Keywords: Intelligent control systems, Robotics, Neural networks
Abstract: Imitation Learning holds significant promise in enabling the automation of complex robotic manipulations tasks which are impossible to explicitly program. Mushroom harvesting is a task of high difficulty requiring weeks of intense training even for humans to master. In this work we present an end-to-end Imitation Learning pipeline that learns to apply the series of motions, namely reaching, grasping, twisting, and pulling the mushroom directly from pixel-level information. Mushroom harvesting experiments are carried out within a simulated environment that models the mushroom dynamics based on von Mises yielding theory with parameters obtained through expert picker demonstration wearing gloves with force sensors. We test the robustness of our technique by performing randomizations on the camera extrinsic and intrinsic parameters as well as on the mushroom sizes. We also evaluate on different kinds of visual input namely grayscale and depth maps. Overall, our technique shows significant promise in automating mushroom harvesting directly from visual input while being remarkably lean in terms of computation intensity. Our models can be trained on a standard Laptop GPU in under one hour while inference of an action takes less than 1.5ms on a Laptop CPU. A brief overview of our experiments in video format is available at: https://bit.ly/41kCH7T

Keywords: Optimisation, Multi-agent systems, Robotics
Abstract: Employing teams of robots to offer services to human operators enables the latter to reduce their physical workload. In this paper, we focus on the problem of optimally allocating and scheduling the robot tasks in order to serve human operators. We formulate a Mixed-Integer Linear Programming problem which aims to minimize the human waiting time and the energy spent by the robots, while ensuring that any velocity constraints of the robots are fulfilled and the task ordering is correct. In addition, we propose an online re-allocation strategy that takes into account the possibility of changing human parameters over time. This strategy determines whether a new optimal solution must be computed. We validate the proposed framework in a simulated precision agriculture setting composed of two robots and four human operators for a harvesting application.

Keywords: Robotics, Assistive technology, Robust control
Abstract: This paper devises a framework to control a quadruped robot tethered to a visually impaired person. The whole-body control of the quadruped robot does not exploit any force sensor. It makes use of two observers: the former for the estimation of the wrench applied on the robot's centre of mass, which is in turn used to handle the human-robot estimation; the latter for the estimation of the external forces acting on the legs to guarantee a stable balance on irregular terrains. Besides, an admittance filter is employed to guarantee a safe human-robot interaction. A supervisor is designed and placed side by side with the quadruped whole-body control to understand human needs and handle lifelike situations. The validity of the approach is tested in a realistic simulation environment.

Keywords: Robotics, Nonlinear control, Hybrid systems
Abstract: A new approach for running by exploiting passive elastic dynamics is addressed in this paper. A control method based on energy and angular momentum regulation is proposed. The controller exploits invariance properties of energy and angular momentum to achieve stabilization of passive gaits with almost zero control effort after convergence. The passive gaits are derived based on a modified version of the well-known SLIP model, which takes into account the pitch dynamics during the flight phase, which is otherwise disregarded by the conventional SLIP model. Based on this model, the interesting phenomenon of running in presence of persistent somersaulting is investigated and persistently-somersaulting running gaits are identified and analyzed.

Numerical simulations validate the method and confirm the effectiveness in ideal conditions.

Keywords: Robotics, Predictive control, Optimisation
Abstract: We introduce optimization-based algorithms that address the problem of robot task scheduling and trajectory planning. Following a two-layer hierarchical control structure, we first decouple the task scheduling from the trajectory planning by introducing two separate optimization problems, a discrete one for task scheduling and a continuous optimization problem for trajectory planning. In a further step, we integrate both planning layers into a monolithic layer in the form of a mixed-integer recursive optimization problem. The algorithms are implemented and validated using Robot Operating System (ROS) on an experimental setup with a robot manipulator performing pick-and-place tasks.

Keywords: System identification, Neural networks, Robotics
Abstract: Flexible manipulators are the core technology to the development of collaborative robotic systems, which is a trend in Industrial Robotics. Hence, the system and parameter identification of these systems is important to develop more accurate strategies of control based on the known dynamic behavior of these joints. The purpose of this paper is to build linear and nonlinear auto-regressive models with exogenous inputs based on experimental data. The data set used was collected with an experimental setup using a multi-sine torque input signal applied to the motor, and the output was the angular velocity of the link. The models employed were the ARX, ARMAX, NARMAX, and NARX-NN, and their performance was measured using the root-mean-squared error and the R^2 score. The results show that all the models presented good performance, with an acceptable R^2 score and an RMSE value close to each other.

