy for dynamic controllers due to their recursive nature. To enable an unlimited operating time, existing implementations thus rely on external ``refreshments'' of the controller state, internal refreshments using bootstrapping, or recurring controller resets. We show in this paper that simple FIR filter-based controllers allow to overcome many drawbacks of the existing approaches. In fact, since FIR filters consider only a finite amount of the most recent input data, the recursion issue is immediately solved and controller refreshments or resets are no longer required. Moreover, well-designed FIR filters are often less complex than and equally effective as IIR controllers.
|
|
12:20-12:40, Paper ThA2.5 | |
Linear System Security: Detection and Correction of Adversarial Attacks on a Quanser AERO System (I) |
|
Liu, Vincent | The University of Melbourne |
Kuijper, Margreta | University of Melbourne |
Keywords: Fault detection and identification, Fault estimation, Linear systems
Abstract: In this paper we consider an experimental setup involving a Quanser AERO system. We construct a model that explicitly accounts for Coulomb friction, which differs from the models used in the literature. We then unleash attacks on the system and implement and demonstrate attack detection and correction strategies from recent theoretical work. We show how the setup provides an experimental demonstration of these strategies and address several practicalities of bridging the gap between theory and practice.
|
|
12:40-13:00, Paper ThA2.6 | |
Fundamental Stealthiness-Distortion Tradeoffs in Dynamical Systems under Injection Attacks: A Power Spectral Analysis (I) |
|
Fang, Song | New York University |
Zhu, Quanyan | New York University |
Keywords: Stochastic control, Stochastic systems
Abstract: In this paper, we analyze the fundamental stealthiness-distortion tradeoffs of linear Gaussian dynamical systems under data injection attacks using a power spectral analysis, whereas the Kullback--Leibler (KL) divergence is employed as the stealthiness measure. Particularly, we obtain explicit formulas in terms of power spectra that characterize analytically the stealthiness-distortion tradeoffs as well as the properties of the worst-case attacks. Furthermore, it is seen in general that the attacker only needs to know the input-output behaviors of the systems in order to carry out the worst-case attacks.
|
|
ThA3 |
|
Nonlinear Systems I |
Regular Session |
Chair: Chaffey, Thomas L. | University of Cambridge |
Co-Chair: Das, Hemjyoti | University of Twente |
|
11:00-11:20, Paper ThA3.1 | |
A Finite Test for the Linearizability of Two-Input Systems by a Two-Dimensional Endogenous Dynamic Feedback |
|
Gstöttner, Conrad | Johannes Kepler University Linz |
Kolar, Bernd | Johannes Kepler University Linz |
Schöberl, Markus | Johannes Kepler University Linz |
|
|
11:20-11:40, Paper ThA3.2 | |
Incremental Nonlinear Dynamic Inversion Control of Long-Stroke Pneumatic Actuators |
|
Das, Hemjyoti | University of Twente |
Pool, Daan Marinus | Delft University of Technology |
van kampen, Erik-Jan | Delft University of Technology |
Keywords: Nonlinear system theory, Emerging control theory, Feedback linearization
Abstract: Pneumatic cylinders provide an environment friendly actuation means by minimizing the leakage of any harmful industrial fluids, as occurs for hydraulic actuators. However, pneumatic actuation has not been utilized widely for industrial servo applications due to its highly nonlinear nature. Incremental nonlinear dynamic inversion (INDI) is a form of nonlinear dynamic inversion (NDI) that relies less on plant model information, and is thus inherently robust to mismatches in the known plant-model, and also to external disturbances. Developing an incremental nonlinear controller for a pneumatic system is the main focus of this research article, which is accomplished by utilizing a cascaded-control approach, where the inner-loop INDI tracks a given force and the outer-loop NDI is for controlling the piston-position. Moreover, realistic sensor noises have been added in the simulation and the robustness of the incremental approach is demonstrated with respect to a baseline PID controller.
|
|
11:40-12:00, Paper ThA3.3 | |
Refining Dichotomy Convergence in Vector-Field Guided Path Following Control |
|
Yao, Weijia | University of Groningen |
Lin, Bohuan | University of Groningen |
Anderson, Brian D.O. | Australian National University/NICTA |
Cao, Ming | University of Groningen |
Keywords: Nonlinear system theory, Robotics, Mechatronics
Abstract: In the vector-field guided path-following problem, the desired path is described by the zero-level set of a sufficiently smooth real-valued function and to follow this path, a (guiding) vector field is designed, which is not the gradient of any potential function. The value of the aforementioned real-valued function at any point in the ambient space is called the level value at this point. Under some broad conditions, a dichotomy convergence property has been proved in the literature: the integral curves of the vector field converge either to the desired path or the singular set, where the vector field attains a zero vector. In this paper, the property is further developed in two respects. We first show that the vanishing of the level value does not necessarily imply the convergence of a trajectory to the zero-level set, while additional conditions or assumptions identified in the paper are needed to make this implication hold. The second contribution is to show that under the condition of real-analyticity of the function whose zero-level set defines the desired path, the convergence to the singular set (assuming it is compact) implies the convergence to a single point of the set, dependent on the initial condition, i.e. limit cycles are precluded. These results, although obtained in the context of the vector-field guided path-following problem, are widely applicable in many control problems, where the desired sets to converge to (particularly, a singleton constituting a desired equilibrium point) form a zero-level set of a Lyapunov(-like) function, and the system is not necessarily a gradient system.
|
|
12:00-12:20, Paper ThA3.4 | |
Monotone RLC Circuits |
|
Chaffey, Thomas L. | University of Cambridge |
Sepulchre, Rodolphe J. | University of Cambridge |
Keywords: Nonlinear system theory, Optimization algorithms, Modeling
Abstract: The circuit-theoretic origins of maximal monotonicity are revisited using modern optimization algorithms for maximal monotone operators. We present an algorithm for computing the periodic behavior of an interconnection of maximal monotone systems using a fixed point iteration. The fixed point iteration may be split according to the interconnection structure of the system. In this preliminary work, the approach is demonstrated on port interconnections of maximal monotone resistors and LTI capacitors and inductors.
|
|
12:20-12:40, Paper ThA3.5 | |
A Tunable Mixed Feedback Oscillator |
|
Che, Weiming | University of Cambridge |
Forni, Fulvio | University of Cambridge |
Keywords: Nonlinear system theory, Emerging control theory, Servo control
Abstract: The interplay of positive and negative feedback loops on different time scales appears to be a fundamental mechanisms for robust and tunable oscillations in both biological systems and electro-mechanical systems. We develop a detailed analysis of a basic three dimensional Lure model to show how controlled oscillations arise from the tuning of positive and negative feedback strengths. Our analysis is based on dominance theory and confirms, from a system-theoretic perspective, that the mixed feedback is a fundamental enabler of robust oscillations. Our results are not limited to three dimensional systems and extend to larger systems via passivity theory, and to uncertain systems via small gain arguments.
|
|
12:40-13:00, Paper ThA3.6 | |
Convergence Conditions for Persidskii Systems |
|
Mei, Wenjie | Inria |
Efimov, Denis | Inria |
Ushirobira, Rosane | Inria |
Aleksandrov, Alexander | Saint-Petersburg State University |
Keywords: Nonlinear system theory
Abstract: A class of generalized Persidskii systems is considered in this work. The conditions of convergence for Persidskii systems are introduced, which can be checked through linear matrix inequalities. Also, the case of almost periodic convergence of this class of dynamics with almost periodic input is studied. The proposed results are applied to a Lotka-Volterra model.
|
|
ThA4 |
|
Learning-Based Control III |
Regular Session |
Chair: Stursberg, Olaf | University of Kassel |
Co-Chair: Sulikowski, Bartlomiej | University of Zielona Gora, Inst. Control and ComputationEng |
|
11:00-11:20, Paper ThA4.1 | |
Online Genetic-Algorithm-Based Model Predictive Control Framework for Multi-Zone Buildings |
|
Mtibaa, Fatma | École De Technologie Supérieure ÉTS |
Nguyen, Kim-Khoa | University |
Dermardiros, Vasken | Company |
Cheriet, Mohamed | École De Technologie Supérieure |
Keywords: Iterative learning control, Optimal control, Neural networks
Abstract: To meet the strong need to reduce energy consumption and environmental impact, a more advanced control system is required for today's buildings. Model predictive control (MPC) is well used to reduce consumption and discomfort. However, to build an online MPC model in real-word building, the dynamics of the physical system must be accurately modeled, which is a time-consuming and costly task. Neural network models help to overcome the modeling problems especially with the availability of historical data. This research presents an implementation of a novel online data-driven control framework named Model Predictive Control via Genetic algorithm (MPC-GA) allowing the optimal operation of the heating, ventilation, and air conditioning system and has been experimentally validated on an existing multi-zone retail building. The MPC-GA combines a neural network time series multivariate prediction model, inspired from a prior work, with a model predictive control framework. A heuristic search algorithm using a genetic algorithm is used to solve the online data-driven MPC models and obtain the optimal combination settings of all controls for all the zones over a prediction horizon. Three costs are minimized including: energy consumption, peak demand and discomfort during occupied hours under self-tuned setpoint, temperature ramp and equipment cycling constraints. The benchmark results showed that the MPC-GA outperforms state-of-the-art control systems with more than 50% and 80% reduction in energy consumption and discomfort respectively in real time deployment mode.
|
|
11:20-11:40, Paper ThA4.2 | |
Polytopic Input Constraints in Learning-Based Optimal Control Using Neural Networks |
|
Markolf, Lukas | University of Kassel |
Stursberg, Olaf | University of Kassel |
Keywords: Neural networks, Constrained control, Optimal control
Abstract: This work considers artificial feed-forward neural networks as parametric approximators in optimal control of discrete-time systems. Two different approaches are introduced to take polytopic input constraints into account. The first approach determines (sub-)optimal inputs by the application of gradient methods. Closed-form expressions for the gradient of general neural networks with respect to their inputs are derived. The approach allows to consider state-dependent input constraints, as well as to ensure the satisfaction of state constraints by exploiting recursive reachable set computations. The second approach makes use of neural networks with softmax output units to map states into parameters, which determine (sub-)optimal inputs by a convex combination of the vertices of the input constraint set. The application of both approaches in model predictive control is discussed, and results obtained for a numerical example are used for illustration.
|
|
11:40-12:00, Paper ThA4.3 | |
Hardware Implementation of Low-Complexity Deep Learning-Based Model Predictive Controller |
|
Mohanty, Nirlipta Ranjan | College of Engineering Pune |
Adhau, Saket | NTNU |
Ingole, Deepak | KU Leuven |
Sonawane, Dayaram Nimba | College of Engineering Pune |
Keywords: Iterative learning control, Neural networks, Predictive control for linear systems
Abstract: Model predictive control (MPC) is an advanced control strategy that predicts the future behaviour of the plant while considering the system dynamics and constraints. This optimization-based control algorithm needs a huge amount of computational resources as it solves the optimization problem at each sampling time. This computational load demands powerful hardware and lightweight algorithms for implementing MPC on an embedded systems with limited computational resources. A deep neural network (DNN) is an attractive alternative for MPC as it provides a close approximation for various linear/nonlinear functions. In this paper, a feed-forward neural network (FFNN) and recurrent neural network (RNN)-based linear MPC is developed. These neural network algorithms, which are trained offline, can efficiently approximate MPC control law which can be easily executed on low-level embedded hardware. The closed-loop performance is verified on a hardware-in-the- loop (HIL) co-simulation on ARM microcontroller. The performance of the proposed DNN-MPC is demonstrated with a case study of a 2 degree-of-freedom (DOF) helicopter. Hardware result shows that the DNN-MPC is faster and consumes less memory as compared to MPC while retaining most of the performance indices.
|
|
12:00-12:20, Paper ThA4.4 | |
Robust Iterative Learning Control for Spatially Interconnected Systems Using 2D Control Theory |
|
Sulikowski, Bartlomiej | University of Zielona Gora, Inst. Control and ComputationEng |
Galkowski, Krzysztof | Univ. of Zielona Gora |
Trzciński, Daniel | University of Zielona Góra |
Rogers, Eric | Univ. of Southampton |
Kummert, Anton | University of Wuppertal |
Keywords: Iterative learning control, Robust control, Uncertain systems
Abstract: This paper develops a class of iterative learning control (ILC) laws for a subclass of uncertain spatially interconnected systems. Model uncertainties appear both in the state and the output equation of the dynamics modeled in the 2D systems setting to which existing results are not applicable. The first stage is to write the dynamics in 2D model form and then the stability theory for a distinct class of 2D systems known as repetitive processes is used to develop the ILC law in the case of differential dynamics. This results in a design algorithm that can be applied using linear matrix inequalities.
|
|
12:20-12:40, Paper ThA4.5 | |
A Neural Network Approach Applied to Multi-Agent Optimal Control |
|
Onken, Derek | Emory University |
Nurbekyan, Levon | UCLA |
Li, Xingjian | Emory University |
Wu Fung, Samy | UCLA |
Osher, Stanley | University of California, Los Angeles |
Ruthotto, Lars | Emory University |
Keywords: Neural networks, Optimal control, Variational methods
Abstract: We propose a neural network approach for solving high-dimensional optimal control problems. In particular, we focus on multi-agent control problems with obstacle and collision avoidance. These problems immediately become high-dimensional, even for moderate phase-space dimensions per agent. Our approach fuses the Pontryagin Maximum Principle and Hamilton-Jacobi-Bellman (HJB) approaches and parameterizes the value function with a neural network. Our approach yields controls in a feedback form for quick calculation and robustness to moderate disturbances to the system. We train our model using the objective function and optimality conditions of the control problem. Therefore, our training algorithm neither involves a data generation phase nor solutions from another algorithm. Our model uses empirically effective HJB penalizers for efficient training. By training on a distribution of initial states, we ensure the controls' optimality is achieved on a large portion of the state-space. Our approach is grid-free and scales efficiently to dimensions where grids become impractical or infeasible. We demonstrate our approach's effectiveness on a 150-dimensional multi-agent problem with obstacles.
|
|
12:40-13:00, Paper ThA4.6 | |
PI^{sigma} -- PI^{sigma} Continuous Iterative Learning Control for Nonlinear Systems with Arbitrary Relative Degree |
|
Cenceschi, Lorenzo | Università Di Pisa |
Angelini, Franco | University of Pisa |
Della Santina, Cosimo | TU Delft |
Bicchi, Antonio | Universita' Di Pisa |
Keywords: Iterative learning control, Robotics
Abstract: Online-Offline Iterative Learning Control provides an effective and robust solution to learn precise trajectory tracking when dealing with repetitive tasks. Yet, these algorithms were developed under the assumption that the relative degree between input and output is one. This prevents applications in many practically meaningful situations - e.g. mechanical systems control. To overcome this issue, this manuscript proposes a PI^{sigma} PI^{sigma} algorithm fusing information from the whole visible dynamics. We provide sufficient convergence conditions when the controlled system has a generic constant relative degree, and it is possibly subject to measurement delay. The controller is validated on several simulation scenarios, including learning to swing-up a soft pendulum.
|
|
ThA5 |
|
Consensus |
Regular Session |
Chair: Sergeenko, Anna | St. Petersburg State University |
Co-Chair: Rao, Sachit | International Institute of Information Technology, Bangalore |
|
11:00-11:20, Paper ThA5.1 | |
Sliding Mode-Based Consensus in the Presence of Byzantine Agents |
|
Rao, Sachit | International Institute of Information Technology, Bangalore |
Bajaj, Vaibhav | International Institute of Information Technology Bangalore |
Rao, Shrisha | International Institute of Information Technology Bangalore |
Keywords: Sliding mode control, Concensus control and estimation, Agents and autonomous systems
Abstract: A consensus protocol is designed for a connected network of cooperative (CO) and Byzantine (BZ) agents. An agent is considered CO if it plays by the rules, and BZ if it does not apply the same consensus protocol as the CO agents, sends different/misleading state values to its neighbors, or forges signatures of CO agents. Such a scenario can arise in networked multi-agent systems, for example, wireless sensor nodes, when some of the nodes become BZ as the result of an external attack. If the consensus protocol is not designed to handle BZ agents, then the CO agents may not reach a consensus. In this paper, agents with first-order continuous dynamics are considered; an extension to agents with higher-order dynamics is also presented. Results from sliding mode control (SMC) theory and the distributed computing systems literature, such as the use of signed messages that are exchanged between agents, are used to design CO agents' inputs that lead to consensus. A novel messaging algorithm is given to handle BZ agents that can forge signatures of CO agents to the messages originating from these agents. The use of SMC theory ensures that consensus is achieved by the CO agents within a finite-time interval. The protocol does not require the knowledge of the number of BZ agents or their locations in the network.
|
|
11:20-11:40, Paper ThA5.2 | |
Consensus Analysis Over Clustered Networks of Multi-Agent Systems under External Disturbances |
|
Pham, Thiem V. | University of Reims Champagne-Ardenne |
Nguyen, Thi Thanh Quynh | University of Reims Champagne Ardenne |
Keywords: Cooperative autonomous systems, Hybrid systems, Agents and autonomous systems
Abstract: This paper studies a consensus problem of multi-agent systems subjected to external disturbances over the clustered network. It considers that the agents are divided into several clusters. They are almost all the time isolated one from another, which has a directed spanning tree. The goal of agents achieves a common value. To support interaction between clusters with a minimum exchange of information, we consider that each cluster has an agent, who can exchange information to any agents outside of its cluster at some discrete instants of time. Our main contribution proposes a consensus protocol, which takes into account the continuous-time communications among agents inside the clusters and discrete-time communication information across clusters. Accordingly, the consensus and the robust mathcal{H}_{infty} consensus over the clustered network are respectively analyzed. Thanks to results from matrix theory and algebraic graph theory, we show that the proposed control protocols can solve the problems mentioned above. Finally, a numerical example is given to show the effectiveness of the proposed theoretical results.
|
|
11:40-12:00, Paper ThA5.3 | |
Synchronizing Live Video Streaming Players Via Consensus |
|
Manfredi, Gioacchino | Politecnico Di Bari |
De Cicco, Luca | Politecnico Di Bari |
Mascolo, Saverio | Politecnico Di Bari |
Keywords: Communication networks, Computer networks, Concensus control and estimation
Abstract: Video streaming is the primary source of the global Internet traffic. Social media applications allow users to experience live streaming events together, even though not being in the same physical place. The current online video delivery architecture cannot ensure a synchronized video playback among geographically distributed users. Since today's online events are often accompanied by comments on social networks or live chats, unsynchronized video playback can become evident and negatively impact the users' feelings of togetherness. In this paper, we propose a distributed control approach to tackle this issue. In particular, we show that the well-known consensus problem of simple integrators with saturated inputs is an appropriate mathematical framework to design a distributed playback synchronization mechanism. Furthermore, we propose a leader-following approach to ensure synchronization both among users and the video provider. Simulations on different network topologies confirm that the proposed approach is effective at enforcing asymptotic synchronization without having detrimental effects on the users' perceived quality.