Keywords: Energy efficient systems, Modelling and simulation, Neural networks
Abstract: Accurately estimating and forecasting building occupancy represents an important tasks for higher level indoor energy management and control routines. Extended availability of public and open datasets reflecting indoor conditions through various sensor measurement and indirect proxies of human activity enable reliable benchmarking of new techniques for pre-processing and learning of occupancy patterns. In this work we present a comparative study between deep learning, such as convolutional neural networks, and conventional machine learning approaches, such as decision trees and random forests, on an a reference occupancy dataset. The various design deci- sion and parametrisation options are discussed. The building occupancy classification task involves generating model outputs for various discrete occupancy categories. Standardised metrics such as accuracy, precision, recall and the F1-score are used for replicable benchmarking of the results. Main finding of the study is that, though generally the deep learning methods offer better overall results, the addition of relevant features (sensors) to the input dataset can yield better results for the conventional machine learning models with significantly lower training time and model size. This results in suitable, fast-inference, models for embedded deployment in physical proximity to the process.

Keywords: Genetic and evolutionary computation, Nonlinear systems, System identification
Abstract: This paper is addressed to a problem of state observation for permanent magnet synchronous motor (PMSM) and its design parameter tuning via evolutionary algorithm. Recently proposed flux, position and speed observer that is based on nonlinear parameterization of motor model and dynamic regressor extension and mixing (DREM) technique is considered. Though global asymptotic convergence of this observer was guaranteed for all positive real values of several design parameters the choice of their values for a particular motor was not well considered. To overcome this drawback a genetic algorithm is used to perform automatic tuning of required coefficients minimizing cost function that is associated with estimation errors. Simulation results supplemented by verification demonstrate the efficiency of the proposed approach resulting in a set of easy-to-implement-in-practice values of design parameters.

Keywords: Multi-agent systems, Networked systems, Switching systems
Abstract: Hegselmann--Krause (HK) models exhibit complex behaviors which are not easily tractable through mathematical analysis. In this paper, a characterization of the steady-state behaviors of homogeneous HK models and sensitivity to confidence thresholds is discussed by commenting on existing and new numerical results. The typical decreasing of number of clusters and convergence time by increasing the confidence thresholds are discussed and motivations for the behavior of some counterexamples are provided. A tighter upper bound for the dependence of the number of clusters with respect to the confidence thresholds is proposed. Differences and analogies between the opinions' evolution for symmetric and asymmetric HK models are commented.

Keywords: Computational intelligence, Distributed systems, Neural networks
Abstract: This paper presents the ability of the federated learning concept to create a collaboration between multiple devices using a shared global model, while still keeping data privacy to meet the General Data Protection Regulation (GDPR). In real-world application scenarios, this concept faces problems related to the defense of the global model from possible attacks and the compatibility with non-independent and identically distributed data (non-IID). This paper presents two aggregation algorithms compatible with non-IID data, which use a refined aggregation of the local model, based on their accuracy. Thus, the proposed algorithms can refine the confidence in each client, eliminate intruders and allow a safe aggregation of the global model. Testing scenarios performed for IID and non-IID data illustrate that the proposed algorithms are able to provide faster training and improved robustness against intruders, w.r.t. the well-known federated average algorithm.

Keywords: Neural networks, Wireless sensor networks, Soft computing
Abstract: The application of Artificial Intelligence (AI) and Machine Learning (ML) in IoT smart sensor technologies has opened wide possibilities in the field of Water Quality Monitoring (WQM). Power saving and price-per-unit requirements, fundamentals for Wide Distributed Sensors Networks (WDSN), drive research in developing AI-model reduction techniques to make algorithms faster and cheaper in terms of hardware resources and battery consumption. Before any optimization process, Feature Selection (FS) is needed to reduce the number of basic operations in smart sensors workflow, thus making lighter the data acquiring phase and decreasing the size of data input for the subsequent AI process. However, selecting the FS method that best fits the specific requirements of the considered application is not trivial, given the numerous available FS methods and the relevant number of possible feature subsets. In this context, this paper presents a generalized and versatile algorithm, based on the concept of ensemble-FS, to support and speed up the AI-unit design process. The method compares different FS methods, effectively providing precise information about the accuracy (and any other requirement) of the selected FS method with respect to the number of acquired features. The proposed methodology is tested on a real WQM case study by analyzing the obtained results when both the popular and high-performer XGBoost algorithm and some ready-to-use FS-ranker methods in the Waikato Environment for Knowledge Analysis (WEKA) are used. Results show that the XGboost is the best performer for the case study in terms of stability and accuracy.

Keywords: Soft computing, Computational intelligence
Abstract: Practical implementation of data analytics in industrial environments has always been a problematic area because of data availability and quality. In this paper, a Gaussian sampling methodology is proposed to address the problem of imbalanced telemetry datasets that is one of the root causes that make modelling less reliable. By generating subsets that achieve homogeneous density distributions this problem is addressed. By comparing the impact of this method with the baseline case of random sampling, this paper aims to address this problem and propose a practical solution. A case study based on an industrial cooling device is used to assess and illustrate the proposed approach.

Keywords: Adaptive control, Unmanned systems, Robust control
Abstract: This paper deals with the constrained position and angle tracking control design for quadrotor under unknown upper bound disturbances. An adaptive integral sliding mode control (AISMC) is proposed to perform the position and angle tracking for the quadrotor subject the severe disturbances and input saturation constraints. The proposed approach that does not require a priori knowledge of disturbance boundaries, allows through an adaptation dynamic law to reduce the computing effort, to obtain a good tracking, and to avoid an overestimation of the gain of the mode sliding that will automatically handle input saturation constraints. Stability and convergence in finite time are proved by Lyapunov theory. The efficiency of the proposed method is shown by simulation.

Keywords: Nonlinear systems, Robust control, Feedback stabilization
Abstract: This article discusses the sliding mode control problem, where the reaching phase is achieved non-monotonically, and the sliding phase can be achieved either monotonically or non-monotonically. Once the reaching phase is completed, the state variables slide on the sliding manifold and then reach the equilibrium point. A practical second-order example of the ball motion model is considered to show the non-monotonic reaching phase. Simulation results verify the non-monotonic behavior of the reaching phase.

Keywords: Neural networks, Robust control, Intelligent control systems
Abstract: In this paper, we propose a novel neural network based state constrained integral sliding mode (NN-SCISM) control algorithm for nonlinear system with partially unknown dynamics in presence of time-varying constraints. In particular, the drift term characterizing the system dynamics is estimated by using a two-layer neural network, whose weights are adjusted according to adaptation laws designed relying on stability analysis. Thanks to a sliding variable which varies depending on the minimum distance between the system state and the current closest constraint, the control algorithm is able to drive the system state to a desired target state, while avoiding the forbidden states contained in the time-varying set delimited by the constraints. The proposal has been theoretical analysed and assessed in simulation.

Keywords: Robust control, Disturbance rejection
Abstract: In this paper, the mixed Finite-Time Stability (FTS)/mathcal H_infty control problem is investigated for the class of nonlinear quadratic systems (NLQSs), which have several relevant applications, e.g., in robotics, systems biology and other domains of applied sciences. Sufficient conditions are provided here to solve synthesis problems, in the presence of both norm-bounded disturbances, constraints on initial and terminal conditions, and finite-time bounds on the output transient. More specifically, taking into account such constraints within the design phase, allows to achieve a desired mathcal H_infty performance with nonzero initial conditions, while simultaneously guaranteeing that a given NLQS is finite-time stable for all admissible uncertainties and disturbances. Such conditions can be formulated as Linear Matrix Inequalities (LMIs) optimization problem. The applicability of the proposed results is illustrated by means of a numerical example.

Keywords: Robust control, Linear systems
Abstract: An approach to the solution of a robust sparse filtering problem via use of a reduced number of outputs under arbitrary bounded external disturbances and norm-bounded system uncertainties using an observer is proposed. The approach is based on the LMI technique and the method of invariant ellipsoids, and made it possible to reduce the initial problem to parameterized semidefinite programming that can be easily solved numerically. Two ways to control sparsity are proposed: controlled relaxation approach and Pareto frontier approach.

Keywords: Robust control, Networked systems, Optimisation
Abstract: In this paper, a discrete time-varying model of sensors network is considered. The external input belongs to the class of sequences of random vectors with bounded anisotropy of the extended vector. The anisotropy-based analysis of the system includes the analysis for the multiplicative noise systems and the boundedness criterion of the anisotropic norm. The considering problem concerns the selection of the estimator, which one guarantees the boundedness anisotropic norm. It is demonstrated how to reduce considering problem to convex optimization one.

Keywords: Robust control, Optimisation, Linear systems
Abstract: The problem of H_infty state estimation is considered for uncertain polytopic linear discrete-time stochastic state-multiplicative systems. We first bring the a unique version of the BRL for the latter systems which allows for vertex-dependent solution in the uncertain case. Following the BRL derivation, we solve the estimation problem for nominal systems which serve as a basis for the extracting the filter parameters in the uncertain case. In both cases: the nominal and the uncertain cases, the filter parameters are extracted by a solving an LMI condition in the former case or a set of LMIs in the latter case, both of which depend on a minimal set of tuning parameters, thus greatly reduce the over-design. The theory presented is demonstrated by a numerical example.