|
|
12:00-12:20, Paper ThA5.4 | |
Optimal Control of Cluster Consensus in Networks with Equitable Partitions |
|
Di Meglio, Anna | University of Naples, Federico II |
Lo Iudice, Francesco | Università Degli Studi Di Napoli Federico II |
Della Rossa, Fabio | Politecnico Di Milano |
Sorrentino, Francesco | University of Naples Parthenope |
Keywords: Concensus control and estimation, Network analysis and control, Optimal control
Abstract: In this paper, we tackle the problem of controlling the cluster consensus state corresponding to an equitable partition in linear networks described by the pair (A,B). The matrix A describes the structure of a directed graph and the matrix B defines the nodes in which control input signals are injected. We show that the presence of equitable partitions corresponds to the existence of an A-invariant subspace, the cluster consensus subspace, along which nodes belonging to the same cluster follow the same time evolution. We show that this subspace encompasses the network controllable subspace. Under the assumptions that the dynamics along the cluster consensus subspace is controllable, we design optimal control actions to control the cluster consensus solution either when the initial condition lies on the cluster consensus subspace or when the dynamics transverse to this subspace is asymptotically stable.
|
|
12:20-12:40, Paper ThA5.5 | |
Weighted SPSA-Based Consensus Algorithm for Distributed Cooperative Target Tracking |
|
Erofeeva, Victoria | Skolkovo Institute of Science and Technology |
Granichin, Oleg | Saint Petersburg State University |
Granichina, Olga | Herzen State Pedagogical University of Russia |
Proskurnikov, Anton | Politecnico Di Torino |
Sergeenko, Anna | St. Petersburg State University |
Keywords: Randomized algorithms, Agents networks, Optimization algorithms
Abstract: In this paper, the new weighted algorithm for distributed multi-target tracking in a sensor network is proposed. The main feature of that algorithm, which is the combination of SPSA and consensus algorithm, is the ability to solve distributed optimization problems in the presence of uncertainties with not 'standard' probabilistic properties. As an example, the multi-target tracking problem is described. In this case, the role of arbitrary external noise can be played by various systematic errors (model errors) during measurements, which are often too difficult to exclude. Moreover, changes in the trajectory of a manoeuvring target also can be described as arbitrary unknown-but-bounded disturbances because usually there is not enough statistics in the variety of possible behaviours. Due to being weighted this algorithm makes it possible to solve the target tracking problem even if targets possess different behaviours. In this work, the covariance matrix of residual estimations under the presence of unknown (but bounded) noise is estimated as the performance index of the algorithm. Theoretical results are illustrated by numerical simulations.
|
|
12:40-13:00, Paper ThA5.6 | |
Distributed Finite-Time Control for Coordinated Circumnavigation with Multiple Agents |
|
Zhang, Chunyan | Nanjing University of Science and Technology |
Kong, Deren | Nanjing University of Science and Technology |
Keywords: Concensus control and estimation, Distributed cooperative control over networks, Stability of nonlinear systems
Abstract: This paper investigates the finite-time coordinated circumnavigation problem for multiple agents. The agents are required to surround a moving target with prescribed circular velocity and formation under velocity constraints. All agents can communicate to their neighbors by a digraph, while the topology of the sensor graph between agents and target may dynamically switch. We first design a distributed finite-time observer to estimate the state of the target. Then, a distributed observer-based control strategy is proposed, so that all agents can converge to their desired trajectories in finite time. Finally, numerical simulations illustrate the effectiveness of the proposed methods.
|
|
ThA6 |
|
Model Predictive Control III |
Regular Session |
Chair: Albalawi, Fahad | Taif University |
Co-Chair: Jungers, Raphaël | Université Catholique De Louvain |
|
11:00-11:20, Paper ThA6.1 | |
Bias Correction in Deterministic Policy Gradient Using Robust MPC |
|
Bahari Kordabad, Arash | Norwegian University of Science and Technology |
Nejatbakhsh Esfahani, Hossein | Norwegian University of Science and Technology |
Gros, Sebastien | NTNU |
Keywords: Predictive control for nonlinear systems, Optimization, Markov processes
Abstract: In this paper, we discuss the deterministic policy gradient using the Actor-Critic methods based on the linear compatible advantage function approximator, where the input spaces are continuous. When the policy is restricted by hard constraints, the exploration may not be Centred or Isotropic (non-CI). As a result, the policy gradient estimation can be biased. We focus on constrained policies based on Model Predictive Control (MPC) schemes and to address the bias issue, we propose an approximate Robust MPC approach accounting for the exploration. The RMPC-based policy ensures that a Centered and Isotropic (CI) exploration is approximately feasible. A posterior projection is used to ensure its exact feasibility, we formally prove that this approach does not bias the gradient estimation.
|
|
11:20-11:40, Paper ThA6.2 | |
Enhanced Flexibility of PWRs (Mode A) Using an Efficient NMPC-Based Boration/Dilution System |
|
Dupré, Guillaume | IMT Atlantique, LS2N (UMR 6004) |
Grossetête, Alain | Framatome |
Chevrel, Philippe | IMT Atlantique, LS2N (UMR 6004) |
YAGOUBI, Mohamed | IMT Atlantique (LS2N) |
Keywords: Power plants, Predictive control for nonlinear systems, Model/Controller reduction
Abstract: This paper presents a novel automatic boration-dilution system aimed at Pressurized Water Reactors which were initially designed for base load operation (most often with the so-called mode A core control system). The main objective is to enhance maneuvering capabilities of already-installed plants by ensuring proper regulation of the axial power distribution during power variations. No substantial modification of the on-site instrumentation and control equipment is required to implement this new control system. In addition, this device can be easily converted into a real-time aid predictive system to help plant operators control the axial offset of the reactor core. Unlike currently available open-loop simulators, a key feature of this control-based system is to periodically benefit from on-site measurements to counteract model discrepancy and unexpected disturbances. The proposed system relies on Nonlinear Model Predictive Control to efficiently control the axial offset of the reactor core with respect to hard actuator constraints. For this purpose, a simplified multi-point core kinetic model of the reactor was developed. This model is then reduced using singular perturbation theory to allow quick computation of the NMPC control law as well as full-state estimation at each recalculation instant. Nevertheless, the reduced model still provides a reasonable level of accuracy in regard to our control objective. Finally, various MATLAB®/Simulink® simulation scenarios were tested to tune the NMPC controller before being successfully validated on a realistic full-scale nuclear power plant simulator. Effectiveness of the control strategy is supported by excellent validation results.
|
|
11:40-12:00, Paper ThA6.3 | |
Immersion-Based Model Predictive Control of Constrained Nonlinear Systems: Polyflow Approximation |
|
Wang, Zheming | UCLouvain |
Jungers, Raphaël | Université Catholique De Louvain |
Keywords: Nonlinear system theory, Predictive control for nonlinear systems, Optimal control
Abstract: In the framework of Model Predictive Control (MPC), the control input is typically computed by solving optimization problems repeatedly online. For general nonlinear systems, the online optimization problems are non-convex and computationally expensive or even intractable. In this paper, we propose to circumvent this issue by computing a high-dimensional linear embedding of discrete-time nonlinear systems. The computation relies on an algebraic condition related to the immersibility property of nonlinear systems and can be implemented offline. With the high-dimensional linear model, we then define and solve a convex online MPC problem. We also provide an interpretation of our approach under the Koopman operator framework.
|
|
12:00-12:20, Paper ThA6.4 | |
Regret-Based Robust Economic Model Predictive Control for Nonlinear Dissipative Systems |
|
Albalawi, Fahad | Taif University |
Dong, Zihang | Imperial College London |
Angeli, David | Imperial College London |
Keywords: Robust control, Stability of nonlinear systems, Optimal control
Abstract: In this note, we introduce a regret-based economic MPC paradigm for nonlinear systems subject to unknown but bounded disturbances. The closed-loop system is optimized with respect to a robust regret function within a tube around the solution of the associated nominal system. The main motivation of the proposed work is the possible improvement of the economic performance when one considers the regret function as the objective function for the robust economic MPC algorithm instead of the worst cost. When the dissipativity of the nominal system with an appropriate supply rate is satisfied, the closed-loop system is proved to be driven to an optimal robust set-point under the proposed Economic MPC. Furthermore, under mild assumptions, we show that the closed-loop asymptotic average regret of the proposed controller is better than or equal to the regret at the robust steady-state. Finally, an illustrative example is utilized to compare the closed-loop stability and the average closed-loop performance (i.e., closed-loop regret and closed-loop economic cost) of the proposed regret based robust EMPC and the worst cost based EMPC.
|
|
12:40-13:00, Paper ThA6.6 | |
A New Tube-Based Output Feedback MPC Scheme for Constrained Linear Systems Based on the Concept of Substitute Estimates |
|
Subramanian, Sankaranarayanan | TU Dortmund |
Engell, Sebastian | TU Dortmund |
Keywords: Output feedback, Robust control, Predictive control for linear systems
Abstract: Model Predictive Control (MPC) is an advanced control strategy for the control of multi-variable and constrained dynamical systems. Tube-based MPC is a robust control strategy used to handle uncertainties that are present in the model. Since the full-state information is rarely available in practical applications, the estimation error must also be taken into account in addition to model uncertainties to achieve robustness and closed-loop stability in the presence of constraints. In this work, we propose a novel way to design tube-based output feedback MPC, which is less conservative than the existing approaches and can work with any estimation scheme for which the estimation error bound can be stated. The scheme employs a novel concept of substitute estimates that replace the actual estimates of the estimator to compute the control law for the system. We also show that the closed-loop is robustly exponentially stable under standard assumptions.
|
|
ThA7 |
|
Robotics III |
Regular Session |
Chair: Doulgeri, Zoe | Aristotle University of Thessaloniki |
Co-Chair: Baioumy, Mohamed | University of Oxford |
|
11:00-11:20, Paper ThA7.1 | |
Fault-Tolerant Control of Robotic Systems with Sensory Faults Using Decoupled Active Inference |
|
Baioumy, Mohamed | University of Oxford |
Pezzato, Corrado | Delft University of Technology |
Ferrari, Riccardo | Delft University of Technology |
Hernandez Corbato, Carlos | Delft University of Technology |
Hawes, Nick | University of Oxford |
Keywords: Robotics, Fault tolerant systems, Applications in neuroscience
Abstract: This work presents a novel fault-tolerant control scheme based on active inference. Specifically, a new formulation of active inference which, unlike previous solutions, provides unbiased state estimation and simplifies the definition of probabilistically robust thresholds for fault-tolerant control of robotic systems using the free-energy. The proposed solution makes use of the sensory prediction errors in the free-energy for the generation of residuals and thresholds for fault detection and isolation of sensory faults, and it does not require additional controllers for fault recovery. Results validating the benefits in a simulated 2-DOF manipulator are presented, and future directions to improve the current fault recovery approach are discussed.
|
|
11:20-11:40, Paper ThA7.2 | |
Exponential Stability of an Attitude Trajectory Tracking Controller Utilizing Unit Quaternions |
|
Koutras, Leonidas | Aristotle University of Thessaloniki |
Doulgeri, Zoe | Aristotle University of Thessaloniki |
Keywords: Robotics
Abstract: The problem of designing a controller for tracking a desired attitude trajectory by a rigid body is addressed. Such a controller design is not trivial since the orientation space is a Riemannian manifold. In this paper we utilize an attitude error stemming from the Lie Algebra of the unit quaternion space in conjunction with an angular velocity error respecting the manifold's geometry. Using these errors we design a controller that guarantees exponential convergence to the desired trajectory. The proposed controller with the particular attitude error is shown through simulations to achieve faster convergence to the desired trajectory than other attitude error selections.
|
|
11:40-12:00, Paper ThA7.3 | |
Magnetometer Aided GPS-Free Localization of an Autonomous Vineyard Drone |
|
Pizzocaro, Solomon | Politecnico Di Milano |
Corno, Matteo | Politecnico Di Milano |
Pantano, Matteo | Technical University of Munich |
Savaresi, Sergio M. | Politecnico Di Milano |
Keywords: Autonomous systems, Autonomous robots, Sensor and signal fusion
Abstract: Autonomous agricultural machines are attracting significant attention from the industry. Automatization helps to reduce costs, and increase productivity. This paper designs a global localization algorithm for an agricultural drone designed for autonomous operation in vineyards. Particularly, the main objective is to obtain an accurate localization also when the GPS localization fails. The scheme has to main components: an Extended Kalman Filter (EKF) and an online magnetometer calibration module. The first fuses information from a inertial sensors, track encoders and azimuth from the calibrated magnetometer. The second module adaptively estimates the magnetometer distortion parameters with a Recursive Least Square (RLS) with a forgetting factor. We validated the proposed scheme during extensive in field tests obtaining centimeter level accuracy.
|
|
12:00-12:20, Paper ThA7.4 | |
Two Key-Frame State Marginalization for Computationally Efficient Visual Inertial Navigation |
|
Thalagala, Ravindu | Memorial University of Newfoundland |
De Silva, Oscar | Memorial University of Newfoundland |
Mann, George K. I. | Memorial University of Newfoundland |
Gosine, Raymond G. | Memorial University of Newfoundland |
Keywords: Autonomous robots, Observers for nonlinear systems, Stochastic filtering
Abstract: In this paper we perform a detailed evaluation of two key-frame state marginalization for visual inertial navigation filters to show that the method is significantly more computationally efficient than generic visual inertial odometry (VIO) methods while being sufficiently accurate for micro aerial vehicle (MAV) navigation. For this purpose, we use the EuRoC MAV dataset for comparing the drift of MSCKF-Generic, MSCKF-Mono, MSCKF-Two way, and Two keyframe VIO filters. The error state formulation of the two key-frame based and multi key-frame based VIO is presented, then the drift, accuracy, and execution time of each filter is compared. The results indicate close to 90% faster execution of two key-frame based VIO algorithm on all datasets compared while having less than 3% drift in position for the total distance traversed.
|
|
12:20-12:40, Paper ThA7.5 | |
Decentralized Reactive Control of Robotic Teams for Cooperative Sweep Coverage of Corridor-Like Environments |
|
Matveev, Alexey S. | St.Petersburg University |
Konovalov, Petr | Spbu |
Keywords: Robotics, Cooperative control, Autonomous robots
Abstract: A group of non-communicating robots has to organize themselves into the densest possible barrier across the width of an unknown corridor-like scene and to ensure that this barrier moves along the corridor with a pre-specified speed. The robots do not know their total number and cannot distinguish between the partners, every of them determines the relative positions of neighboring objects only within a short range of visibility. A decentralized control law is presented that solves this mission, individually operates at any robot, is common for all of them and reactive, i.e., immediately converts observation data into control in a reflex-like fashion. Scenarios where the team size is increased or decreased, respectively, in the course of operation are also elaborated. The control law is justified via a mathematically rigorous global convergence result for corridors with straight parallel walls, and its performance is confirmed for more general corridors via computer simulation tests.
|
|
12:40-13:00, Paper ThA7.6 | |
A Modified Recursive Newton-Euler Algorithm Embedding a Collision Avoidance Module |
|
Trotti, Francesco | University of Verona |
Ghignoni, Eros | University of Verona |
Muradore, Riccardo | University of Verona |
Keywords: Robotics
Abstract: Planning collision-free trajectories for robotic manipulators is a quite old problem that has recently re-gained importance thanks to collaborative robotics. In scenarios where a robot is moving close to an operator, it is of paramount importance to detect obstacles (e.g. human' arms) that must be avoided. When the environment is dynamic this problem is still challenging. In this paper we integrate a recent collision avoidance algorithm based on the efficient computation of distances between capsules and the environment into the well-known recursive Newton-Euler algorithm for computing the inverse dynamics of serial-link manipulators. This approach allows to compute repulsive torques at the joint level that guarantee collision-free motion at run-time. The proposed approach has been validated on a simulated environment using a six degrees of freedom UR5 robot.
|
|
ThA8 |
|
Hybrid Systems |
Regular Session |
Chair: AHMED ALI, Sofiane | IRSEEM Rouen |
Co-Chair: Trenn, Stephan | University of Groningen |
|
11:00-11:20, Paper ThA8.1 | |
Gaussian Mixture Probability Hypothesis Density Filter with State-Dependent Probabilities |
|
Sun, Yi-Chieh | Purdue University |
Hwang, Inseok | Purdue University |
Keywords: Linear systems, Hybrid systems, Automotive
Abstract: The Gaussian mixture probability hypothesis density (GM-PHD) filter has been successfully applied to various multiple target tracking (MTT) applications due to its ability to estimate both the number of targets and target states from noisy measurements effectively and efficiently. However, since the GM-PHD filter assumed the target detection and survival probabilities and the birth rate to be constant, its performance could be degraded when the targets are temporarily occluded or when measurements are missing. To address this, we propose the state-dependent GM-PHD filter which explicitly considers the state-dependent probability of detection and survival as well as the state-dependent birth rate. In addition, in order to consider the different kinds of maneuvers of the target, we use a hybrid system model for a target. The performance of the proposed algorithm is illustrated with an example in a road traffic simulation, which includes target occlusion, and compared with that of the original GM-PHD filter.
|
|
11:20-11:40, Paper ThA8.2 | |
A Randomized Algorithm for the Stabilization of Switched Nonlinear Systems under Restricted Switching |
|
Kundu, Atreyee | Indian Institute of Science Bangalore |
Keywords: Switched systems, Stability of nonlinear systems, Randomized algorithms
Abstract: This paper deals with input/output-to-state stability (IOSS) of switched nonlinear systems in the discrete-time setting. We present an algorithm to construct periodic switching signals that obey pre-specified restrictions on admissible switches between the subsystems and admissible dwell times on the subsystems, and identify sufficient conditions on the individual subsystems, the admissible switches and admissible dwell times under which a switching signal obtained from our algorithm preserves stability of a switched system with overwhelming probability. We recover our earlier result on probabilistic techniques for the design of switching signals that preserve global asymptotic stability of switched linear systems under sufficient conditions on the properties of the individual subsystems and the admissible dwell times on the subsystems.
|
|
11:40-12:00, Paper ThA8.3 | |
Minimal Realization for Linear Switched Systems with a Single Switch |
|
Hossain, Md Sumon | University of Groningen |
Trenn, Stephan | University of Groningen |
Keywords: Switched systems, Linear time-varying systems, Reduced order modeling
Abstract: We discuss the problem of minimal realization for linear switched systems with a given switching signal and present some preliminary results for the single switch case. The key idea is to extend the reachable subspace of the second mode to include nonzero initial values (resulting from the first mode) and also extend the observable subspace of the first mode by taking information from the second mode into account. We provide some simple examples to illustrate the approach.
|
|
12:00-12:20, Paper ThA8.4 | |
Robust Model Predictive Control for Switched Nonlinear Dynamic Systems |
|
Ebrahim, Taher | Technical University of Dortmund |
Engell, Sebastian | TU Dortmund |
Keywords: Switched systems, Predictive control for nonlinear systems, Robust control
Abstract: This paper discusses a new approach to model predictive control (MPC) of switched nonlinear dynamic systems. Optimal control schemes that are based on relaxation followed by integrality restoration, have been proven to be computationally efficient in handling switched systems and therefore are promising candidates for use in MPC algorithms. The main disadvantage of such schemes, however, is the inability to guarantee optimality or even feasibility of the generated solution after the integrality restoration step. For solving this problem, in this paper an upper bound of the expected integer approximation error is computed and integrated as an additive disturbance into the relaxed model used in predictions. By using robust MPC schemes, e.g., multi-stage MPC, the resulting uncertain system can be handled in order to guarantee feasibility of the switched system. The development of the scheme is described and its performance is illustrated via simulation studies of a nonlinear switched system with parametric uncertainty.