Keywords: Guidance
Abstract: In stochastic interception scenarios, an intercepting missile only has uncertain information about the target state, as this information is obtained from noisy measurements. The true dynamics of the target are also unavailable to the intercepting missile, so, instead, the interceptor can assume that the target possesses ideal dynamics, which amounts to adopting the worst-case scenario. Moreover, even when linear models and Gaussian noises are assumed, the notorious curse of dimensionality renders the straightforward optimal solution to this problem intractable in real-time. To alleviate the computational burden, this work uses an approach based on the notion of terminal sets to present an optimal interception strategy for stochastic scenarios. We show that using this approach greatly reduces the computational effort, as the number of modes diverges quadratically in time instead of exponentially. Another computational burden reduction is achieved via a novel decomposition of the interceptor’s terminal set. These results render the proposed strategy implementable in real-time, as the horizon is sufficiently short at the endgame stage of the engagement. A Monte Carlo simulation study is used to demonstrate the performance of the novel guidance law in stochastic scenarios, and to show that it achieves real-time performance despite its (still) considerable computational burden.

Keywords: Intelligent transportation systems, Distributed systems, Decentralized control
Abstract: Vehicle platooning is deemed a promising solution to improve traffic safety, reduce fuel consumption and increase traffic throughput and road capacity.Road throughput can be increased by driving at small inter-vehicle distances. Adjusting the spacing with both the preceding and the succeeding vehicles using bidirectional communication can potentially improve the platoon cohesiveness and is crucial when considering the synchronized merging scenario. In this paper, a distributed consensus controller is presented which relies on bidirectional platoon communication. Along with the consensus-based controller synthesis, platoon stability proof and string stability analysis, effectiveness of the controller is verified in a simulation study.

Keywords: Intelligent transportation systems, Distributed systems, Linear systems
Abstract: Vehicle platooning is deemed a promising solution to improve traffic safety, reduce fuel consumption and increase traffic throughput and road capacity. Road throughput can be increased by driving at small inter-vehicle distances. Adjusting the spacing with both the preceding and the succeeding vehicles using bidirectional communication can potentially improve the platoon cohesiveness and is crucial when considering the synchronized merging scenario. This paper presents vehicle controllers for cooperative adaptive cruise control (CACC) for heterogeneous vehicle platoons. To achieve desired inter-vehicle spacing with respect to a predecessor and a follower, a bidirectional communication scheme is employed. Moreover, conditions for vehicle stability are provided and string stability properties of the platoon analyzed. Finally, effectiveness of the controller is verified in a simulation study.

Keywords: Intelligent transportation systems, Multi-agent systems, Distributed systems
Abstract: Virtual Coupling has been included among the most relevant innovations to be studied in the European Horizon 2020 Shift2Rail Joint Undertaking for increasing the railway lines capacity through the dynamic connection of two or more trains to form a convoy while preserving safety. Within this framework, this paper addresses the virtual coupling control problem for heterogeneous nonlinear uncertain connected high-speed trains. Leveraging the Multi-Agent Systems framework, a novel distributed robust and adaptive PI-like control scheme is proposed to solve the control problem. In order to provide suitable adaptive mechanisms for the control gains, we exploit the Lyapunov theory and Barbalat's lemma and prove the asymptotic stability for the overall networked control system, as well as the boundedness of these signals. The virtual coupling objective is achieved in a fully-distributed fashion by limiting the amount of time-varying information necessary for the computation of the control action and, hence, saving communication channel bandwidth while reducing the computational burden. Exemplary numerical simulations are given to support the theoretical derivations and to prove the effectiveness of the proposed control strategy in a real driving scenario.

Keywords: Networked systems, Distributed systems, Multi-agent systems
Abstract: In epidemic networks, walk-based centrality indices are often used to identify the nodes that are significantly contributing to the spread of disease. While the network topology can provide a good insight into how the disease might propagate throughout the network, epidemic-related factors can change the ranking results as well. This paper presents a dynamics-based node centrality that incorporates epidemic characteristics, internal time delays, and network structure at the same time. This centrality allows for dynamic identification of the nodes that are more sensitive to external shocks, which in turn can help prevent performance degradation in the network. It is shown that some of the prominent walk-based centralities, such as local and eigenvector centralities, are in fact correlated with dynamics-based centrality for certain epidemic parameters.

Keywords: Swarms, Networked systems, Distributed systems
Abstract: In this paper, we investigate the performance deterioration of commensurate fractional-order consensus networks under exogenous stochastic disturbances. We formulate fractional-order differential equations for the network dynamics using Caputo derivatives and the Laplace transform, and employ the H_2 norm of the dynamical system as a performance measure. By developing a graph-theoretic methodology, we relate the structural specifications of the underlying graphs to the performance measure and explicitly quantify fundamental limits on the best achievable levels of performance in fractional-order consensus networks. We also establish new connections between the sparsity of the network and the performance measure, characterizing fundamental tradeoffs that reveal the interplay between the two. Finally, we provide numerical illustrations to verify our theoretical results, which could help in the design of robust fractional-order control systems in the presence of disturbances.

Keywords: Marine control, Predictive control, Nonlinear control
Abstract: Shipping goods using international waterways as well as inland waterways is one of the most important kind of transportation. With the increase of global networking, also the amount of goods exchange increases and has increased in the last decades. While sailing on open sea doesn't pose difficulties due to good assisting systems anymore, especially for low-speed maneuvering in confined waters exists a lack of reasonable control solutions. This is due high requirements in control quality and obstructive thereby non-linearities in ship dynamics and non-linear transformation from body- to earth-fixed frame. In this paper, a model predictive approach to control acting forces and moments for offset-free trajectory tracking is applied. To deal with non-linear behavior, successive linearization at every sampling interval is used. Furthermore, offset-freeness is guaranteed by reformulating the model as incremental system. The performance is validated simulatively and compared to a state of the art state controller. It turns out, that the model predictive approach outperforms the state controller in terms of control error and disturbance rejection.

Keywords: Marine control, Robotics, Modelling and simulation
Abstract: Manipulation tasks in underwater operations were so far only realizable with large work class remotely operated vehicles (ROV). However, new developments in the field of miniaturized fully-electric multi-joint manipulators are opening up applications for small and medium size ROVs. Despite miniaturization of the manipulators, the manipulator-to-vehicle ratio is not identical to that of large work class ROVs. Coupling effects between manipulator and vehicle must be taken into account to ensure good controller performance of the position and attitude of the vehicle and the tool center point (TCP) of the manipulator. In this paper, the modelling of an underwater manipulator based on an extended Newton-Euler method is presented. The proposed method is applied to a Reach Alpha 5 manipulator and validated by static measurements. Furthermore, a method for collision detection between arm parts is presented and applied to the Reach Alpha 5 for workspace analysis.

Keywords: Robotics, Predictive control, Marine control
Abstract: This paper presents the design of a robust tube-based nonlinear Model Predictive Control (MPC) law for a triangular marine platform, that is over-actuated with three rotating jets. The goal is safe navigation and dynamic positioning of the platform under realistic wind and wave environmental disturbances, as well as real-time obstacle avoidance employing Control Barrier Functions (CBF) as constraints in the robust MPC strategy. Extensive Monte Carlo simulations have been conducted under a control allocation scheme, taking into account the actuator thrust and rotation dynamics, sensor noise, as well as additional state and input constraints. The simulation results show that the nonlinear controller ensures robust and safe navigation with obstacle avoidance and accomplishes accurate positioning of the floating platform at a given goal pose, while satisfying the actuator limits.

Keywords: Robotics, Unmanned systems, Marine control
Abstract: This paper presents a novel control strategy for image-based visual servoing (IBVS) of underwater vehicle manipulator systems (UVMS) using control barrier functions (CBFs) to handle field of view (FoV) constraints and system's operational limitations such as manipulator joint limits and vehicle velocity performances. The proposed approach combines the advantages of IBVS, which provides visual feedback for control, with CBFs, which can formally enforce visibility and safety constraints on the UVMS's motion. A CBF-based control law is derived and integrated with the IBVS algorithm, which guarantees the satisfaction of FoV and system's operational constraints and ensure stability of the closed-loop system. To deal with FoV constraints, the proposed method uses a FoV index to estimate the degree of visibility of the scene, which is used to adjust the control inputs accordingly. The effectiveness of the proposed strategy is demonstrated through realistic simulation results, showing improved performance and safety of the UVMS under FoV and operational constraints compared to traditional IBVS methods. The results indicate that the proposed approach can handle the challenging underwater environment, UVMS dynamics and the operational constraints effectively, making it a valuable control strategy for practical applications of UVMS.

Keywords: Robotics
Abstract: Algorithms for autonomous navigation in environments without Global Navigation Satellite System (GNSS) coverage mainly rely on onboard perception systems. These systems commonly incorporate sensors like cameras and Light Detection and Rangings (LiDARs), the performance of which may degrade in the presence of aerosol particles. Thus, there is a need of fusing acquired data from these sensors with data from Radio Detection and Rangings (RADARs) which can penetrate through such particles. Overall, this will improve the performance of localization and collision avoidance algorithms under such environmental conditions. This paper introduces a multimodal dataset from the harsh and unstructured underground environment with aerosol particles. A detailed description of the onboard sensors and the environment, where the dataset is collected are presented to enable full evaluation of acquired data. Furthermore, the dataset contains synchronized raw data measurements from all onboard sensors in Robot Operating System (ROS) format to facilitate the evaluation of navigation, and localization algorithms in such environments. In contrast to the existing datasets, the focus of this paper is not only to capture both temporal and spatial data diversities but also to present the impact of harsh conditions on captured data. Therefore, to validate the dataset, a preliminary comparison of odometry from onboard LiDARs is presented.