|
|
12:20-12:40, Paper ThA8.5 | |
Continuous–Discrete Time High Gain Observer Design for State and Unknown Inputs Estimations of Quadrotor UAV |
|
Dam, Quang Truc | Normandy University, UNIROUEN, ESIGELEC, IRSEEM |
THABET, RIHAB EL HOUDA | NORMANDY UNIVERSITY Univ, UNIROUEN |
AHMED ALI, Sofiane | IRSEEM Rouen |
Guerin, François | Université Le Havre |
Hugo, Antoine | ESIGELEC |
Keywords: UAV's, LMI's/BMI's/SOS's, Sampled data control
Abstract: In this paper a Novel continuous-discrete (Sampled data) time High Gain Observer (NSHGO) for state and unknown-input estimation of a quadrotor is proposed. The proposed observer aims to provide states/unknown input estimation of the quadrator UAVs in the case of lacking of state measurement which is characterized by losing sensor signals between two sampling times. To achieve this task, a new structure of a continuous-discrete time observer, which combines the advantages of high gain observer with the LMI-based observer technique, is proposed. Using the extended state methodology and based on the proposed observer, an unknown input estimator is derived. To overcome the problem of discrete time measurements, a new structure of the output predictor with a correction term is introduced in the structure of the HGO. The performance of the proposed observer is illustrated through a simulation and experimental results.
|
|
12:40-13:00, Paper ThA8.6 | |
Tuning of a Class of Reset Elements Using Pseudo-Sensitivities |
|
Ahmadi Dastjerdi, Ali | Delft University Technology |
Saikumar, Niranjan | Delft University of Technology |
Hosseinnia, S. Hassan | Delft University of Technology |
Keywords: Mechatronics, Hybrid systems
Abstract: Currently, the demand for a better alternative to linear PID controllers is increasing due to the rising expectations of the high-tech industry. In literature, it has been shown that Constant in gain Lead in phase (CgLp) compensators, which are a type of reset element, have high potential to improve the performance of systems. Although there are few works which investigate tuning of these compensators, the high order harmonics and steady-state performances have not yet been considered in these methods. Recently, a frequencydomain framework has been developed to analyze closed-loop performances of reset control systems which includes high order harmonics. In this paper, this frequency-domain framework is combined with loop-shaping constraints to provide a reliable frequency-domain tuning method for CgLp compensators. Finally, different performance metrics of a CgLp compensator are compared with those of a PID controller on a precision positioning stage. The results show that the presented tuning method is effective, and the system with the CgLp compensator achieves superior dynamic performance to that of the PID controller.
|
|
ThA9 |
|
Wind Energy Systems |
Regular Session |
Chair: Yazdanpanah, M. J. | University of Tehran |
Co-Chair: van Wingerden, Jan-Willem | Delft University of Technology |
|
11:00-11:20, Paper ThA9.1 | |
Comparison of Hardware-In-The-Loop Control Methods for Wind Turbine Drive Trains on System Test Benches |
|
Kaven, Lennard | RWTH Aachen University |
Leisten, Christian | RWTH Aachen University |
Basler, Maximilian | RWTH Aachen University |
Jassmann, Uwe | Conatys GmbH |
Abel, Dirk | RWTH Aachen University |
Keywords: Energy systems, Predictive control for linear systems, Constrained control
Abstract: The current test process in the design and certification of wind turbines (WTs) is time and cost-intensive, as it depends on the wind conditions and requires setting up the WT in the field. Efforts are made to transfer the test process to a system test bench (STB) to simplify the installation and enable arbitrary and reproducible loads. The missing rotor and tower on an STB affect the drive train’s behavior and require a Hardware-in-the-Loop (HiL) operation. This contribution investigates suitable HiL control methods for applying aerodynamic torque loads and compares them in terms of performance, actuator requirements, and robustness. We survey three methods of different complexity: the inertia emulation (IE) method, the model reference control (MRC) method, and the model predictive control (MPC) method. Our comparison identifies a correlation between increased complexity of the HiL control method and enhanced performance. HiL operation reproduces the actual WT drive train’s inertia and dynamics up to 1, 6, and 9 Hz for IE, MRC, and MPC method, respectively. Yet, for HiL control methods with high performance, the dynamic torque requirements proliferate, and the controller’s robustness to model deviations decreases. Altogether, our quantitative comparison of HiL control methods facilitates selecting the appropriate method for respective testing and certification demands.
|
|
11:20-11:40, Paper ThA9.2 | |
An Optimal Reeling Control Strategy for Pumping Airborne Wind Energy Systems without Wind Speed Feedback |
|
Berra, Andrea | Politecnico Di Milano |
Fagiano, Lorenzo | Politecnico Di Milano |
Keywords: Energy systems, Adaptive control, Emerging control applications
Abstract: Pumping airborne wind energy (AWE) systems employ a kite to convert wind energy into electricity, through a cyclic reeling motion of the tether. The problem of computing the optimal reeling speed for the sake of maximizing the average cycle power is considered. The difficulty stems from two aspects: 1) the uncertain, time- (and space-) varying nature of wind speed, which can not be measured accurately, and 2) the need to consider, in the same optimization problem, the different operational phases of the power cycle. %, each requiring a different reeling control strategy. A new, model-based approach that solves this problem is proposed. In the design phase, a model of the AWE system is employed to collect data pertaining to the cycle power obtained with various reel-in/reel-out speed pairs, assuming known wind speed. Then, a nonlinear map, identified from these data, is used as cost function in an optimization program that computes the best reel-in and -out speed pairs for each wind speed. Finally, the optimization results are exploited to infer the link between optimal reeling speed and tether force, which are both measured with high accuracy. Such a link is used to design a feedback controller that computes the reeling speed based on the measured tether force, in order to converge on the optimal force-speed manifold. Simulation results with a realistic model illustrate the effectiveness of the approach.
|
|
11:40-12:00, Paper ThA9.3 | |
Blade Effective Wind Speed Estimation: A Subspace Predictive Repetitive Estimator Approach |
|
Liu, Yichao | Delft University of Technology |
Pamososuryo, Atindriyo Kusumo | Delft University of Technology |
Ferrari, Riccardo | Delft University of Technology |
Hovgaard, Tobias Gybel | Vestas Technology R&D |
van Wingerden, Jan-Willem | Delft University of Technology |
Keywords: Energy systems, Iterative learning control, Optimization algorithms
Abstract: Modern wind turbine control algorithms typically utilize rotor effective wind speed measured from an anemometer on the turbine's nacelle. Unfortunately, the measured wind speed from such a single measurement point does not give a good representation of the effective wind speed over the blades, as it does not take the varying wind condition within the entire rotor area into account. As such, Blade Effective Wind Speed (BEWS) estimation can be seen as a more accurate alternative. This paper introduces a novel Subspace Predictive Repetitive Estimator (SPRE) approach to estimate the BEWS using blade load measurements. In detail, the azimuth-dependent cone coefficient is firstly formulated to describe the mapping between the out-of-plane blade root bending moment and the wind speed over blades. Then, the SPRE scheme, which is inspired by Subspace Predictive Repetitive Control (SPRC), is proposed to estimate the BEWS. Case studies exhibit the proposed method's effectiveness at predicting BEWS and identifying wind shear in varying wind speed conditions. Moreover, this novel technique enables complicated wind inflow conditions, where a rotor is impinged and overlapped by wake shed from an upstream turbine, to be estimated.
|
|
12:00-12:20, Paper ThA9.4 | |
Turbulence-Based Load Alleviation Control for Wind Turbine in Extreme Turbulence Situation |
|
Dong, Liang | Technical University of Denmark |
Lio, Wai Hou | Technical University of Denmark |
Keywords: Energy systems, Electrical power systems, Power plants
Abstract: The extreme loads experienced by the wind turbine in the extreme wind events are critical for the evaluation of structural reliability. Hence, the load alleviation control methods need to be designed and deployed to reduce the adverse effects of extreme wind events. This work demonstrates that the extreme loads are highly correlated to wind conditions such as turbulence-induced wind shears. Based on this insight, this work proposes a turbulence-based load alleviation control strategy for adapting the controller to changes in wind condition. The estimation of the rotor averaged wind shear based on the rotor loads is illustrated, and is herein used to statistically characterize the extreme wind events for control purpose. To demonstrates the benefits, simulations are carried out using high-fidelity aero-elastic tool and the DTU 10 MW reference turbine in normal and extreme turbulence wind conditions. The results indicate that the proposed method can effectively decrease the exceedance probability of the extreme loads. Meanwhile, the method can minimize the loss of annual energy production in normal operating condition.
|
|
12:20-12:40, Paper ThA9.5 | |
Distributed Optimal Load Frequency Control with Stochastic Wind Power Generation |
|
Silani, Amirreza | University of Groningen, University of Tehran |
Cucuzzella, Michele | University of Groningen |
Scherpen, Jacquelien M.A. | Fac. Science and Engineering, University of Groningen |
Yazdanpanah, M. J. | University of Tehran |
Keywords: Electrical power systems, Nonlinear system theory, Stochastic control
Abstract: Motivated by the inadequacy of conventional control methods for power networks with a large share of renewable generation, in this paper we study the (stochastic) passivity property of wind turbines based on the Doubly Fed Induction Generator (DFIG). Differently from the majority of the results in the literature, where renewable generation is ignored or assumed to be constant, we model wind power generation as a stochastic process, where wind speed is described by a class of stochastic differential equations. Then, we design a distributed control scheme that achieves load frequency control and economic dispatch, ensuring the stochastic stability of the controlled network.
|
|
12:40-13:00, Paper ThA9.6 | |
Active Power Control of Waked Wind Farms: Compensation of Turbine Saturation and Thrust Force Balance |
|
Gonzalez Silva, Jean | Delft University of Technology |
Matthijs Doekemeijer, Bart | Delft University of Technology |
Ferrari, Riccardo | Delft University of Technology |
van Wingerden, Jan-Willem | Delft University of Technology |
Keywords: Energy systems, Supervisory control
Abstract: Active power control regulates the total power generated by wind farms with the power consumed on the electricity grid. Due to wake effects, the available power is reduced and turbulence is increased at downstream wind turbines. Such effects lead to a design challenge for wind farm control, where the delicate balance between supply and demand should be maintained, while considering the load balancing in the wind turbine structures. We propose a control architecture based on simple feedback controllers that adjusts the demanded power set points of individual wind turbines to compensate for turbine saturations and to balance thrust forces. For compensation purposes, the dynamics of power tracking in the wind turbines is approximated as a pure time-delay process, and the thrust force balance design is based on an identified linear model of the turbines. In this paper, we show that using the control architecture the generated power tracks its reference even when turbines saturate, while the thrust forces are balanced. In addition, the result shows that the proposed power dispatch strategy, which considers thrust force balance, also avoids turbine saturation, being thus beneficial for energy production. The effectiveness of the proposed feedback controller is demonstrated using high-fidelity computational fluid dynamics simulations of a small wind farm.
|
|
ThA10 |
|
Control for Agriculture |
Regular Session |
Chair: Ocampo-Martinez, Carlos | Technical University of Catalonia (UPC) |
Co-Chair: Keviczky, Tamas | Delft University of Technology |
|
11:00-11:20, Paper ThA10.1 | |
Predictive Control of Autonomous Greenhouses: A Data-Driven Approach |
|
Kerkhof, Lars | Hoogendoorn Growth Management |
Keviczky, Tamas | Delft University of Technology |
Keywords: Manufacturing processes, Predictive control for nonlinear systems, Adaptive systems
Abstract: In the past, many greenhouse control algorithms have been developed. However, the majority of these algorithms rely on an explicit parametric model description of the greenhouse. These models are often based on physical laws such as conservation of mass and energy and contain many parameters which should be identified. Due to the complex and nonlinear dynamics of greenhouses, these models might not be applicable to control greenhouses other than the ones for which these models have been designed and identified. Hence, in current horticultural practice these control algorithms are scarcely used. Therefore, the need rises for a control algorithm which does not rely on a parametric system representation but rather on input/output data of the greenhouse system, hereby establishing a way to control the system with unknown or unmodeled dynamics. A recently proposed algorithm, Data- Enabled Predictive Control (DeePC), is able to replace system identification, state estimation and future trajectory prediction by one single optimization framework. The algorithm exploits a non-parametric model constructed solely from input/output data of the system. In this work, we apply this algorithm in order to control the greenhouse climate. It is shown that in numerical simulation the DeePC algorithm is able to control the greenhouse climate while only relying on past input/output data. The algorithm is bench-marked against the Nonlinear Model Predictive (NMPC) algorithm in order to show the differences between a predictive control algorithm that has direct access to the nonlinear greenhouse simulation model and a purely data-driven predictive control algorithm. Both algorithms are compared based on reference tracking accuracy and computational time. Furthermore, it is shown in numerical simulation that the DeePC algorithm is able to cope with changing dynamics within the greenhouse system throughout the crop cycle.
|
|
11:20-11:40, Paper ThA10.2 | |
Model-Based PI Design for Irrigation Canals with Faulty Communication Networks |
|
Arauz, Teresa | University of Sevilla |
Maestre, J. M. | University of Seville |
Cetinkaya, Ahmet | National Institute of Informatics |
Camacho, Eduardo F. | University of Sevilla |
Keywords: Fault tolerant systems, Predictive control for linear systems, Constrained control
Abstract: A PI design method for faulty networks is provided based on Linear Matrix Inequalities (LMIs). Feedback controllers for irrigation canals are designed based on LMIs, but sparsity constraints are also imposed to make zero the feedback control law elements not corresponding to the tuning PI parameters. Therefore, the design method is halfway between a PI controller and an optimal feedback control law, also providing stability guarantees up to a maximum probability of packet losses. The objective of the downstream controller is to maintain the water levels upstream from each downstream check structure of each canal pool, while gravity-offtake gates satisfy downstream water demands. The proposed approach is tested using the irrigation system of ASCE Test Canal 1 and compared with other tuning methods via simulation. Our results show that the design method can be a useful tool when dealing with control systems under faulty networks.
|
|
11:40-12:00, Paper ThA10.3 | |
Mechanistic Crop Growth Model Predictive Control for Precision Irrigation in Rice |
|
Cabrera, John Audie | Electrical and Electronics Engineering Institute University of T |
Pedrasa, Jhoanna Rhodette | University of the Philippines Diliman |
Radanielson, Ando | University of Southern Queensland |
Aswani, Anil | University of California, Berkeley |
Keywords: Emerging control applications, Optimization, Identification for control
Abstract: Rapid urbanization and climate change exacerbate water scarcity. Thus, irrigation conservation is an important endeavor in food security . Conservation efforts have centered on the ability to monitor and manage the amount of water in rice fields. In this framework, there have been advances in simple rule-based irrigation scheme such as safe alternate wetting and drying (Safe-AWD). Safe-AWD provides a robust rule set that mitigates yield reduction due to the decrease in irrigation. However, the extent of the adverse effects from induced drought stress are often hard to predict across different factor combinations of crop variety, environment, and management practices. In this light, the integration of crop models offers additional precision in the amount of irrigation for specific conditions. Moreover, the availability of crop models allows the use of methods in control theory for irrigation management. In particular, Model Predictive Control (MPC) is robust and able to handle multi-objective and practical real world constraints. Importantly, it has been studied for irrigation set point tracking in water balance based models (WB-MPC). Thus to address variable crop response, this work augments MPC irrigation with a crop model to track growth trajectories of biomass, leaf area index, and grain formation throughout the planting season (CG-MPC). Presented are simulations that compare irrigation management techniques - namely traditional ponding, Safe-AWD, WB-MPC, and CG-MPC. Based on water savings and yield reduction, results show that Safe-AWD has the greatest irrigation conservation while WB-MPC and CG-MPC can be designed to have comparable water usage. However, CG-MPC produced the best yield reduction and has minimal variance across different field scenario simulations. This presents design opportunities for tuning risk trade-offs between yield reduction and water savings.
|
|
12:00-12:20, Paper ThA10.4 | |
An Unknown Input Moving Horizon Estimator for Open Channel Irrigation Systems |
|
Conde, Gregory | Universidad Central |
Quijano, Nicanor | Universidad De Los Andes |
Ocampo-Martinez, Carlos | Universitat Politécnica De Catalunya (UPC) |
Keywords: Fluid flow systems, Process control, Large-scale systems
Abstract: The use of modeling and estimation strategies appears as a valuable tool to increase the efficiency of the open channel irrigation systems (OCIS). This paper is focused on exploring the feasibility, advantages, and conditions in the implementation of a moving horizon estimation (MHE) approach designed from an approximated model that contemplates mass and energy balances of the channels, which is useful to differentiate when a change of level is a conduction change effect, or when the change is due to an unknown input. The estimation strategy is evaluated via simulation using a test case reported in the literature. The results show that, with the use of the proposed estimation strategy, it is possible to reach an optimal estimation of the total amount of unknown inputs.
|
|
12:20-12:40, Paper ThA10.5 | |
Economic Model Predictive Control for Smart and Sustainable Farm Irrigation |
|
Cáceres Rodríguez, Gabriela | Universidad Loyola De Andalucía |
Millán Gata, Pablo | Universidad Loyola Andalucía |
Pereira Martin, Mario | Universidad Loyola Andalucia |
Lozano, David | IFAPA (Junta De Andalucía) |
Keywords: Predictive control for nonlinear systems, Nonlinear system identification
Abstract: The joint effects of rise of global population, climate change and water scarcity makes the shift towards an efficient and sustainable agriculture more and more ur- gent. Fortunately, recent developments in low-cost, IoT-based sensors and actuators can help us to incorporate advanced control techniques for efficient irrigation system. This paper proposes the use of an economic model predictive control at a farm scale. The controller makes use of soil moisture data sent by the sensors, price signals, operative restrictions, and accurate dynamical models of water dynamics in the soil. Its performance is demonstrated through simulations based on a real case-study, showing that it is possible to obtain significant reductions in water and energy consumption and operation costs.
|
|
12:40-13:00, Paper ThA10.6 | |
Human-In-The-Loop Approach to Agri-Environment Control in Small-Scale Greenhouses |
|
Sugai, Mayu | Keio University |
Inoue, Masaki | Keio University |
Nakamura, Hisakazu | Tokyo University of Science |
Keywords: Predictive control for linear systems, Modeling
Abstract: In this manuscript, we address the design problem of agri-environment control systems constructed and operated at low costs. The key of the cost-reduction is the human-in-the-loop approach: farmers participate in the control system and work on agri-environment in greenhouses instead of automated actuators. A controller gives a recommendation that contains multiple options of working schedule to farmers. Based on the recommendation, farmers determine their schedule with control actions such as watering and/or ventilating to control the agri-environment in the house. The proposed control system is applied to an actual experimental environment in a greenhouse, and its effectiveness is demonstrated.
|
|
ThA11 |
|
Modeling |
Regular Session |
Chair: Teodorescu, Catalin Stefan | University College London |
Co-Chair: Jayawardhana, Bayu | University of Groningen |
|
11:00-11:20, Paper ThA11.1 | |
Analysis of an Extended Model of Bio-Inspired Source Seeking |
|
Pequeno Zuro, Alejandro | Istituto Italiano Di Tecnologia |
Shaikh, Danish | University of Southern Denmark |
Rano, Inaki | University of Southern Denmark |
Keywords: Modeling, Stability of linear systems, Differential algebraic systems
Abstract: Braitenberg vehicles are widely known models of sensor-driven source seeking behavior in animals, known in biology as taxes. While the original vehicles are reactive and move according to the instantaneously perceived stimulus, experiments have found that animals also exploit temporal changes in the stimulus for source seeking. For instance, cockroaches rely on the temporal changes in the odour concentration to perform odour localisation, which has proven a highly challenging task for robots. In this paper, we present a new mathematical model, the dynamic Braitenberg vehicle, that incorporates the temporal rate of change of the stimulus in the source seeking behavior. Our theoretical analysis of the model stability shows that including a term dependent on the temporal evolution of the stimulus significantly improves the behaviour of the closed-loop system compared to its reactive counterpart. The paper illustrates the validity of the new model and the theoretical results of the stability analysis through a set of simulations. Our proposed dynamic vehicle represents a more accurate model of animal behavior and has the potential to improve the performance of mobile robots in challenging tasks like chemical source localisation.
|
|
11:20-11:40, Paper ThA11.2 | |
A Switched Electrical Model with Thermal Effects for Li-Ion Batteries |
|
Baccari, Silvio | Università Del Sannio |
Sagnelli, Salvatore | SITAEL |
Vasca, Francesco | University of Sannio |
Iannelli, Luigi | University of Sannio in Benevento |
Keywords: Modeling, Electrical power systems, Identification
Abstract: A dynamic model for Li-ion batteries based on an equivalent switched electrical circuit is proposed. The switching topology of the model allows one to discriminate the model parameters for charging, discharging and relaxation phases. Thermal effects and the reversible entropic heat are taken into account, too. The proposed model is shown to require a lower number of parameters to be identified with respect to the classical equivalent circuit battery model. An identification procedure for determining the electrical and thermal model parameters is defined and its effectiveness is verified through experimental results.
|
|
11:40-12:00, Paper ThA11.3 | |
Parameter Estimation in Type 2 Diabetes in the Presence of Unannounced Meals and Unmodelled Disturbances |
|
Al Ahdab, Mohamad | Aalborg University |
Glavind Clausen, Henrik | Aalborg University |
Knudsen, Torben | Aalborg University, Denmark |
Björk Aradóttir, Tinna | Novo Nordisk |
Schmidt, Signe | Steno Diabetes Center, Denmark and Hvidovre University Hospital |
Norgaard, Kirsten | Hvidovre University Hospital |
Leth, John | Aalborg University |
Keywords: Metabolic systems, Identification, Optimization
Abstract: A least squares strategy to estimate states and parameters for type 2 diabetes (T2D) subjects based only on continuous glucose measurements and injected insulin in the presence of unannounced meals and disturbances, e.g., physical activity and stress, is presented. The strategy is based on a simple T2D subject model and tested with clinical data in addition to simulated data generated by using jump diffusion models for meals and disturbances. Three parameters are estimated together with the states, meals, and disturbances. The estimated meal states were shown to follow the trend of the unannounced meals. The strategy can be used to obtain a model with the estimated parameters for predictive control design. In addition, the strategy can also be used to test different insulin and meal plans with the estimated disturbances and parameters. Moreover, the paper demonstrates the ability of jump diffusion models to simulate meals and disturbances.
|
|
12:00-12:20, Paper ThA11.4 | |
Model-Based Sensor Fusion and Filtering for Localization of a Semi-Autonomous Robotic Vehicle |
|
Teodorescu, Catalin Stefan | University College London |
Caplan, Irving | University College London |
Eberle, Harry | University College London |
Carlson, Tom | University College London |
Keywords: Mechatronics, Transportation systems, Robotics
Abstract: This paper refines a physically-inspired model governing the dynamic motion of a vehicle. We present a method used to perform experimental parameter calibration, and then use this model to build an observer (an extended Kalman filter). Experimental results with a robotic vehicle fitted with a prototype kit focus on recovering the truthful real-world information in the context of systematic errors (a faulty wheel encoder sensor), randomly occurring errors (a faulty ultrasonic sensor) and simplifying model assumptions (e.g. usage of two identical motors). We show that our model-based approach is able to perform reasonably well even under these extreme circumstances.
|
|
12:20-12:40, Paper ThA11.5 | |
On the One-Shot Data-Driven Verification of Dissipativity of LTI Systems with General Quadratic Supply Rate Function |
|
Rosa, Tábitha E. | University of Groningen |
Jayawardhana, Bayu | University of Groningen |
Keywords: Linear systems, LMI's/BMI's/SOS's, Identification for control
Abstract: Based on a one-shot input-output set of data from an LTI system, we present a verification method of dissipativity property based on a general quadratic supply-rate function. We show the applicability of our approach for identifying suitable general quadratic supply-rate function in two numerical examples, one regarding the estimation of mathcal{L}_2-gains and one where we verify the dissipativity of a mass-spring-damper system.
|
|
ThA12 |
|
Safety Critical Systems |
Regular Session |
Chair: Zhang, Ping | University of Kaiserslautern |
Co-Chair: Incremona, Gian Paolo | Politecnico Di Milano |
|
11:00-11:20, Paper ThA12.1 | |
A Safe Control Architecture Based on a Model Predictive Control Supervisor for Autonomous Driving |
|
Nezami, Maryam | University of Lübeck |
Männel, Georg | Universität Zu Lübeck |
Abbas, Hossam | Assiut University, Faculty of Engineering |
Schildbach, Georg | University of Lübeck |
Keywords: Safety critical systems, Predictive control for nonlinear systems, Automotive
Abstract: This paper presents a novel, safe control architecture (SCA) for controlling an important class of systems: safety-critical systems. Ensuring the safety of control decisions has always been a challenge in automatic control. The proposed SCA aims to address this challenge by using a Model Predictive Controller (MPC) that acts as a supervisor for the operating controller, in the sense that the MPC constantly checks the safety of the control inputs generated by the operating controller and intervenes if the control input is predicted to lead to a hazardous situation in the foreseeable future invariably. Then an appropriate backup scheme can be activated, e.g., a degraded control mechanism, the transfer of the system to a safe state, or a warning signal issued to a human supervisor. For a proof of concept, the proposed SCA is applied to an autonomous driving scenario, where it is illustrated and compared in different obstacle avoidance scenarios. A major challenge of the SCA lies in the mismatch between the MPC prediction model and the real system, for which possible remedies are explored.
|
|
11:20-11:40, Paper ThA12.2 | |
Tractable Compositions of Discrete-Time Control Barrier Functions with Application to Driving Safety Control |
|
Khajenejad, Mohammad | Arizona State University |
Cavorsi, Matthew | Harvard University |
Niu, Ruochen | Arizona State University |
Shen, Qiang | Shanghai Jiao Tiong University |
Yong, Sze Zheng | Arizona State University |
Keywords: Constrained control, Safety critical systems, Autonomous systems
Abstract: This paper introduces control barrier functions for discrete-time systems, which can be shown to be necessary and sufficient for the controlled invariance of a given set. In particular, we propose nonlinear discrete-time control barrier functions for control affine systems with an additional structure that lead to controlled invariance conditions that are affine in the control input, resulting in a tractable formulation that enables us to handle the safety optimal control problem for a broader range of applications with more complicated safety conditions than existing approaches. Further, we develop alternative mixed-integer formulations for basic and secondaryBoolean compositions of multiple control barrier functions and further provide mixed-integer constraints for piecewise control barrier functions. Finally, we apply these proposed tools to driving safety problems of lane-keeping and obstacle avoidance, which are shown to be effective in simulation.
|
|
11:40-12:00, Paper ThA12.3 | |
Safety Design of Control Systems According to International Standards |
|
Suyama, Koichi | Tokyo University of Marine Science and Technology |
Sebe, Noboru | Kyushu Inst. of Tech |
Keywords: Fault tolerant systems, Switched systems, Linear systems
Abstract: We propose safety design of control systems under the authorization by international standards, such as IEC 61508, via controllers for reducing the frequency of events that physical values in a control system deviates from the normal operating range.
|
|
12:00-12:20, Paper ThA12.4 | |
A Configuration Space Reference Generation Approach for Real-Time Collision Avoidance of Industrial Robot Manipulators |
|
Sacchi, Nikolas | University of Genova |
Sangiovanni, Bianca | University of Pavia |
Incremona, Gian Paolo | Politecnico Di Milano |
Ferrara, Antonella | University of Pavia |
Keywords: Safety critical systems, Robotics, Optimization algorithms
Abstract: In this work a novel approach for target reaching and collision avoidance in industrial robot manipulators is proposed. This method solely relies on the computation of the direct kinematics of the considered industrial manipulator in order to generate joint reference positions to perform real-time tracking in the operative space and avoidance of obstacles moving in the proximity of the robot. A comparison with a conventional model-based collision avoidance method has been carried out in a simulated industrial setting under different conditions, showing satisfactory results even in case of coarse sampling times. This makes the proposal suitable for filed real-time operations executed by industrial robots performing a task. The proposed approach has been deployed on the EPSON VT6 6-axis industrial manipulator, whose proprietary software has been interfaced with general-purpose robotic simulators in order to emulate complex and dynamic environments.
|
|
12:20-12:40, Paper ThA12.5 | |
Comparison of the Approaches for PFD Evaluation of Safety Instrumented Systems with Non-Exponential Distributions |
|
Mikhaylenko, Dina | University of Kaiserslautern |
Zhang, Ping | University of Kaiserslautern |
Keywords: Safety critical systems, Markov processes, Stochastic systems
Abstract: The probability of failure on demand (PFD) is an important performance index to evaluate the reliability of safety instrumented systems (SIS) and determine their safety integrity level. In practice, some failure processes and repair processes in the SIS have a non-exponential distribution in nature. A SIS with non-exponential distributions can be described by a semi-Markov model. This paper aims to provide an overview of current developments in the evaluation of the PFD for SIS described by a semi-Markov model. Different approaches to calculate the PFD are compared in terms of accuracy of calculation results, computational efforts, and required pre-knowledge. Finally, a numerical example illustrates the application, advantages and disadvantages of each approach.
|
|
12:40-13:00, Paper ThA12.6 | |
A Fixed-Time Stable Adaptation Law for Safety-Critical Control under Parametric Uncertainty |
|
Black, Mitchell | University of Michigan |
Arabi, Ehsan | Ford Motor Company |
Panagou, Dimitra | University of Michigan |
Keywords: Safety critical systems, Robust adaptive control, Uncertain systems
Abstract: We present a novel technique for solving the problem of safe control for a class of nonlinear, control-affine systems subject to parametric model uncertainty. Invoking Lyapunov analysis and the notion of fixed-time stability (FxTS), we introduce a parameter adaptation law which guarantees convergence of the estimates of unknown parameters in the system dynamics to their true values within a fixed-time independent of the initial error. We then synthesize this law with a robust, adaptive control barrier function (RaCBF)-based quadratic program to compute safe control inputs despite the considered uncertainty. To corroborate our results, we undertake a comparative case study on the efficacy of this result versus other recent approaches in the literature to safe control under uncertainty, and close by highlighting the value of our method in the context of an automobile overtake scenario.
|
|
ThPB3 |
|
Plenary 2: Karen Wilcox |
Plenary Session |
Chair: Scherpen, Jacquelien M.A. | Fac. Science and Engineering, University of Groningen |
|
14:00-15:00, Paper ThPB3.1 | |
A Probabilistic Graphical Model Foundation for Enabling Predictive Digital Twins at Scale |
|
Willcox, Karen Elizabeth | University of Texas at Austin |
Keywords: Computational methods
Abstract: A digital twin is a set of coupled computational models that evolves over time to persistently represent the structure, behavior, and context of an individual physical asset such as a component, system, or process. Digital twins have the potential to bring value to decision-making in a broad range of societal, natural, and engineering systems. This talk presents a probabilistic graphical model as a formal mathematical representation of a digital twin and its associated physical asset. We create an abstraction of the asset-twin system as a set of coupled dynamical systems, evolving over time through their respective state-spaces and interacting via observed data and control inputs. The abstraction is realized computationally as a dynamic decision network. Predictive capabilities are enabled by physics-based reduced-order models. We demonstrate how the approach is instantiated to create, update and deploy a structural digital twin of an unmanned aerial vehicle.
|
|
ThB1 |
|
Optimal Control in Automotive Systems |
Regular Session |
Chair: Evangelou, Simos | Imperial College London |
Co-Chair: Matusko, Jadranko | Uni Zagreb |
|
15:30-15:50, Paper ThB1.1 | |
Model Predictive Path-Following Control for General N-Trailer Systems with an Arbitrary Guidance Point |
|
Lukassek, Markus | Friedrich-Alexander-Universität Erlangen-Nürnberg |
Völz, Andreas | Friedrich-Alexander-Universität Erlangen-Nürnberg |
Szabo, Tomas | ZF Friedrichshafen AG |
Graichen, Knut | University Erlangen-Nürnberg (FAU) |
Keywords: Automotive, Predictive control for nonlinear systems, Autonomous robots
Abstract: This paper presents a method to stabilize a general vehicle-n-trailer configuration so that a specific point on the rearmost trailer follows a given reference path. The system can consist of any combination of on- and off-axle couplings. To achieve stabilization in forward and reverse driving, a model predictive control is used, which is based on a kinematic vehicle-trailer model. In order to follow the path with a specific point on the trailer, a coordinate transformation is applied to the cost function. This allows, amongst various applications, both the exact positioning of agricultural implements such as field sprayers and maneuvering double-turntable trailers with the rear edge of the trailer during alley docking. The optimization problem is solved with a real-time capable tailored gradient-based augmented Lagrangian approach. To validate the algorithm, simulations are performed against a dynamic two-track model with additive sensor noise. In one of the scenarios, an agricultural setup with one trailer in forward direction and a logistical setup with four trailers in reverse direction is demonstrated.
|
|
15:50-16:10, Paper ThB1.2 | |
Predictive Approach to Torque Vectoring Based on the Koopman Operator |
|
Svec, Marko | University of Zagreb, Faculty of Electrical Engineering and Comp |
Iles, Sandor | University of Zagreb, Faculty of Electrical Engineering and Comp |
Matusko, Jadranko | Uni Zagreb |
Keywords: Predictive control for nonlinear systems, Automotive, Identification for control
Abstract: Torque Vectoring (TV) system uses an individually controlled electric powertrain to improve the dynamic behavior and enhance the handling and stability of a vehicle. In this paper, a Model Predictive Control (MPC) algorithm with a model of the vehicle identified using the Koopman operator theory is proposed. The Koopman operator is a linear predictor for nonlinear dynamical systems based on the lifting of the nonlinear dynamics in a higher-dimensional space where its evolution is linear. Using such a model may allow for achieving similar performance to those of a nonlinear MPC with the computational efficiency of a linear MPC. The Koopman MPC was compared to a Linear Time-Variant (LTV) MPC, a common approach in the existing literature, and showed increased performance.
|
|
16:10-16:30, Paper ThB1.3 | |
Optimal Energy Management for Fuel and Emissions Minimization of Series Hybrid Electric Vehicles with Consideration of Engine Preheating |
|
Hu, Ruikuan | Imperial College London |
Pan, Xiao | Imperial College London |
Chen, Boli | Unversity College London |
Evangelou, Simos | Imperial College London |
Keywords: Automotive, Optimization, Optimal control
Abstract: As a result of multiple energy sources, hybrid electric vehicles (HEVs) provide additional flexibility of the engine operating point, which enables optimization of fuel economy and emissions reduction. This paper introduces an energy management (EM) control strategy by an optimal control approach that jointly optimizes fuel consumption and various vehicle pollutant emissions. Engine thermal dynamics are modeled and integrated into the engine-out emission and fuel consumption models for enhanced modeling accuracy. The proposed method is formulated as an optimal control problem (OCP) that is benchmarked against a baseline EM strategy for fuel consumption only optimization. The proposed temperature sensitive emission and fuel consumption models enable a thorough investigation of engine temperature-emissions relationships, which provide important insights into the optimal power split during the engine preheating phase. Simulation results validate the effectiveness of the proposed approach and highlight the importance of analyzing the fuel consumption-emissions trade-off, as small compromises in fuel consumption lead to significant reductions in emissions.
|
|
16:30-16:50, Paper ThB1.4 | |
Synchronized Ecodriving Control for Convoys |
|
Schauer, Christian | Johannes Kepler University Linz |
Obereigner, Gunda | JKU Linz |
Del Re, Luigi | Johannes Kepler University Linz |
Keywords: Optimization, Automotive, Optimal control
Abstract: The adoption of an energy aware driving style, often referred to as ecodriving, is a promising way to save, besides of costs, non-renewable fossil fuels. In a previous paper we presented a two-layer control approach for ecodriving which tracks an optimal speed profile resulting from the first layer under the presence of traffic in the second layer. In this paper, we extend the developed control approach to a convoy of vehicles and determine, whether it leads to safe and stable convoys and whether an increasing number of vehicles within the convoy leads to an increased fuel efficiency. Thereby, we focus on two different prediction methods of the preceding vehicle: in the first case we assume perfect knowledge of the future movement of the preceding vehicle, whereas in the second case we assume the current speed of the preceding vehicle to be constant during the prediction horizon. It is shown that the higher the number of vehicles the more fuel saving from the optimal solution can be recovered.
|
|
16:50-17:10, Paper ThB1.5 | |
Development of Analytical Eco-Driving Cycles for Electric Vehicles |
|
Wulf Ribelles, Luis Alfredo | Université D’Orléans |
Gillet, Kristan | University of Orléans |
Colin, Guillaume | University of Orléans |
chamaillard, yann | University of Orléans |
Simon, Antoine | Groupe PSA |
Nouillant, Cedric | PSA Peugeot Citroen |
Keywords: Automotive, Optimal control, Transportation systems
Abstract: This paper presents the development of analytical solutions for the computation of Eco-Driving cycles for electric vehicles. The task of defining an Eco-Driving strategy is formulated as an Optimal Control Problem aiming to minimize the energy consumed during a trip subject to input and speed constraints. Here, the final time of the driving mission is set as a free parameter and only the relevant terms for the optimization are taken into account. The problem is solved using Pontryagin’s Minimum Principle in a systematic way, allowing the derivation of closed-form expressions for the different (un)constrained solutions. The results obtained with the proposed approach are compared to the optimal solution given by Dynamic Programming, where a minor deviation from the optimal consumption is achieved while drastically reducing the computation time of the solution.
|
|
17:10-17:30, Paper ThB1.6 | |
Predictive Energy Management of a HEV Considering Engine Torque Dynamic |
|
Kuchly, Jean | Stellantis |
Nelson-Gruel, Dominique | University of ORLEANS |
Charlet, Alain | Université D'Orléans |
Simon, Antoine | Groupe PSA |
JAINE, Thierry | PSA Groupe |
Nouillant, Cedric | PSA Peugeot Citroen |
chamaillard, yann | University of Orléans |
Keywords: Automotive, Optimal control
Abstract: Although the Equivalent Consumption Minimization Strategy (ECMS) provides excellent performances in the context of energy management of parallel Hybrid Electric Vehicles (HEVs), it usually approximates the vehicle powertrain mainly as static maps, considering only the dynamic of the battery State Of Charge (SOC). In particular, neglecting the dynamic of the Internal Combustion Engine (ICE) is problematic and leads to unnecessary engine starts and a worse overall energy efficiency. This paper aims to address this deficiency by taking into account the energy impact of the engine torque transient and the switch from an electric mode to a hybrid or thermal mode, through an approach based on both ECMS and Model Predictive Control (MPC). The resolution of a quadratic optimization problem taking into account short-term future leads to allow or forbid the start of the ICE, depending on the energy balance between the potential gain of a hybrid mode and the cost of the engine start and torque raise. By avoiding the consumption cost of unnecessary engine starts and by handling properly the ICE torque dynamic, this mixed approach yields fuel consumption gain up to 0.6%. This gain is important in a context of very challenging environmental constraints in the automotive industry.