Keywords: Intelligent control systems, Optimisation, Modelling and simulation
Abstract: This paper deals with the morning commute problem when two classes of commuters co-exist in an urban transportation network. While each school starts at the same time, traffic flow enters the network simultaneously, leading to the formation of a peak demand value that the network cannot fully accommodate. The task becomes more challenging when commuters head to their workplace after reaching their respective school. To tackle this issue, we propose the School Demand Allocation Paradigm (SDAP), which allows schools to have different starting times. Subsequently, we deploy an optimization framework, the target of which is to determine the optimal pair of the shifted school start time for each school that maintains operation under free-flow conditions. We utilize a macroscopic MFD-based traffic model, which can account for the coupling of the classes mentioned above. The effectiveness of the proposed approach is verified through macroscopic simulations.

Keywords: Networked systems, Decentralized control, Intelligent control systems
Abstract: We consider networked systems comprised of interconnected sets of linear subsystems and propose a decentralized and compositional approach to stabilize or dissipativate such linear networked systems via optimally modifying some existing interconnections and/or creating entirely new interconnections. We also extend this interconnection topology synthesis approach to ensure the ability to stabilize or dissipativate such networks under distributed (local) feedback control. To the best of the authors' knowledge, this is the first work that attempts to address the optimal interconnection topology synthesis problem for general linear networked systems. The proposed approach only involves solving a sequence of linear matrix inequality problems (one at each subsystem), and thus, it can be implemented efficiently and scalably in a decentralized and compositional manner. We also include a case study where the proposed interconnection topology synthesis approach is compared with an alternative dissipativity-based approach.

Keywords: Renewable energy and sustainability, Modelling and simulation, Intelligent control systems
Abstract: Photovoltaic-thermal (PV-T) systems are expected to fulfil an increasingly vital role in future energy production. The current research endeavors to showcase machine learning modeling and control of a water-based PV-T collector. In this work, the PV-T collector is modeled using a decision tree algorithm and artificial neural network (ANN). The predicted outputs are compared with the actual outputs to validate the models. The ANN-based model performed better and proved its efficacy in training and testing. Further, various control strategies are implemented and their performance is compared. All the techniques presented are illustrated through simulation results.

Keywords: Process control, Intelligent control systems, Mechatronic systems
Abstract: The current paper introduces a hybrid data-driven control algorithm obtained by the mix of two data-driven algorithms, namely intelligent proportional-integral controllers as representative Model-Free Control (MFC) algorithms, whose parameters are optimally tuned via Fictitious Reference Iterative Tuning (FRIT) algorithms using metaheuristic Slime Mould Algorithm. The purpose of the current mix is to combine the advantages of MFC and FRIT. The efficiency of the novel data-driven algorithms is proved using real-time experiments by controlling the 3 degrees of freedom tower crane system equipment.

Keywords: Modelling and simulation, System identification, Swarms
Abstract: Fractional-order differential equations are increasingly used to model systems in engineering for purposes such as control and health-monitoring. Because of the nature of a fractional derivative, mechanistically fractional-order dynamics will most naturally arise when there are non-local features in the dynamics. Even if there are no non-local effects, however, when searching for an approximate model for a very high order system, it is worth considering whether a fractional-order model is better than an integer-order model. This work is motivated by the challenges presented by very large scale systems, which will be increasingly common as integration of the control of formerly decoupled systems occurs such as in cyber-physical systems. Because fractional-order differential equations are more difficult to numerically compute, justifying the use of a fractional-order model is a balance between accuracy of the approximation and ease of computation. This paper constructs large, random networks and compares the accuracy of integer-order and fractional-order models for their dynamics. Over the range of parameter values considered, fractional-order models generally provide a more accurate approximation to the response of the system than integer order models. To ensure a fair comparison, both the fractional-order and integer-order models considered had two parameters.