|
|
ThB2 |
|
Security & Privacy of Cyber-Physical Systems II |
Invited Session |
Chair: Sinopoli, Bruno | Washington University in St Louis |
Co-Chair: Shames, Iman | University of Melbourne |
Organizer: Murguia, Carlos | Eindhoven University of Technology |
Organizer: Ruths, Justin | University of Texas at Dallas |
Organizer: Farokhi, Farhad | The University of Melbourne |
Organizer: Shames, Iman | University of Melbourne |
|
15:30-15:50, Paper ThB2.1 | |
Designing Privacy Filters for Hidden Markov Processes (I) |
|
Cavarec, Baptiste | KTH Royal Institute of Technology |
Stavrou, Photios A. | KTH Royal Institute of Technology |
Skoglund, Mikael | Royal Institute of Technology |
Bengtsson, Mats | KTH Royal Institute of Technology |
Keywords: Markov processes, Optimization, Computational methods
Abstract: We address the problem of releasing a utility process correlated with a hidden sensitive source (both modeled through a hidden Markov model) by designing a privacy filter hiding the sensitive data while maintaining a fidelity criterion on the utility process. The problem is formulated as a constrained minimization of a variant of relative entropy between the sensitive hidden process and the output of the privacy filter. We first explain that in its initial form, the problem is suffering from the curse of dimensionality. Then, we propose a relaxation of it taking into account the information structure of the decoder policies. Such relaxation leads to tractable privacy filtering policies that are illustrated via a simulation study.
|
|
15:50-16:10, Paper ThB2.2 | |
Robust Resilient Sparse Voltage Control of Low-Voltage DC Microgrids |
|
Sadabadi, Mahdieh S. | University of Sheffield |
Bahavarnia, Mirsaleh | Lehigh University |
Keywords: Distributed control, LMI's/BMI's/SOS's, H2/H-infinity methods
Abstract: This paper presents a robust resilient sparse voltage control framework for low-voltage DC microgrids consisting of multiple distributed generation (DG) units with uncertain loads. To this end, a multiple-DG DC microgrid with arbitrary topology is cast as a linear time-invariant (LTI) interconnected system subject to structured uncertainty. By means of this modeling framework, a systematic convex-optimization-based approach for the design of sparse voltage control is proposed. The proposed voltage control strategy regulates the voltage of DG units at the point of common couplings and is simultaneously robust, resilient, and sparse. The effectiveness of the proposed sparse control scheme is verified through detailed simulation case studies.
|
|
16:10-16:30, Paper ThB2.3 | |
Analyzing Cyber-Resiliency of a Marine Navigation System Using Behavioral Relations |
|
Nissov, Morten Christian | Technical University of Denmark |
Dagdilelis, Dimitrios | Technical University of Denmark |
Galeazzi, Roberto | Technical University of Denmark |
Blanke, Mogens | DTU |
Keywords: Maritime, Fault diagnosis, Behavioural systems
Abstract: Marine vessels need trustworthy navigation data for safe manoeuvring, but threats exist for external manipulation of signals and on-board systems. This paper employs analysis of behaviours to cross-validate that instruments provide correct information. Deviations from normal behaviour could be effects of malicious cyber-attack or instrument malfunction. Independent of the root cause, faulty information need be disregarded for navigation. This paper shows how instruments' violation of correct behaviour can be detected and isolated during near-coast navigation. The approach is to analyse topology of information flow and information processing, also referred to as structural analysis. The paper addresses the diagnosis potential for isolation of erroneous information about state of own ship and of surrounding objects. The analysis includes position, ship speed, and heading, which could lead to errors in navigation, to collision or grounding. The paper addresses required sensors, according to the International Maritime Organizations (IMO) Safety of Life at Sea (SOLAS), and also presents potential gains by inclusion of computer vision. Showing that all single and several cases of simultaneous defects are discovered, for own ship and in surroundings, the results demonstrate that resilience of navigation information can be obtained for vessels sailing in coastal waters.
|
|
16:30-16:50, Paper ThB2.4 | |
Resilient Consensus against Epidemic Malicious Attacks |
|
Wang, Yuan | Tokyo Institute of Technology |
Ishii, Hideaki | Tokyo Institute of Technology |
Bonnet, Francois | Tokyo Institute of Technology |
Defago, Xavier | Tokyo Institute of Technology |
Keywords: Distributed parameter systems, Agents networks, Fault tolerant systems
Abstract: This paper addresses novel consensus problems for multi-agent systems operating in a pandemic environment where infectious diseases are spreading. The dynamics of the diseases follows the susceptible-infected-recovered (SIR) model, where the infection induces faulty behaviors in the agents and affects their state values. To ensure resilient consensus among the noninfectious agents, the difficulty is that the number of infectious agents changes over time. We assume that a high-level policy maker announces the level of infection in real time, which can be adopted by the agents for their preventative measures. It is demonstrated that this problem can be formulated as resilient consensus in the presence of the so-called mobile malicious models, where the mean subsequence reduced (MSR) algorithms are known to be effective. We characterize sufficient conditions on the network structures for different policies regarding the announced infection levels and the strength of the pandemic. Numerical simulations are carried out for random graphs to verify the effectiveness of our approach.
|
|
16:50-17:10, Paper ThB2.5 | |
Data-Injection Attacks Using Historical Inputs and Outputs (I) |
|
Alisic, Rijad | KTH Royal Institute of Technology |
Sandberg, Henrik | KTH Royal Institute of Technology |
Keywords: Behavioural systems, Statistical learning, Linear systems
Abstract: Data-driven, model-free control has become popular in recent years, due to their ease of implementation and minimal information requirement about the system. In this paper, we investigate whether the same methods could be used by an adversary to synthesize undetectable data-injection attacks on cyber-physical systems using Willems' Fundamental Lemma. We show that if the adversary is able to upper bound the order of a linear, time-invariant system and read all its inputs and outputs, then the adversary will be able to generate undetectable attack signals in the form of covert attacks. Additionally, we provide conditions on the disclosed data set that enable the adversary to generate zero dynamics attacks. These conditions give operators insights into when enough information about the system has been revealed for an adversary to conduct an undetectable attack. Finally, the different attack strategies are verified through a numerical example.
|
|
17:10-17:30, Paper ThB2.6 | |
Physical Watermarking for Replay Attack Detection in Continuous-Time Systems (I) |
|
Yaghooti, Bahram | Washington University in St. Louis |
Romagnoli, Raffaele | Carnegie Mellon University |
Sinopoli, Bruno | Washington University in St Louis |
Keywords: Fault detection and identification, Stochastic systems, Sampled data control
Abstract: Physical watermarking is a well established technique for replay attack detection in cyber-physical systems (CPSs). Most of the watermarking methods proposed in the literature are designed for discrete-time systems. In general real physical system evolve in continuous time. In this paper, we analyze the effect of watermarking on sampled-data continuous-time systems controlled via a Zero-Order Hold. We investigate the effect of sampling on detection performance and we provide a procedure to find a suitable sampling period that ensures detectability and acceptable control performance. Simulations on a quadrotor system are used to illustrate the effectiveness of the theoretical results.
|
|
ThB3 |
|
Nonlinear Systems II |
Regular Session |
Chair: Peaucelle, Dimitri | CNRS |
Co-Chair: Antoulas, Athanasios C. | Rice Univ |
|
15:30-15:50, Paper ThB3.1 | |
Fast and Robust Stability Region Estimation for Nonlinear Dynamical Systems |
|
Berthier, Eloïse | Inria - Ecole Normale Supérieure |
Carpentier, Justin | Inria |
Bach, Francis | Inria |
Keywords: Robust control, Lyapunov methods, Computational methods
Abstract: A linear quadratic regulator can stabilize a nonlinear dynamical system with a local feedback controller around a linearization point, while minimizing a given performance criteria. An important practical problem is to estimate the region of attraction of such a controller, that is, the region around this point where the controller is certified to be valid. This is especially important in the context of highly nonlinear dynamical systems. In this paper, we propose two stability certificates that are fast to compute and robust when the first, or second derivatives of the system dynamics are bounded. Associated with an efficient oracle to compute these bounds, this provides a simple stability region estimation algorithm compared to classic approaches of the state of the art. We experimentally validate its application to both polynomial and non-polynomial systems of various dimensions, including standard robotic systems, for estimating region of attractions around equilibrium points, as well as for trajectory tracking.
|
|
15:50-16:10, Paper ThB3.2 | |
L2 Induced Norm Analysis of Discrete-Time LTI Systems for Nonnegative Input Signals and Its Application to Stability Analysis of Recurrent Neural Networks |
|
Ebihara, Yoshio | Kyushu University |
Waki, Hayato | Institute of Mathematics for Industry, Kyushu University |
Magron, Victor, Liev | CNRS |
Mai, Ngoc Hoang Anh | LAAS-CNRS |
Peaucelle, Dimitri | CNRS |
Tarbouriech, Sophie | LAAS-CNRS |
Keywords: LMI's/BMI's/SOS's, Stability of nonlinear systems, Machine learning
Abstract: In this paper, we focus on the ``positive'' l2 induced norm of discrete-time linear time-invariant systems where the input signals are restricted to be nonnegative. To cope with the nonnegativity of the input signals, we employ copositive programming as the mathematical tool for the analysis. Then, by applying an inner approximation to the copositive cone, we derive numerically tractable semidefinite programming problems for the upper and lower bound computation of the ``positive'' l2 induced norm. This norm is typically useful for the stability analysis of feedback systems constructed from an LTI system and nonlinearities where the nonlinear elements provide only nonnegative signals. As a concrete example, we illustrate the usefulness of the ``positive'' l2 induced norm for the stability analysis of recurrent neural networks with activation functions being rectified linear units.
|
|
16:10-16:30, Paper ThB3.3 | |
Learning Reduced-Order Models of Quadratic Dynamical Systems from Input-Output Data |
|
Gosea, Ion Victor | Max Planck Institute for Dynamics of Complex Technical Systems |
KARACHALIOS, DIMITRIOS | Max Planck Institute for Dynamics of Complex Technical Systems |
Antoulas, Athanasios C. | Rice Univ |
Keywords: Reduced order modeling, Nonlinear system theory, Model/Controller reduction
Abstract: In this paper, we address an extension of the Loewner framework for learning quadratic dynamical systems from input-output data. The proposed method first constructs a reduced-order linear model from measurements of the classical transfer function. Then, this surrogate model is enhanced by incorporating a term that depends quadratically on the state. More precisely, we employ an iterative procedure based on least squares fitting that takes into account measured or computed data. Here, data represent transfer function values inferred from higher harmonics of the observed output, when the control input is purely oscillatory.
|
|
16:30-16:50, Paper ThB3.4 | |
Lyapunov-Based Temperature Regulation by Flow Reorientation |
|
Lensvelt, Ruud | Eindhoven University of Technology |
Speetjens, Michel | Eindhoven University of Technology |
Nijmeijer, Hendrik | Eindhoven University of Technology |
Keywords: Lyapunov methods, Fluid flow systems, Switched systems
Abstract: Transport of scalars, in the form of heat or chemicals, by fluid flow is a key feature for the effective operation of applications that range from chemical species mixing to subsurface resource extraction. Therefore, enhancing transport of these scalars by improving their dispersion will prove beneficial to a large variety of industries. Systems that involve scalar transfer from the boundary and have a substantial influence of diffusion and/or chemical reactions on heat/chemical transport are of particular interest. The system considered in this work intends to rapidly homogenize a scalar field by reorientating a laminar base flow. In conventional heating/mixing approaches a periodic reorientation scheme is designed towards effective fluid mixing and thus lacks robustness to perturbations required for widespread application. In this work we present two novel methods that accomplishes transport acceleration by adjusting the fluid flow reorientation. Rationale behind these methods is that influencing transport rates by fluid flow is analogous to influencing the decay rate of a Lyapunov function. This reasoning leads to the design of a bang-bang regulator and a general nonlinear regulator. We numerically investigate the performance of these regulators on a representative thermal flow problem: boundary heating of an initially cold fluid by reorientating of a 2D flow fields. Results show that the proposed regulators improve heating rates by upto 80 % compared to mere diffusive heating.
|
|
16:50-17:10, Paper ThB3.5 | |
Observability Analysis for Spacecraft Attitude Determination Using a Single Temperature Sensor |
|
Posielek, Tobias | German Aerospace Center (DLR) |
Reger, Johann | TU Ilmenau |
Keywords: Observers for nonlinear systems, Aerospace
Abstract: We consider the problem of spacecraft attitude determination using temperature data. Common algorithms fuse multiple temperature measurements to reconstruct the attitude. However, a single sensor already contains a lot of information due to the temperature dynamics inherent to the non-linear structure. In this work a rigorous observability analysis is carried out to determine the configurations in which the combination of a single temperature sensor with angular velocity measurements is sufficient for estimating the entire attitude. We evaluate the observability properties of the different configurations and give recommendations for the best configurations.
|
|
17:10-17:30, Paper ThB3.6 | |
Design of Positive Unknown Input Observer for a Class of Positive Nonlinear Systems |
|
Shafai, Bahram | Northeastern Univ |
Moradmand, Anahita | Northeastern University |
Keywords: Observers for nonlinear systems, Nonlinear system theory, Lyapunov methods
Abstract: This paper considers the design of nonlinear positive unknown input observer (NPUIO) for a class of positive nonlinear systems. Positivity and stability of nonlinear systems are analyzed and a possible design procedure for positive nonlinear systems without unknown input is provided as a preliminary exposition. To estimate the states of nonlinear system in the presence of unknown input, two classes of nonlinearity are considered. For the class of additive nonlinearity, the design equations for NPUIO are derived and reformulated in terms of LMI. The requirement of maintaining the positivity of observer parameters was a major challenge, which has been realized through the constraints imposed on design equations.
|
|
ThB4 |
|
Learning-Based Control IV |
Regular Session |
Chair: Nemeth, Balazs | SZTAKI Institute for Computer Science and Control |
Co-Chair: Mounier, Hugues | Université Paris Sud 11 |
|
15:30-15:50, Paper ThB4.1 | |
Stable and Robust Neural Network Controllers |
|
Sterud, Camilla | SINTEF Digital |
Moe, Signe | Norwegian University of Science and Technology, SINTEF |
Gravdahl, Jan Tommy | Norwegian Univ. of Science & Tech |
Keywords: Nonlinear system theory, Neural networks, Machine learning
Abstract: Neural networks are expressive function approimators that can be employed for state estimation in control problems. However, control systems with machine learning in the loop often lack stability proofs and performance guarantees, which are crucial for safety-critical applications. In this work, a feedback controller using a feedforward neural network of arbitrary size to estimate unknown dynamics is suggested. The controller is designed for solving a general trajectory tracking problem for a broad class of two-dimensional nonlinear systems. The controller is proven to stabilize the closed-loop system, such that it is input-to-state and finite-gain Lp-stable from the neural network estimation error to the tracking error. Furthermore, the controller is proven to make the tracking error globally and exponentially converge to a ball centered at the origin. When the neural network estimate is updated discretely, or the state measurements are affected by bounded noise, the convergence bound is shown to be dependent on the Lipschitz constant of the neural network estimator. In light of this, we demonstrate how regularization techniques can be beneficial when utilizing deep learning in control. Experiments on simulated data confirm the theoretical results.
|
|
15:50-16:10, Paper ThB4.2 | |
Learning Control Barrier Functions with High Relative Degree for Safety-Critical Control |
|
Wang, Chuanzheng | University of Waterloo |
Meng, Yiming | University of Waterloo |
Li, Yinan | University of Waterloo |
Smith, Stephen L. | University of Waterloo |
Liu, Jun | University of Waterloo |
Keywords: Lyapunov methods, Machine learning, Uncertain systems
Abstract: Control barrier functions have shown great success in addressing control problems with safety guarantees. These methods usually find the next safe control input by solving an online quadratic programming problem. However, model uncertainty is a big challenge in synthesizing controllers. This may lead to the generation of unsafe control actions, resulting in severe consequences. In this paper, we develop a learning framework to deal with system uncertainty. Our method mainly focuses on learning the dynamics of the control barrier function, especially for high relative degree with respect to a system. We show that for each order, the time derivative of the control barrier function can be separated into the time derivative of the nominal control barrier function and a remainder. This implies that we can use a neural network to learn the remainder so that we can approximate the dynamics of the real control barrier function. We show by simulation that our method can generate safe trajectories under parametric uncertainty using a differential drive robot model.
|
|
16:10-16:30, Paper ThB4.3 | |
Approximating a Deep Reinforcement Learning Docking Agent Using Linear Model Trees |
|
Gjærum, Vilde | NTNU |
Rørvik, Ella-Lovise | TrønderEnergi |
Lekkas, Anastasios | Norwegian University of Science and Technology |
Keywords: Autonomous systems, Neural networks, Maritime
Abstract: Deep reinforcement learning has led to numerous notable results in robotics. However, deep neural networks (DNNs) are unintuitive, which makes it difficult to understand their predictions and strongly limits their potential for real-world applications due to economic, safety, and assurance reasons. To remedy this problem, a number of explainable AI methods have been presented, such as SHAP and LIME, but these can be either be too costly to be used in real-time robotic applications or provide only local explanations. In this paper, the main contribution is the use of a linear model tree (LMT) to approximate a DNN policy, originally trained via proximal policy optimization, for an autonomous surface vehicle with five control inputs performing a docking operation. The two main benefits of the proposed approach are: a) LMTs are transparent which makes it possible to associate directly the outputs (control actions, in our case) with specific values of the input features, b) LMTs are computationally efficient and can provide information in real-time. In our simulations, the opaque DNN policy controls the vehicle and the LMT runs in parallel to provide explanations in the form of feature attributions. Our results indicate that LMTs can be a useful component within digital assurance frameworks for autonomous ships.
|
|
16:30-16:50, Paper ThB4.4 | |
A Hierarchical Primitive-Based Learning Tracking Framework for Unknown Observable Systems Based on a New State Representation |
|
Radac, Mircea-Bogdan | Politehnica University of Timisoara |
Lala, Timotei | Politehnica University of Timisoara, Department of Automation An |
Keywords: Intelligent systems, Neural networks, Robotics
Abstract: A three-level learning-based hierarchical tracking framework is validated in this work and aims at endowing control systems with generalization capabilities specific to artificial intelligence. The framework operates at three levels: the low-level L1 is concerned with ensuring a linear time-invariant (LTI) behavior from the reference input to the controlled output in terms of reference tracking. The second level L2 acts on top of the linearized closed-loop dynamics, to learn primitive pairs (reference inputs-controlled outputs pairs), in an entirely experimentally-driven style, using Iterative Learning Control (ILC). These primitives are optimally learned for tracking a desired trajectory and they naturally learn by trials, due to the ILC principle. Finally, the third level L3 uses the learned primitive pairs to extrapolate the optimal behavior to new desired trajectories, without relearning by trials. Level L3 is able to predict the optimal reference inputs that ensure accurate tracking in new tracking scenarios, which is a feature specific to intelligent beings. The framework is validated on a multivariable nonlinear two-joints rigid planar manipulator.
|
|
16:50-17:10, Paper ThB4.5 | |
Systematic Comparison of Numerical Differentiators and an Application to Model-Free Control |
|
Othmane, Amine | Saarland University |
Rudolph, Joachim | Saarland University |
Mounier, Hugues | Université Paris Sud 11 |
Keywords: Signal processing, Filtering, Intelligent systems
Abstract: Recent tuning guidelines for algebraic differentiators based on the analysis of the Fourier transform of the kernels are reviewed. A region of validity for the previous analyses carried out for high frequencies is proposed and the results are related to those based on the L2 -norm of the filters’ amplitudes. These results are then used for a systematic comparison with established approaches from the literature (high-gain and sliding-mode differentiators) for the estimation of the second derivative of a known signal corrupted by a known disturbance. It is shown that properly tuned and discretized algebraic differentiators outperform the other analysed approaches in terms of robustness, convergence time, and tuning simplicity. The tuning guidelines are then used in the context of control of a partially known system, thus bridging the gap to the tuning of recent model-free control approaches.
|
|
17:10-17:30, Paper ThB4.6 | |
Stability and Tracking Performance Analysis for Control Systems with Feed-Forward Neural Networks |
|
Lelkó, Attila | SZTAKI Institute for Computer Science and Control |
Nemeth, Balazs | SZTAKI Institute for Computer Science and Control |
Gaspar, Peter | SZTAKI |
Keywords: Autonomous systems, Neural networks, V&V of control algorithms
Abstract: Application of learning approaches in control systems can pose challenges against the reliability of neural networks, e.g. in safety-critical applications provable guarantees on the achieved safety level are requested. It motivates the development of analysis methods, with which stability and performance level of the controlled systems with neural networks can be verified. This paper provides an analysis method on stability and tracking performance of control systems, which contain feed-forward neural networks with one hidden layer. The method is based on the approximation of the neural networks in the form of a set of discrete linear systems. The core of the analysis is an optimization method, whose constraints contain stability conditions on the set of systems and the result of the optimization is in relation with the decay rate for tracking performance. For the selection of number of linear systems in the optimization the scenario approach is used, with which the probability for preserving stability and performance level of the closed-loop system is scaled. The effectiveness of the proposed method is illustrated through simulation examples, e.g. on a robotic arm control and on a vehicle cruise control.
|
|
ThB5 |
|
Networked Systems |
Regular Session |
Chair: Cao, Ming | University of Groningen |
Co-Chair: Kouvelas, Anastasios | ETH Zurich |
|
15:30-15:50, Paper ThB5.1 | |
Modelling Behavioural Preferences in Epidemic Models for Sexually Transmitted Infections on Temporal Networks |
|
Frieswijk, Kathinka | University of Groningen, ENgineering and TEchnology Institute Gr |
Zino, Lorenzo | University of Groningen |
Cao, Ming | University of Groningen |
Keywords: Network analysis and control, Control over networks, Agents networks
Abstract: In this paper, we propose a temporal model for the spreading of curable sexually transmitted infections (STIs). The model is developed within the framework of activity-driven networks, which allows to model the time-varying pattern of sexual encounters and the individuals' heterogeneity in their proclivity to initiate them. Our model explicitly includes the delay between infectiousness and symptoms onset, and individuals' behavioural preferences for the use of protection during encounters. Behavioural preferences evolve according to a nontrivial mechanism that accounts for the perceived risks, the cost of adopting protective measures, and the persuasive effect of interactions with individuals who have a different preference. In the limit of large-scale populations, we use a mean-field approach to derive the epidemic threshold and study the effect of two control measures on the spread of STIs: i) routine screening at STI clinics, and ii) condom (social) marketing campaigns. Our results reveal the important effect of routine screening for STIs, which has emerged as a key factor to favour stability of the disease-free equilibrium, while marketing campaigns can be very effective in mitigating endemic diseases.
|
|
15:50-16:10, Paper ThB5.2 | |
Averaging and Cluster Synchronization of Kuramoto Oscillators |
|
Kato, Rui | Tokyo Institute of Technology |
Ishii, Hideaki | Tokyo Institute of Technology |
Keywords: Network analysis and control, Linear time-varying systems, Applications in neuroscience
Abstract: In this paper, we investigate cluster synchronization of heterogeneous Kuramoto oscillators, where multiple synchronized groups of oscillators coexist in a connected network. Motivated by recent studies on brain networks, we provide a framework to analyze stability of the cluster synchronization manifold via timescale separation. A condition known as almost equitable partitions is employed to characterize an invariant manifold of the Kuramoto dynamics. Relying on averaging methods, we show that fast inter-cluster oscillations help the oscillators to achieve synchronization within clusters. In order to prove the main results, we extend conventional averaging methods to accommodate our problem setting. Then, we derive a general stability condition for multi-cluster synchronization. To find a less conservative condition, two-cluster synchronization with a special structure is also considered. Our results are demonstrated through a numerical example of a network with three clusters.
|
|
16:10-16:30, Paper ThB5.3 | |
H-Infinity Network Optimization for Edge Consensus |
|
Farhat, Omar | American University of Beirut |
Abou Jaoude, Dany | American University of Beirut |
Hudoba de Badyn, Mathias | ETH, Zürich |
Keywords: Network analysis and control, H2/H-infinity methods, Concensus control and estimation
Abstract: This paper examines the H-infinity performance problem of the edge agreement protocol for networks of agents operating on independent time scales, connected by weighted edges, and corrupted by exogenous disturbances. H-infinity-norm expressions and bounds are computed that are then used to derive new insights on network performance in terms of the effect of time scales and edge weights on disturbance rejection. We use our bounds to formulate a convex optimization problem for time scale and edge weight selection. The applicability of the derived results is illustrated via numerical examples.
|
|
16:30-16:50, Paper ThB5.4 | |
Strong Structural Functional Controllability of Networks |
|
Mousavi, Shima Sadat | ETH Zurich |
Kouvelas, Anastasios | ETH Zurich |
Keywords: Network analysis and control, Control over networks, Large-scale systems
Abstract: In this paper, we examine functional controllability for a family of linear time-invariant (LTI) networks having the same wiring diagrams. When a system is functional controllable, with a suitable choice of input, its output can follow the desired trajectory over a time period. In this work, we present sufficient combinatorial conditions for strong structural functional controllability of networks, that ensure the functional controllability of all LTI systems defined over the same system graph. Also, by defining a set of target nodes, a procedure for selecting a set of control nodes, rendering the network strongly structurally functional controllable, is proposed. A comparison between the results of this work and the existing results on strong structural output controllability and weak structural functional controllability is illustrated through some examples.
|
|
16:50-17:10, Paper ThB5.5 | |
A Binary Homophily Model for Opinion Dynamics |
|
De Pasquale, Giulia | University of Padova |
Valcher, Maria Elena | Universita' Di Padova |
Keywords: Network analysis and control, Agents networks, Modeling
Abstract: In this paper we propose a discrete time binary model, based on the homophily social mechanism, that dynamically reduces the cognitive dissonance among the agents in a social network. We show that the binary homophily model can drive an initially structurally unbalanced network towards a socially balanced one. In order to characterise non- structurally balanced equilibrium points, we introduce a (V, Σ)- factorization that finds an interesting interpretation in terms of structurally balanced classes, and can be used to investigate the case of 3 classes and to provide a complete analysis of the converge to equilibrium for small networks.
|
|
17:10-17:30, Paper ThB5.6 | |
On an Optimal Control Approach Toward Mitigating an SIS Epidemic Model |
|
Cenedese, Carlo | University of Groningen |
Cucuzzella, Michele | University of Groningen |
Zino, Lorenzo | University of Groningen |
Cao, Ming | University of Groningen |
Scherpen, Jacquelien M.A. | Fac. Science and Engineering, University of Groningen |
van der Schaft, Arjan J. | University of Groningen |
Keywords: Modeling, Cooperative autonomous systems, Optimization
Abstract: Understanding how to effectively control an epidemic spreading on a network is a problem of paramount importance for the scientific community. The ongoing COVID-19 pandemic has highlighted the need for policies that mitigate the spread, without relying on pharmaceutical interventions, that is, without the possibility of controlling the recovery process. These policies typically entail lockdowns and mobility restrictions, having thus nonnegligible socio-economic consequences for the population. Here, we propose an optimal control approach to design effective policies to ``flatten the epidemic curve," limiting the negative consequences for the society, providing thus control-theoretic tools to assist the public health authorities to balance safety and normalcy during an epidemic outbreak.
|
|
ThB6 |
|
Nonlinear Model Predictive Control |
Regular Session |
Chair: Sename, Olivier | Grenoble INP / GIPSA-Lab |
Co-Chair: Werner, Herbert | Hamburg University of Technology |
|
15:30-15:50, Paper ThB6.1 | |
Short-Sighted Robust LPV Model Predictive Control: Application to Semi-Active Suspension Systems |
|
Menezes Morato, Marcelo | Universidade Federal De Santa Catarina |
Normey-Rico, Julio Elias | Federal University of Santa Catarina |
Sename, Olivier | Grenoble INP / GIPSA-Lab |
Keywords: Predictive control for nonlinear systems, Linear parameter-varying systems, Automotive
Abstract: This paper develops a novel Linear Parameter Varying (LPV) Model Predictive Control (MPC) algorithm for Semi-Active Suspension systems. The current state-of-the- art comprises two possible implementations: a) to consider the future variations of the LPV scheduling variables as uncertainties, thereby solving a robust optimization, which is usually time-consuming; or b) to estimate the future scheduling variables and solve a sub-optimal quadratic program, which can be evaluated rapidly. This paper proposes a control paradigm in between these paths, considering a robust min-max procedure with small predictions horizons, being implementable within the short 5 ms sampling period of the suspension system. The method includes terminal ingredients, derived via LMIs, that ensure input-to-state stability and recursive feasibility. Realistic simulations show the effectiveness of the proposed method, when compared against a nonlinear MPC and a sub-optimal LPV MPC. The results show that the method is indeed able to run in real-time (in the order of milliseconds), almost as fast as the sub-optimal MPC, while still guaranteeing good safety and comfort performances for the vehicle.
|
|
15:50-16:10, Paper ThB6.2 | |
Regional Control Laws As Fallback Strategy for Nonlinear MPC: Application to Wind Turbine Control |
|
Dyrska, Raphael | Ruhr-Universität Bochum |
Mitze, Ruth | Ruhr-Universität Bochum |
Mönnigmann, Martin | Ruhr-Universität Bochum |
Keywords: Predictive control for nonlinear systems, Constrained control, Electrical power systems
Abstract: We propose a method for determining fallback control laws for nonlinear MPC problems. A fallback strategy is needed whenever a nonlinear MPC problem cannot be solved in due time. In this paper, we use the concept of regional MPC to compute local control laws that mimic the solution expected from the original problem. In regional MPC, the piecewise affine structure of the solution of linear-quadratic optimization problems is exploited by computing the current polytope and control law online. We use the regional control law, which is naturally stabilizing for more states than its corresponding polytope, and dismiss its region. In every time step, we use the solution of the original MPC problem as a basis to determine the regional control law. If in one time step no solution can be found, the regional control law from the previous time step is used as a fallback strategy. We show the effectiveness with an example of a wind turbine.
|
|
16:10-16:30, Paper ThB6.3 | |
Implementation of Fast Predictive Controllers on FPGA Platforms Based on Parallel Lipschitz Interpolation |
|
Nadales, J.M. | Universidad De Sevilla |
Manzano, Jose Maria | Universidad Loyola Andalucía |
Barriga Barros, Angel | Universidad De Sevilla |
Limon, Daniel | Universidad De Sevilla |
Keywords: Predictive control for nonlinear systems, Machine learning, Emerging control applications
Abstract: The implementation of nonlinear model predictive controllers for systems operating at high frequencies constitutes a significant challenge, mainly because of the complexity and time consumption of the optimization problem involved. An alternative that has been proposed is the employment of data-driven techniques to offline learn the control law, and then to implement it on a target embedded platform. Following this trend, in this paper we propose the implementation of predictive controllers on FPGA platforms making use of a parallel version of the machine learning technique known as Lipschitz interpolation. By doing this, computation times can be enormously accelerated while ensuring a given bound on the approximation error. The results are compared to those obtained when the sequential algorithm runs on standard CPU platforms, and when the system is controlled by solving the optimization problem online, in terms of the error made and computing times. This method is validated in a case study where the nonlinear model predictive controller is employed to control a self-balancing two-wheel robot.
|
|
16:30-16:50, Paper ThB6.4 | |
Koopman Operator-Based Model Predictive Control with Recursive Online Update |
|
Martínez Calderón, Horacio | IAV GmbH |
Schulz, Erik | IAV GmbH |
Oehlschlägel, Thimo | IAV GmbH |
Werner, Herbert | Hamburg University of Technology |
Keywords: Predictive control for nonlinear systems, Adaptive control, Linear parameter-varying systems
Abstract: The Koopman operator framework allows to embed a nonlinear system into a linear one. This enables the analysis, estimation, and control of nonlinear dynamics with linear methods. Controllers based on the Koopman operator (KO) are often model predictive control (MPC) schemes. The performance of an MPC depends on the prediction accuracy of its model. Hence, it is meaningful to update the model online if the predictions are not sufficiently accurate. In this work, we approach this problem by using a recursive least squares (RLS) algorithm with forgetting factor. Furthermore, we show in an empirical case study that combining the KO with an online update and the recently proposed quasi-linear parameter varying model predictive control (qLMPC) algorithm results in an efficient control scheme.
|
|
16:50-17:10, Paper ThB6.5 | |
A Velocity quasiLPV-MPC Algorithm for Wind Turbine Control |
|
Dittmer, Antje | TU Hamburg |
Sharan, Bindu | Hamburg Technical University |
Werner, Herbert | Hamburg University of Technology |
Keywords: Predictive control for nonlinear systems, Linear parameter-varying systems, Modeling
Abstract: This paper presents a velocity quasi linear parameter varying model predictive control (quasiLPV-MPC) algorithm for the whole wind turbine operating range, lever-aging MIMO pitch and torque control. Rigorous modelling of wind turbine dynamics based on Lagrange’s equation is presented in detail and validated with FAST. Implementing the resulting non-linear wind turbine model in Simulink instead of including the FAST environment accelerates the simulationby a factor of hundred and hence yields an efficient controller test environment. In this work, the undesirable power variance as well as fore-aft tower motions are decreased considerably, up to a ratio of hundred, compared to a baseline PI controller. The implementation as velocity quasiLPV-MPC mitigates the need to store equilibrium input and state vectors, decreasing memory usage considerably. Moreover, the proposed predictive algorithm converges in one iteration, requiring 2 ms calculationtime, making it suitable for real-time applications.
|
|
17:10-17:30, Paper ThB6.6 | |
Model Predictive Control for Micro Aerial Vehicles: A Survey |
|
Nguyen, Huan | Norwegian University of Science and Technology |
Kamel, Mina | ETHZ |
Alexis, Konstantinos | NTNU |
Siegwart, Roland Y. | ETH Zürich |
Keywords: Predictive control for nonlinear systems, Predictive control for linear systems
Abstract: This paper presents a review of the design and application of model predictive control strategies for Micro Aerial Vehicles and specifically multirotor configurations such as quadrotors. The diverse set of works in the domain is organized based on the control law being optimized over linear or nonlinear dynamics, the integration of state and input constraints, possible fault-tolerant design, if reinforcement learning methods have been utilized and if the controller refers to free-flight or other tasks such as physical interaction or load transportation. A selected set of comparison results are also presented and serve to provide insight for the selection between linear and nonlinear schemes, the tuning of the prediction horizon, the importance of disturbance observer-based offset-free tracking and the intrinsic robustness of such methods to parameter uncertainty. Furthermore, an overview of recent research trends on the combined application of modern deep reinforcement learning techniques and model predictive control for multirotor vehicles is presented. Finally, this review concludes with explicit discussion regarding selected open-source software packages that deliver off-the-shelf model predictive control functionality applicable to a wide variety of Micro Aerial Vehicle configurations.
|
|
ThB7 |
|
Autonomous Robots |
Regular Session |
Chair: Persson, Niklas | Mälardalen University |
Co-Chair: Nikolakopoulos, George | Luleå University of Technology, Sweden |
|
15:30-15:50, Paper ThB7.1 | |
Data-Driven, Ground Truth-Free Tuning of an Adaptive Monte Carlo Localization Method for Urban Scenarios |
|
Giovagnola, Jessica | Politecnico Di Milano |
Rigamonti, Davide | Politecnico Di Milano - Dipartimento Di Elettronica, Informazion |
Corno, Matteo | Politecnico Di Milano |
Chen, Weidong | Shanghai Jiao Tong University |
Savaresi, Sergio M. | Politecnico Di Milano |
Keywords: Autonomous robots, Autonomous systems, Agents and autonomous systems
Abstract: Robust robot localization in outdoor urban environments is a challenging task. One of the most popular localization approaches is Adaptive Monte Carlo Localization (AMCL). AMCL can accurately locate a robot in a mapped environment. However, tuning the many parameters that characterize the algorithm is more of an art than a science. Furthermore, trial-and-error tuning requires access to a ground truth which, especially in outdoor urban scenarios, may not be available. In this paper, we propose a tuning method for AMCL. The proposed method tunes the most important AMCL parameters without the need of a continuous ground truth by optimizing the estimated path smoothness and using the passage through a finite number of gateways as constraints. The optimization algorithm exploits Bayesian Optimization in order to limit the number of tuning runs. Data collected with an instrumented robot on a public road validate the approach. The proposed tuning yields a robust localization with minimal manual intervention in the tuning.
|
|
15:50-16:10, Paper ThB7.2 | |
A Comparative Analysis and Design of Controllers for Autonomous Bicycles |
|
Persson, Niklas | Mälardalen University |
Andersson, Tom | Mälardalen University |
Fattouh, Anas | Mälardalen University |
Ekström, Martin | Mälardalen University |
Papadopoulos, Alessandro Vittorio | Mälardalen University |
Keywords: Autonomous robots, Mechatronics, Modeling
Abstract: In this paper, we develop and compare the performance of different controllers for balancing an autonomous bicycle. The evaluation is carried out both in simulation, using two different models, and experimentally, on a bicycle instrumented with only lightweight components, and leaving the bicycle structure practically unchanged. Two PID controllers, a Linear Quadratic Regulator (LQR), and a fuzzy controller are developed and evaluated in simulations where both noise and disturbances are induced in the models. The simulation shows that the LQR controller has the best performance in the simulation scenarios. Experimental results, on the other hand, show that the PID controllers provide better performance when balancing the instrumented bicycle.
|
|
16:10-16:30, Paper ThB7.3 | |
Autonomous UAV Navigation in an Unknown Environment Via Multi-Trajectory Model Predictive Control |
|
Saccani, Danilo | Politecnico Di Milano |
Fagiano, Lorenzo | Politecnico Di Milano |
Keywords: Autonomous systems, UAV's, Predictive control for linear systems
Abstract: A novel model predictive control (MPC) formulation, named multi-trajectory MPC (mt-MPC), is presented and applied to the problem of autonomous navigation of an unmanned aerial vehicle (UAV) in an unknown environment. The UAV is equipped with a LiDAR sensor, providing only a partial description of the surroundings and resulting in time varying constraints as the vehicle navigates among the obstacles. The control system layout is hierarchical: the low-level loops stabilize the vehicle’s trajectories and track the set-points commanded by the high-level, mt-MPC controller. The latter is required to plan the UAV trajectory trading off safety, i.e. to avoid collisions with the uncertain obstacles, and exploitation, i.e. to reach an assigned target location. To achieve this goal, mt-MPC considers different future state trajectories in the same Finite Horizon Optimal Control Problem (FHOCP), enabling a partial decoupling between constraint satisfaction (safety) and cost function minimization (exploitation). Recursive feasibility and, consequently, persistent obstacle avoidance guarantees are derived under the assumption of a time invariant environment. The performance of the approach is studied in simulation and compared with that of a standard MPC, showing good improvement.