Keywords: Multi-agent systems, Networked systems, Neural networks
Abstract: This paper studies the spread of opinions in social media networks through the lens of opinion dynamics. As more human interactions and public discourse move online, understanding opinion formation and evolution in social media is crucial for issues such as virtual marketing, information dissemination, and social security. We introduce a novel approach using recurrent neural networks (RNN) to monitor and predict interactions in these networks. Our method uses two configurations of RNN algorithms to predict the opinions of agents in an online social network, with results showing its effectiveness in predicting diverse opinions. The first configuration uses a sigmoid activation function to predict the binary opinions output (agree, disagree), while the second configuration uses the softmax function to predict more detailed opinions. For the simulation results, we considered a group of five agents interacting in the Twitter network on the subject of COVID-19. The social interaction for a 30-day period was captured and opinion dynamics prediction using the RNN was verified.

Keywords: Nonlinear systems, Nonlinear control, Optimisation
Abstract: We present the principled design of a control pipeline for the synthesis of policies from examples data. The pipeline, based on a discretized design, expounds the algorithm introduced in [1] to synthesize policies from examples for constrained, stochastic and nonlinear systems. The pipeline: (i) does not need the constraints to be fulfilled in the possibly noisy example data; (ii) enables control synthesis even when the data are collected from an example system that is different from the one under control. The design is benchmarked on an example that involves controlling an inverted pendulum with actuation constraints. The data that are used to synthesize the policy are collected from a pendulum that: (i) is different from the one under control; (ii) does not satisfy the actuation constraints.

Keywords: Hybrid systems, Nonlinear control
Abstract: In this work we consider uncertain impulsive systems in Brunovsky canonical form with possibly aperiodic impulses. Following the prescribed performance control methodology, a state feedback controller is designed to guarantee that between any two consecutive impulses, the output tracking error will converge to a neighborhood of zero of predefined size, in no greater than a user selected fixed-time. In addition, all signals in the closed-loop are bounded. Simulations clarify and verify the approach.

Keywords: Nonlinear control, Optimisation, Nonlinear systems
Abstract: Control Lyapunov function (CLF) paves the way for designing a certified controller with a known stable region, which is the out-most importance in control systems. Sum-of-Squares (SOS) optimization is one method to construct the CLF with this stable region known as a region of attraction (ROA). However, existing methods yield quite conservative results. A new approach for constructing CLF overcoming existing limitations is proposed in this paper. The proposed method is based on the Union Theorem in sum-of-squares optimization, which enables the application of more than one variable size region generated by positive functions known as the Shape Function. Numerical simulations demonstrate the effectiveness of the proposed method, which outperforms the existing methods and provides a significantly enhanced ROA.

Keywords: Nonlinear systems, Nonlinear control, Robust control
Abstract: For a class of uncertain, single-input single-output (SISO) nonlinear systems, we consider the problem of prescribed performance tracking control under strict control input constraints. By prescribed performance, we mean that the system's output should track a given reference signal, with an error confined within a user-prespecified time-varying envelope. The desired reference signal is also predefined by the user; however, when input saturation occurs, it is temporarily modified so that the tracking task becomes feasible given the provided level of saturation. With the proposed approach, it is proved that if the saturation level exceeds a given threshold that we explicitly quantify, then all closed-loop signals remain bounded and the desired reference can be tracked with arbitrary precision. Simulations verify and clarify theoretical results.

Keywords: Nonlinear systems, Discrete-event systems, Optimisation
Abstract: This article explores the theory of discrete-time gradient systems that converge in a finite amount of time and are governed by a difference equation with minima. Two algorithms with distinct structures are discussed, both aimed at achieving finite-time stabilization of these systems. These gradient-based algorithms have significant applications in solving optimization problems. Using the finite-time convergent techniques discussed in the article, a quadratic programming problem is solved, and an optimal solution is obtained within a finite time frame. The effectiveness of these proposed methods is demonstrated through simulation results.

Keywords: Adaptive control, System identification, Linear systems
Abstract: Exponentially stable extended adaptive observer is proposed for a class of linear time-invariant systems with unknown parameters and overparameterization. It allows one to reconstruct unmeasured states and bounded external disturbance produced by a known linear exosystem with unknown initial conditions if a weak requirement of regressor finite excitation is met. In contrast to the existing solutions, the proposed observer reconstructs the original (physical) states of the system rather than the virtual one of its observer canonical form. Simulation results to validate the developed theory are presented.

Keywords: Adaptive control, System identification, Linear systems
Abstract: An adaptive state observer is proposed for a class of overparametrized uncertain linear time-invariant systems without restrictive requirement of their representation in the observer canonical form. It evolves the method of Generalized Parameters Estimation-Based Observer design and, therefore, (i) does not require to identify Luenberger correction gain parameters, (ii) forms states using algebraic rather than differential equation. Additionally, the developed observer is applicable to systems with unknown output matrix and ensures exponential convergence of unmeasured state observation error under weak requirement of the regressor finite excitation. The effectiveness of the proposed solution is supported by simulation results.