|
|
16:30-16:50, Paper ThB7.4 | |
Online Motion Planning Based on Nonlinear Model Predictive Control with Non-Euclidean Rotation Groups |
|
Rösmann, Christoph | TU Dortmund University |
Makarow, Artemi | TU Dortmund University |
Bertram, Torsten | Technische Universität Dortmund |
Keywords: Autonomous robots, Automotive, Predictive control for nonlinear systems
Abstract: This paper proposes an online motion planning approach to robot navigation based on nonlinear model predictive control. In robot navigation, state spaces often include rotational components which span over non-Euclidean rotation groups. The proposed approach applies nonlinear increment and difference operators in the entire optimization scheme to explicitly consider these groups. Realizations include but are not limited to quadratic form and time-optimal objectives. A complex parking scenario for the kinematic bicycle model demonstrates the effectiveness and practical relevance of the approach. In case of simpler robots (e.g. differential drive), a comparative analysis in a hierarchical planning setting reveals comparable computation times and performance. The approach is available in a modular and highly configurable open-source C++ software framework.
|
|
16:50-17:10, Paper ThB7.5 | |
An Adaptive Observer Approach to Slip Estimation for Agricultural Tracked Vehicles |
|
Tazzari, Roberto | University of Bologna |
Azzollini, Ilario Antonio | University of Bologna |
Marconi, Lorenzo | Univ. Di Bologna |
Keywords: Autonomous robots, Adaptive systems
Abstract: The paper deals with autonomous Unmanned Ground Vehicles developed for precision agriculture contexts. The focus of the paper is on the design of an adaptive observer for slip estimation ensuring exponential convergence to the real slip coefficients. Uniform global exponential stability of the origin of the error system is shown via Lyapunov analysis and persistency of excitation arguments. Furthermore, robustness to additive perturbations is shown in terms of Input-to-State Stability. Experimental results validate the effectiveness of the proposed estimator even in presence of noisy measurements.
|
|
17:10-17:30, Paper ThB7.6 | |
Geometry Aware NMPC Scheme for Morphing Quadrotor Navigation in Restricted Entrances |
|
Papadimitriou, Andreas | Luleå University of Technology |
Sharif Mansouri, Sina | Lulea University |
Kanellakis, Christoforos | Luleå University of Technology |
Nikolakopoulos, George | Luleå University of Technology, Sweden |
Keywords: Autonomous robots, UAV's
Abstract: Geometry-morphing Micro Aerial Vehicles (MAV) are gaining more attention lately, since, their ability to modify their geometric morphology in-flight increases their versatility while expanding their application range. In this novel research field, most of the works focus on the platform design and on the low-level control for maintaining stability after changing its shape. Nevertheless, another aspect of geometry morphing MAV is the association of its morphology with respect to the shape and structure of the environment. In this article, we propose a novel Nonlinear Model Predictive Control (NMPC) structure that modifies the morphology of a quadrotor based on the environment entrances' geometrical shape. The proposed method considers restricted entrances as a constraint in the NMPC and modifies the arm configuration of the MAV to provide a collision-free path from the initial position to the desired goal while passing through the entrance. Multiple simulation results present the performance and efficiency of the proposed scheme in scenarios where the quadrotor is commanded to pass through narrow entrances.
|
|
ThB8 |
|
Sliding Mode Control |
Regular Session |
Chair: Cavallo, Alberto | Università Degli Studi Della Campania "L. Vanvitelli" |
Co-Chair: BOUKAL, YASSINE | Altran by Capgemini |
|
15:30-15:50, Paper ThB8.1 | |
A Novel Integral Nonlinear Hyperplane-Based Sliding Mode Controller for a Quadrotor Vehicle Subjected to Disturbances |
|
Labbadi, Moussa | Mohammed V University, Mohammadia School of Engineers, Engineeri |
BOUKAL, YASSINE | Altran by Capgemini |
ZERROUGUI, Mohamed | Aix Marseille University |
Boudaraia, Karima | Mohammed V University in Rabat, EMI, Engineering 3S Research Cen |
Cherkaoui, Mohamed | Université Mohamed V, EMI |
Keywords: Sliding mode control, UAV's, Autonomous systems
Abstract: One of the most critical issues for a precision flight and safety of quadrotor UAV is external disturbances. To this end, a novel integral nonlinear hyperplane-based sliding mode control (INH-SMC) technique is presented for a quadrotor system against wind disturbances. The proposed INH-SMC method is designed for the attitude/position of a quadrotor. The proposed controller enhances the tracking performance, the response of quadrotor states, and copes with wind disturbances. The stability of the INH-SMC technique for the quadrotor is verified and analyzed using the Lyapunov theory. The results of the simulation demonstrate the effectiveness of the suggested method against wind disturbances. Also, a comparative study between the proposed controller and super-twisting proportional-derivative-integral sliding mode controller (STPIDSMC) has been carried out to verify the superiority of the proposed method in terms of accurate tracking, fast convergence, and robust performance against wind disturbances.
|
|
15:50-16:10, Paper ThB8.2 | |
Generalized Super-Twisting Control of a Dual Active Bridge for More Electric Aircraft |
|
Russo, Antonio | Università Della Campania L. Vanvitelli |
Canciello, Giacomo | Department of Industrial and Information Engineering, University |
Cavallo, Alberto | Università Degli Studi Della Campania "L. Vanvitelli" |
Keywords: Sliding mode control, Stability of nonlinear systems, Electrical power systems
Abstract: In this work the control of a Dual Active Bridge (DAB) is presented. The application is within the innovative framework of the More Electric Aircraft (MEA) concept, and its goal is to recharge a battery pack in the case of standard loading conditions, while using the batteries to supply energy to an extra-load in the case of overload. The control objective is achieved through the adoption of the Generalized Super-Twisting algorithm. Rigorous mathematical proofs of stability are given for the controlled system and for supervisor switching between normal condition and overload. The effectiveness of the proposed strategy is shown by detailed simulations in Matlab/Stateflow/Electronics.
|
|
16:10-16:30, Paper ThB8.3 | |
Enhanced Variable-Gain Sliding Mode Control for Robot Manipulators |
|
Incremona, Gian Paolo | Politecnico Di Milano |
Rubagotti, Matteo | Nazarbayev University |
Tanelli, Mara | Politecnico Di Milano |
Ferrara, Antonella | University of Pavia |
Keywords: Sliding mode control, Robust adaptive control, Robotics
Abstract: Robotic manipulators must operate in complex scenarios, which make the overall operational space quite large, and the system dynamics within that space subject to significant variations and uncertainties. Sliding mode control (SMC) strategies have been successfully applied in this context, yet, if a worst-case approach is taken in the face of large operational variations, the resulting performance may happen to be suboptimal. Moreover, attention must be paid, in an industrial setting, to vibrations that may be induced on the robot joints by the presence of chattering induced by the SMC algorithm. This work tests, in a challenging application context, a recently-proposed r-order SMC strategy that encompasses both continuous and discrete adaptation strategies to adjust its parameters, giving rise to an overall switched approach. In particular, two realistic case studies of robot motion control are discussed, proving its effectiveness in enhancing both performance and robustness in complex operational scenarios.
|
|
16:30-16:50, Paper ThB8.4 | |
Design of Sliding Mode Controllers for Quadrotor Vehicles Via Flatness-Based Feedback and Feedforward Linearization Strategies |
|
Bascetta, Luca | Politecnico Di Milano |
Incremona, Gian Paolo | Politecnico Di Milano |
Keywords: Sliding mode control, UAV's, Uncertain systems
Abstract: Unmanned Aerial Vehicles (UAVs) have to operate in complex environments, characterized by disturbances of different nature that affect the system performance. Moreover, system dynamics can be altered by unavoidable modeling uncertainties, that can further decrease the control performance. This motivates the introduction of robust control strategies and, among them, Sliding Mode Control (SMC) represents a viable solution, provided that the UAV model is led back to a normal form, suitable for control design purposes. This paper investigates two flatness-based linearization approaches, a feedback and a feedforward one, that transform the nonlinear and coupled quadrotor model into a canonical form eligible to design a trajectory tracking controller based on a battery of Higher-Order Sliding Mode (HOSM) regulators. Simulation results, based on a realistic model of a quadrotor, are presented to assess the performance of the proposed control system.
|
|
16:50-17:10, Paper ThB8.5 | |
Sliding Mode Control of a Dc-Dc Dual Active Bridge Using the Generalized Space State Averaging Description |
|
Dòria-Cerezo, Arnau | Technical Univ. of Catalonia (UPC) |
Serra, Federico | Universidad Nacional De San Luis |
Biel, Domingo | Universitat Politècnica De Catalunya |
Griño, Robert | Universitat Politecnica De Catalunya |
Keywords: Sliding mode control, Power electronics
Abstract: This paper presents a sliding mode control strategy for a dc-dc dual active bridge converter. The controller is based on a truncated model obtained using the generalized state space averaging method that transforms the mixed dc-ac dynamics of the converter into a regulation problem. The proposed controller, that uses a dynamic extension to overcome the structural problem of the non-affine control input, provides good results in terms of performance and robustness. Numerical simulations are included to validate the proposed modelling methodology and the control design.
|
|
17:10-17:30, Paper ThB8.6 | |
Fault Identification Via Sliding Mode Observers in Nonlinear Systems Not Satisfying Matching and Minimum Phase Conditions |
|
Zhirabok, Alexey | Far Eastern Federal University |
Zuev, Alexander | Institute of Automationa and Control Processes FEB RAS |
Shumsky, Alexey | Far Eastern Federal University |
Keywords: Fault estimation, Nonlinear system theory
Abstract: The paper is devoted to the problem of fault identification (reconstruction) in systems described by nonlinear models under the unmatched disturbances. A novel approach to construct sliding mode observer is suggested for systems which do not satisfy the general conditions required for fault identification, in particular, minimum phase and matching conditions. The suggested approach is based on the reduced order model of the original system insensitive to the disturbances. This allows to reduce the dimension of sliding mode observer and relax the limitations imposed on the original system.
|
|
ThB9 |
|
Energy Systems |
Regular Session |
Chair: Vasca, Francesco | University of Sannio |
Co-Chair: Alvarez, Miguel | University of Maryland |
|
15:30-15:50, Paper ThB9.1 | |
SOC and Diffusion Rate Estimation in Redox Flow Batteries: An I&I-Based High-Gain Observer Approach |
|
Clemente, Alejandro | Institut De Robòtica I Informàtica Industrial, CSIC-UPC |
Cecilia, Andreu | Institut De Robòtica I Informàtica Industrial, CSIC-UPC |
Costa-Castello, Ramon | Universitat Politècnica De Catalunya (UPC) |
Keywords: Energy systems, Observers for nonlinear systems, Adaptive control
Abstract: This paper presents an adaptive non-linear observer for the state of charge estimation in vanadium redox flow batteries. The study is based directly on the use of nonlinear equations that describe the evolution of the species concentration inside the system, and a nonlinear cell voltage expression that takes into account the effect of overpotentials. Moreover, a more realistic approach is used which does not consider that the electrolyte concentration is the same in the catholyte and anolyte sides of the system. It is shown that the state of charge can be estimated through the measurement of the output voltage and a high-gain observer. Nonetheless, the accuracy of the estimation is affected by uncertainty in the system diffusion rates. For this reason, the observer is robustified by means of an immersion and invariance adaptive parameter estimation. The results are validated in a numerical simulation.
|
|
15:50-16:10, Paper ThB9.2 | |
Improved State of Charge Estimation of Lithium-Ion Battery Cells |
|
Su, Jiayi | Marquette University |
Strandt, Alia | Marquette University |
Schneider, Susan | Marquette University |
Yaz, Edwin | Marquette University |
Josse, Fabien | Marquette University |
Keywords: Energy systems, Stochastic filtering, Adaptive systems
Abstract: Lithium-ion battery cells are widely used in a variety of applications. An accurate online estimation technique to determine the State of Charge (SOC) can improve the safety of Lithium-ion cells, their performance, and their life cycle. However, available nonlinear models make estimation challenging. In this work, an estimation technique which combines the Multiple Model Adaptive Estimation (MMAE) and the Extended Kalman filter (EKF) is introduced. The combination of these two techniques improves the SOC estimation accuracy and avoids the drawbacks of each technique. Simulation results for a LiFePO4 Lithium-ion cell demonstrate that this combined technique provides smaller estimation error compared to either the MMAE or the EKF alone.
|
|
16:10-16:30, Paper ThB9.3 | |
An Integrated Model of the Kite and Tether Dynamics of a Marine Hydrokinetic Energy Harvesting System |
|
Alvarez, Miguel | University of Maryland |
Bhattacharjee, Debapriya | University of Maryland |
Vermillion, Christopher | University of North Carolina at Charlotte |
Fathy, Hosam K. | The Pennsylvania State University |
Keywords: Energy systems, Modeling, Differential algebraic systems
Abstract: This paper models the dynamics of a marine tethered energy harvesting system focusing on exploring the sensitivity of the kite dynamics to tether parameters. These systems repetitively reels a kite out at high tension, then reels it in at low tension, in order to harvest energy. The kite’s high lift-to-drag ratio makes it possible to maximize net energy output through periodic cross-current flight. Significant modeling efforts exist in the literature supporting such energy maximization. The goal of this paper is to address the need for a simple model capturing the interplay between the system’s kite and tether dynamics. The authors pursue this goal by coupling a partial differential equation (PDE) model of tether dynamics with a point mass model of translational kite motion. One can simplify this model significantly by neglecting tether mass and compliance, effectively transforming the tether into a kinematic constraint. Simulation results show that the coupling effects discarded through such a simplification are non-trivial. For example, even when the tether is neutrally buoyant, its transverse vibrations can still cause significant oscillations in net kite forcing.
|
|
16:30-16:50, Paper ThB9.4 | |
Battery State of Health Estimation Via Reinforcement Learning |
|
Natella, Domenico | University of Sannio |
Vasca, Francesco | University of Sannio |
Keywords: Energy systems, Identification, Iterative learning control
Abstract: The state of health of a battery characterizes its performance in terms of loss of capacity compared to the beginning of its life. This paper proposes a reinforcement learning algorithm for identifying the capacity of lithium-ion batteries. The training phase of the algorithm is based on data derived from constant current and constant voltage charging operations. The technique exploits a state observer based on a dynamic model of the battery and on the capacity estimation obtained with the reinforcement learning technique. The reward is defined as the error between the estimated and measured battery voltage. The effectiveness of the proposed solution is validated by considering different C-rates battery charging.
|
|
16:50-17:10, Paper ThB9.5 | |
Comparison of Control Strategies for HCPV Sun Tracking |
|
Garrido Satue, Manuel | Universidad De Sevilla |
Castaño, Fernando | University of Seville |
Ortega, M. G. | Universidad De Sevilla |
Rubio, Francisco R. | Universidad De Sevilla |
Keywords: Energy systems, Electrical power systems, Power plants
Abstract: This paper presents a comparison between five control strategies for HCPV sun tracking that comprise open loop by means of Solar Equations, closed loop by Sun position feedback (using an electro-optical sensor and a DC power sensor), and pointing ahead of the Sun. The performance of the control strategies is tested under a fully calibrated suntracker conditions and not-calibrated conditions to expose the problematic of the high accuracy sun pointing. When the suntracker is calibrated, all control strategies perform in a very similar way in terms of efficiency, with the exception that one of the control strategies does not need prior calibration. The measured loss in efficiency when the sun-tracker is not calibrated precisely is approximately equal to 6%.
|
|
17:10-17:30, Paper ThB9.6 | |
MPC for Collaborative Heat Transfer in a District Heating Network with Distributed Renewable Energy Generation and Storage |
|
Frison, Lilli | Fraunhofer Institute for Solar Energy Systems ISE |
Oliva, Axel | Fraunhofer Institute for Solar Energy Systems ISE |
Herkel, Sebastian | Fraunhofer Institute for Solar Energy Systems ISE |
Keywords: Energy systems
Abstract: District heating networks are ideally suited to include a high share of renewable energy sources into the heat production in urban areas with limited space. A powerful concept is the decentralized heat and electricity production where each household produces and stores heat, e.g., with solar thermal plants on the rooftops. To exploit the full potential, the households which have an excess of heat should be able to transfer heat to other households that have heat demand. We formulate the problem of a cooperative heat exchange and optimized district heating operation into a nonlinear MPC problem with binary decision variables that determine the heat exchange between the households and the central district heating plant. The resulting mixed-integer nonlinear problem is solved with a gradient-based optimization algorithm in combination with a suitable integer approximation scheme.
|
|
ThB10 |
|
Aerospace Systems |
Regular Session |
Chair: Alwi, Halim | University of Exeter |
Co-Chair: Yang, Hao | University of Leicester |
|
15:30-15:50, Paper ThB10.1 | |
Orbit Keeping about the Martian Moons with a Robust Path Following Control |
|
Rodolfo Batista Negri, Rodolfo | National Institute for Space Research |
Prado, Antonio | INPE |
Keywords: Aerospace, Sliding mode control, Autonomous systems
Abstract: In the last years, a considerable number of missions were proposed to study the Martian moons Phobos and Deimos. These satellites are known to have a highly perturbed environment, due to high-order terms of the moons' gravity fields, and mainly to non-negligible third-body effects from Mars. This work intends to address an important part of missions to such bodies that is the orbital maintenance. A feedback control for orbit keeping can decrease the mission operational cost, make the spacecraft readily respond to changes in the environment and increase the science outcome by allowing different orbital configurations and a more audacious operation. However, under a practical perspective, this is not a trivial task, as this orbit keeping law should be robust, accommodate idle-thrusters periods, and applicable to any orbital configuration. In order to accomplish these goals, we apply a robust path following control law recently proposed in the literature. This control law is derived under the frame of the sliding mode control theory, with a novel set of sliding surfaces specially designed for the problem, to robustly cope with bounded unknown disturbances. Because it is a path following law, the spacecraft can safely accommodate long term idle-thrusters periods, allowing it to make scientific measurements with no thrust interference and to reduce the fuel expenditure. We prove the efficacy of our path following law for different orbital configurations about Phobos and Deimos, considering a restricted three-body problem composed of the moon, Mars and the spacecraft. We also consider a polyhedron model to simulate higher order terms of the moons' gravity field. Our results indicate that the proposed autonomous orbit keeping is a reliable and efficient tool for Deimos exploration. In the case of Phobos, because of the large deviation from a perturbed Keplerian orbit due to third-body effects, a relative large amount of fuel is expected to be necessary to maintain a Kepler
|
|
15:50-16:10, Paper ThB10.2 | |
Flight Control Design for a Validated XV-15 Tilt Rotor Model: H-Infinity vs Linear Parameter Varying |
|
Yang, Hao | University of Leicester |
Morales, Rafael Mauricio | University of Leicester |
Keywords: Aerospace, Robust control, Linear parameter-varying systems
Abstract: This manuscript develops full envelope flight control systems for a validated tilt rotor vehicle by using 2 different methods: H-Infinity and Linear Parameter Varying. The former one leads to a single controller operating at a wide range of flight conditions, while the latter one results in a gain-scheduled controller, whose dynamics is varying and parameter dependent. Simulation results on the trimmed vehicle model show that both methods perform very well. Moreover, the comparison of the 2 strategies are carried out at several test points, illustrating pros and cons for each method.
|
|
16:10-16:30, Paper ThB10.3 | |
Energy Optimal Attitude Control for a Solar-Powered Spacecraft |
|
Kristiansen, Bjørn Andreas | NTNU Norwegian University of Science and Technology |
Gravdahl, Jan Tommy | Norwegian Univ. of Science & Tech |
Johansen, Tor Arne | Norweigian Univ. of Sci. & Tech |
Keywords: Aerospace, Optimal control, Optimization
Abstract: In this article we aim to maximize the net energy a solar-powered spacecraft gains when performing a maneuver. The net energy can be defined as the integral of the power supplied by the solar panels minus the power used by the attitude control system, and is important since energy is a scarce resource in space. Previous research on optimal attitude control has focused on optimization with respect to other costs, such as time-optimal control and optimal attitude control with respect to the integral of the square of the input. The energy flow depends on both the power spent on actuation and the power received from the solar panels. Thus, the optimal attitude control problem should be formulated in such a way that the attitude of the spacecraft relative to the Sun during the maneuver is included in the calculations. This paper proposes a cost function based on net power to address this problem, introducing a new cost function that incorporates the incoming energy from the solar irradiance and the outgoing energy due to actuation. A simulation study comparing an optimal control solution of the proposed net power cost function using IPOPT in CasADi is presented for a 6U CubeSat equipped with solar cell arrays, where the net power based optimal control maneuver is shown to compare favorably to a sun-pointing PD controller.
|
|
16:30-16:50, Paper ThB10.4 | |
An Interpolated Model Recovery Anti-Windup for a Canard-Guided Projectile Subject to Uncertainties |
|
Thai, Sovanna | Institut Franco-Allemand De Recherches De Saint-Louis (ISL) |
Roos, Clément | ONERA |
BIANNIC, Jean-Marc | ONERA |
THEODOULIS, Spilios | French-German Research Institute of Saint-Louis (ISL) |
Keywords: Aerospace, Uncertain systems, Robust control
Abstract: This paper presents an autopilot design for a dual-spin guided projectile subject to aerodynamic uncertainties and actuator saturations. The proposed design consists of a gain-scheduled baseline controller together with an interpolated dynamic anti-windup compensator based on model recovery. Evaluation of the closed-loop is done through IQC-based analysis at an operating point, and through Monte Carlo simulations. Both assessments show that the addition of an anti-windup compensator can drastically improve the behaviour of the system in degraded flight conditions.
|
|
16:50-17:10, Paper ThB10.5 | |
Fault Tolerant Control of a Quadplane UAV Using Sliding Modes |
|
Mizrak, Ibrahim | University of Exeter |
Alwi, Halim | University of Exeter |
Edwards, Christopher | University of Exeter |
Keywords: Aerospace, Fault tolerant systems, Sliding mode control
Abstract: This paper presents a FTC scheme for a quadplane using sliding mode control allocation. The scheme presented in this paper takes full advantage of the redundant vertical rotors to handle total actuator failures, especially during forward flight. The scheme utilises a combination of sliding modes and control allocation which requires only a single baseline control to be designed that operates in both fault free and fault/failure conditions. The scheme exploits the robustness of sliding modes to deal with actuator faults and control allocation to redistribute the control signals in the event of total actuator failures. Simulation results on a nonlinear model of a quadplane shows the efficacy of the scheme.
|
|
17:10-17:30, Paper ThB10.6 | |
An Automated Approach for the Design of a Fault Tolerant Controller |
|
Vile, Liam | University of Exeter |
Alwi, Halim | University of Exeter |
Edwards, Christopher | University of Exeter |
Keywords: Aerospace, Fault tolerant systems, Sliding mode control
Abstract: In this paper a fault tolerant sliding mode control allocation law is proposed. The controller is designed through a particle swarm optimization method which ensures the closed loop system is optimal for multiple design objectives - including robustness to uncertainties in the actuator fault and failure information. The approach is tested for the control of blended wing body aircraft’s lateral dynamics.
|
|
ThB11 |
|
Autonomous Systems |
Regular Session |
Chair: Del Re, Luigi | Johannes Kepler University Linz |
Co-Chair: Sjoberg, Jonas E. | Chalmers Univ. of Techn |
|
15:30-15:50, Paper ThB11.1 | |
Probabilistic Inverse Velocity Obstacle for Free Flying Quadrotors |
|
Khare, Ishaan | IIIT-Hyderabad |
Poonganam, Jyotish | IIIT-Hyderabad |
GOPALAKRISHNAN, BHARATH | IIIT HYDERABAD |
Krishna, Madhava | IIIT-Hyderabad |
Keywords: Adaptive control
Abstract: In this paper, we explore Probabilistic Inverse Velocity Obstacle (PIVO) as an alternative to probabilistic versions of Velocity Obstacles (PVO) for free flying quadrotor systems. Inverse Velocity Obstacles compute effective controls from a sequence of observations on other agents without the need to access ego state information. As a direct consequence of this the ego state noise is not entailed in probabilistic formulations bringing in verifiable advantages in the form of reduced path lengths, less conservative manoeuvres, reduced occurrences of stopping/hovering to let others pass. These advantages are vividly tabulated in this paper, showcasing the efficacy of PIVO as an alternative to probabilistic versions of Velocity Obstacles. In particular we show the benefits of PIVO over PVO in relation to sample complexity as well as overall trajectory lengths. We also show the efficacy of our probabilistic formulation in handling non-parametric and often multimodal noise distributions.
|
|
15:50-16:10, Paper ThB11.2 | |
A Two-Layer Switching Based Trajectory Prediction Method |
|
Reisinger, Stefan | Johannes Kepler University Linz |
Adelberger, Daniel | Johannes Kepler University Linz |
Del Re, Luigi | Johannes Kepler University Linz |
Keywords: Autonomous systems, Optimization, Modeling
Abstract: Safety-critical situations in road traffic often result from incorrect estimation of the future behavior of other road users. Therefore, many Advanced Driver Assistance Systems (ADAS) need prediction models to ensure safety. Physical prediction models offer the advantage of general use and work quite well for short prediction horizons, while for longer periods of time, maneuver based models offer better performance which, however, strongly depends on the data used to train them. An additional challenge for prediction is the fact that the surrounding traffic can change its path, i.e. for safety not only one maneuver should be considered but regular updates are required. Against this background, we propose a method that uses three physics-based predictions – corresponding to different prediction assumptions and models – combined with possible maneuver-based trajectories derived from environmental knowledge. Continuous monitoring is used to select the most likely of the three physics-based models. This choice then influences the environment-based prediction and the output of both models is fused afterwards. The output of the resulting Multiple Model Trajectory Prediction (MMTP) has been validated with measured data from two different scenarios – a city junction and a highway – with a good prediction performance and without the need for special measurements as commonly required for maneuver-based prediction.
|
|
16:10-16:30, Paper ThB11.3 | |
Risk-Sensitive Motion Planning Using Entropic Value-At-Risk |
|
Dixit, Anushri | California Institute of Technology |
Ahmadi, Mohamadreza | Caltech |
Burdick, Joel W. | California Inst. of Tech |
Keywords: Stochastic control, Predictive control for linear systems, Robotics
Abstract: We consider the problem of risk-sensitive motion planning in the presence of randomly moving obstacles. To this end, we adopt a model predictive control (MPC) scheme and pose the obstacle avoidance constraint in the MPC problem as a distributionally robust constraint with a KL divergence ambiguity set. This constraint is the dual representation of the Entropic Value-at-Risk (EVaR). Building upon this viewpoint, we propose an algorithm to follow waypoints and discuss its feasibility and completion in finite time. We compare the policies obtained using EVaR with those obtained using another common coherent risk measure, Conditional Value-at-Risk (CVaR), via numerical experiments for a 2D system. We also implement the waypoint following algorithm on a 3D quadcopter simulation.
|
|
16:30-16:50, Paper ThB11.4 | |
Depth-First Coupled Sensor Configuration and Path-Planning in Unknown Static Environments |
|
St. Laurent, Chase | Worcester Polytechnic Institute |
Cowlagi, Raghvendra V. | Worcester Polytechnic Institute |
Keywords: Autonomous systems, Sensor and mesh networks, Machine learning
Abstract: We address path-planning for a mobile agent in an unknown static environment. The environment is observed by a sensor network where each sensor has a configurable location and field of view. We propose a depth-first coupled sensor configuration and path-planning (DF-CSCP) iterative method, which iteratively finds an optimal sensor configuration (location and FoV), applies Gaussian Process Regression to construct a threat field estimate, and then finds a candidate optimal path with minimum expected threat exposure. The DF-CSCP method uses a two stage procedure, (1) Explore and (2) Exploit, to drive the uncertainty of the candidate path cost variance below a prespecified threshold. To maintain tractability of GPR with increasing number of measurements, we present a sparse-update scheme. The proposed method relies on novel task-driven information gain (TDIG) metrics, the maximization of which provides sensor configurations. The TDIG metric quantifies the importance of acquiring sensor data of highest relevance to the path-planning task. Through numerical studies, we demonstrate the technical results that the DF-CSCP algorithm finds near-optimal paths with significantly fewer sensor measurements compared to traditional information-maximization methods.
|
|
16:50-17:10, Paper ThB11.5 | |
Resources to Support a First Course in Feedback, Dynamics and Control |
|
Rossiter, J. Anthony | University of Sheffield |
Keywords: Control education, Computer aided learning, Control courses and labs
Abstract: A recent survey of the global control community [13-15] identified views on the most important topics that should be covered when engineering students take just a single control related course. The next requirement for staff is to have efficient ways of locating suitable resources to support the teaching of such a course. Due to the lack, for now, of an effective international database for such resources, the author has produced his own website to capture the most important topics, including some online quizzes, tutorial sheets, videos, PDF notes and videos. This paper gives an overview of the resource as that may be useful to colleagues and doubles as an invite to the community to work together to form and systematically organise generic free resources available to all.
|
|
17:10-17:30, Paper ThB11.6 | |
Road Boundary Modeling for Autonomous Bus Docking Subject to Rectangular Geometry Constraints |
|
Elawad, Amal | Chalmers University of Technology |
Murgovski, Nikolce | Chalmers University of Technology |
Jonasson, Mats | Chalmers |
Sjoberg, Jonas E. | Chalmers Univ. of Techn |
Keywords: Autonomous systems, Optimal control, Optimization
Abstract: This paper studies the optimization problem of autonomous bus parallel parking subjected to rectangular geometry constraints. The parking space is a non-smooth and non-convex irregular polygon. We propose a novel method for modeling geometry constraints, which allows using the exact non-smooth parking space, while still being able to formulate the problem as a smooth nonlinear program. The focus of this paper is to compare our method to two other approaches of modeling rectangular geometry constraints, where one uses mixed-integers while the other approximates the rectangle as a smooth function. We show that our novel method prevents collisions and requires shorter computation time, where collision constraints are imposed on a limited number of points on the vehicle contour.
|
|
ThB12 |
|
Process Control |
Regular Session |
Chair: Tanaskovic, Marko | Singidunum University |
Co-Chair: Beilkin-Sirota, Lea | Tel Aviv University, School of Mechanical Engineering |
|
15:30-15:50, Paper ThB12.1 | |
A Causal Model-Based Planner for the Reconfiguration of Continuous Processes |
|
Reinartz, Christopher | Technical Univeristy of Denmark |
Enevoldsen, Thomas Thuesen | Technical University of Denmark 228487 |
Galeazzi, Roberto | Technical University of Denmark |
Ravn, Ole | Technical University of Denmark |
Keywords: System reconfiguration, Supervisory control, Chemical process control
Abstract: This paper presents a planning framework for discrete planning of human operations in highly automated industrial plants, where manual and automatic control coexist. Based solely on qualitative knowledge of the controlled continuous process represented through signed directed graphs, the planner exploits a greedy algorithm to determine the optimal action for the human operator upon changes in the process' operating conditions or reconfiguration of the control system due to faulty conditions. The planner results in a decision support tool that instructs the human operator on the best course of action. The planning framework and the resulting planner are demonstrated on a quadruple-tank system.
|
|
15:50-16:10, Paper ThB12.2 | |
Gaussian Processes for Improved Dynamic Modeling in the Predictive Control of an Arduino Temperature Control Lab |
|
Sadik, Idris | KU Leuven |
Kueper, Armin | KU Leuven |
Waldherr, Steffen | KU Leuven |
Keywords: Process control, Machine learning, Nonlinear system identification
Abstract: A practical application of Gaussian processes (GPs) as an alternative nonlinear system identification approach in model predictive control (MPC) is presented. By means of using an Arduino Temperature Control Lab, setpoint tracking accuracy for a Gaussian process-based MPC scheme is compared to state space MPC and a proportional-integral-derivative (PID) controller. Foregoing parameterized system identification, GPs are proven to offer superior accuracy, while eliminating the tedious developments plaguing first-principle nonlinear alternatives. By further utilizing GPs probabilistic framework, estimates for variance are interpreted as system-specific uncertainty and used to better select control solutions that remain in training regions. Including variance within the optimal control problem (as opposed to its exclusion), improved overall setpoint tracking and affords a more cautious controller.
|
|
16:10-16:30, Paper ThB12.3 | |
Real-Time Controlled Acoustic Metamaterials Imitating Quantum Wave Phenomena |
|
Beilkin-Sirota, Lea | Tel Aviv University, School of Mechanical Engineering |
Keywords: Complex systems
Abstract: In this work I demonstrate how acoustic and mechanical metamaterials that are based on embedded active feedback mechanism, can realize unique waveguding inspired by quantum-mechanical phenomena. With the embedded feedback design, the couplings between the metamaterial unit cells and the consequent dynamical properties are determined in a real-time closed loop operation, and can be reprogrammed at will by the user. In particular, a controller can create structural couplings that are impossible to create with fixed elements, either passive or active. I present two examples. The first is a realization of unidirectional, topologically-protected waveguiding along the boundary of a mechanical metamaterial constrained to out-of-plane vibration. The controller achieves this by creating non-reciprocal couplings that violate Newton's third law. The second is emulation of quantum tunneling in an effectively continuous acoustic metamaterial. The controller mimics the tunneling in real-time by creating effectively negative constitutive parameters of a specific form, resulting in negative refraction of waves unimpeded in the normal direction.
|
|
16:30-16:50, Paper ThB12.4 | |
From Alarm System Events towards Quality Inspection of the Final Product: Application to a Semiconductor Industry |
|
AL-KHARAZ, Mohammed | Laboratoire d'Informatique Et Systèmes - Aix Marseille Universit |
Ananou, Bouchra | LSIS |
Ouladsine, Mustapha | Université D'aix Marseille III |
combal, Michel | ST-Microelectronics |
PINATON, JACQUES | STMicroelectronics |
Keywords: Process control, Manufacturing processes
Abstract: Process diagnostic and monitoring during production is a fundamental task of the control and alarm system. However, many defected products are still related to various issues of health states of production equipment. Therefore, quality inspection is a crucial step during the manufacturing process, ensuring that a product's quality is maintained or improved with a reduced or total absence of errors. The final product quality determines whether or not a product unit satisfies its intended use. In this paper, we propose a final quality inspection framework based on alarm events data. In this framework, we first transform the textual alarm data into numeric using binary scoring. Then, we reduce the dimension of the obtained numeric matrix using an appropriate alarms grouping method. After that, we apply the reduced data to learn a classifier and to make a decision. Finally, we compare several machine learning algorithms' performance in the prediction of scrap-per-lot, namely Decision Tree, Logistic Regression, K-nearest neighbors, Linear Support Vector Machine, and Multi-Layer Perceptron. The results show a satisfactory performance of the compared models that we effectively prove on a dataset collected over the whole semiconductor fabrication facility.
|
|
16:50-17:10, Paper ThB12.5 | |
Advanced Hierarchical Predictive Routing Control of a Smart De-Manufacturing Plant |
|
Boffadossi, Roberto | Politecnico Di Milano |
Fagiano, Lorenzo | Politecnico Di Milano |
Cataldo, Andrea | Cnr - Stiima |
Tanaskovic, Marko | Singidunum University |
Lauricella, Marco | Politecnico Di Milano |
Keywords: Manufacturing processes, Predictive control for nonlinear systems
Abstract: The application of a novel approach to the routing control problem of a real de-manufacturing plant is presented. Named Hierarchical Predictive Routing Control (HPRC) and recently proposed in the literature, the approach deals with large number of integer inputs and complex temporal-logic constraints by adopting a low-level path-following strategy and a high-level predictive path allocation. Several improvements are presented, including a novel search tree exploration method, lockout detection routines, and plant-specific handling constraints. Simulation results show very good performance and small computational times even with high number of pallets and long prediction horizon values.
|
|
17:10-17:30, Paper ThB12.6 | |
Output Maneuvering for Cartesian 3D Printer |
|
Moltumyr, Andreas | Norwegian University of Science and Technology |
Arbo, Mathias Hauan | Norwegian University of Science and Technology |
Gravdahl, Jan Tommy | Norwegian Univ. of Science & Tech |
Keywords: Manufacturing processes, Robotics
Abstract: 3D printing, also known as additive manufacturing, is a production technique that can create highly customized parts and is therefore ideal for product prototyping and customized orders. An important aspect of 3D printer systems is the ability to accurately and precisely move the extruder along a planned path, ensuring the production of parts with low dimensional error. In this paper, output maneuvering is considered for the purpose of steering the extruder of a Cartesian 3D printer along a desired path. As slicing software provides waypoints with minimal change in angle between the line segments, a novel speed profile adjustment is introduced which prioritizes maintaining the current along-path speed when the angle between line segments is sufficiently low. Through a design example, a nonlinear maneuvering controller consisting of a geometric and a dynamic task is deduced. Positive and negative aspects of applying output maneuvering to additive manufacturing are discussed.
|
|
ThTS |
|
Stability and Robust Control of PDEs and Large Scale Networks |
Tutorial Session |
Chair: Mironchenko, Andrii | University of Passau |
Co-Chair: Prieur, Christophe | CNRS |
|
15:30-17:30, Paper ThTS.1 | |
Stability and Robust Control of PDEs and Large Scale Networks (I) |
|
Mironchenko, Andrii | University of Passau |
Prieur, Christophe | CNRS |
Keywords: Large-scale systems, Control over networks
Abstract: In this tutorial, we introduce to a broad audience key concepts, results and applications of the infinite-dimensional stability theory, with a particular focus on input-to-state stability and robustness analysis. The scope of techniques that we discuss, includes Lyapunov functions, nonlinear systems theory, semigroup theory, spectral decompositions and boundary control. We discuss the applications of these methods to robust stability of boundary control systems, robust control of partial differential equations and to the stability of large-scale and infinite networks. The tutorial consists of 4 talks: Stability analysis of large-scale and infinite networks (Andrii Mironchenko) Input-to-state stability of boundary control systems (Andrii Mironchenko) Pole shifting theorem for parabolic systems (Christophe Prier) Stabilization of unstable parabolic systems via saturated controls (Christophe Prieur)
|
| |