Keywords: Linear systems, Algebraic and geometric methods, Computational methods
Abstract: The linear matrix inequality approach is proposed to state control design of discrete-time linear positive systems, guaranteeing the closed-loop system positiveness, enabling attenuation of the impact of disturbances on the system and, if it is necessary, also giving possibility to mount limiting quadratic constraints on state variables into design conditions. Constructing the set of linear matrix inequalities warranting the strictly positive structure and the Lyapunov inequality forcing quadratic stability of the controlled system, the design conditions outlined and proven are the main results of the paper. The diagonal stabilizability had to be included into the set of linear matrix inequalities to construct the closed-loop schemes with a positive control law gain. The proposed approach is numerically illustrated.

Keywords: Linear systems, Nonlinear control, Disturbance rejection
Abstract: The paper proposes a method for synthesizing the control of linear plants with a guarantee of finding the controlled variable in a given set under the condition that only the system output is measurable. In this work, the output feedback control is not used because of its complexity of synthesis, but the observer-based control using the Luenberger observer is used. A change of coordinates is applied to transfer the original problem with output constraints to a problem of control by an auxiliary variable without constraints. The controller’s adjustable parameter is selected from the solution of linear matrix inequalities, which enhances the practical applicability of the proposed method. Numerical simulations using Matlab confirm the effectiveness of the proposed method by demonstrating the boundedness of all signals in the control system and the presence of controlled signals within the given set.

Keywords: Nonlinear systems
Abstract: The paper is devoted to the problem of finite-time observer design for linear descriptor systems. The scheme of observer parameters selection is presented by linear matrix equations and inequalities. The proposed observer does not require system transformation to a canonical form and guarantees convergence of the observation error to zero in a finite time.

Keywords: Multi-agent systems, Event based systems, Decentralized control
Abstract: This paper considers the leader-following consensus control for linear multi-agent systems. Two improved dynamic event-triggered control frameworks are proposed. The first is based on a moving average approach, whereas the second is a fully-distributed control scheme based on a well-chosen Lyapunov function with rigorous proof of adjustable inter-event time. The proposed methods involve model-based estimation and clock-like auxiliary dynamic variables to eventually increase the inter-event time as long as possible. Compared to the static event-triggered strategy and the existing state-of-the-art dynamic event-triggered mechanism, the proposed approach significantly reduces the communication frequency while guaranteeing asymptotic convergence. Numerical simulations demonstrate the validity of the proposed theoretical results.

Keywords: Multi-agent systems, Neural networks, Cyber-physical systems
Abstract: This paper proposes a method to reduce the malicious agent's negative effects on a multi-agent reinforcement learning (MARL) system, including actor-critic architecture. The method achieves the overall goal of the MARL system, which is to increase the cumulative reward of all individual agents and reduce the malicious agents' harmful effects on the entire MARL system. Assuming that the adverse agent is detectable, we propose to change the malicious agent's neural network (NN) structure. By leveraging a comparative methodology, we have demonstrated that a specific NN architecture using a linear activation function surpasses another utilizing a sigmoid activation function in minimizing loss. Our analysis indicates that this performance differential is attributable to the utilization of distinct activation functions within the models. This approach involves calculating the gradient of the loss function with respect to the activation function. The claims have been proven theoretically, and the simulation confirms theoretical findings.

Keywords: Multi-agent systems, Nonlinear systems, Distributed systems
Abstract: We consider the distributed leader-follower consensus problem with event-triggered communications. The system under consideration is a non-linear input-affine multi agent system. The agents are assumed to have identical dynamics structure with uncertain parameters and satisfying an incremental stabilisability condition. A distributed control law is proposed which achieves consensus based on two novel Communication Triggering Conditions (CTCs): the first one to achieve an asymptotic consensus but without any guarantees on Zeno behaviour and the second one to exclude Zeno behaviour but with practical consensus.

Keywords: Multi-agent systems, Nonlinear systems, Intelligent transportation systems
Abstract: This paper presents control strategies based on timevarying convergent higher order control barrier functions for the coordination of networks of platoons. This network could be modelled by a class of leader-follower multi-agent systems, where the leaders have knowledge on the associated tasks and control the performance of their platoon involved vehicles. The followers are not aware of the tasks, and do not have any control authority to reach them. They follow their platoon leader commands for the task satisfaction. Signal temporal logic (STL) tasks are defined for the platoons coordination. Robust solutions for the task satisfaction, based on the leader’s accessibility to the follower vehicles’ states are suggested. In addition, using the notion of higher order barrier functions, decentralized barrier certificates for each vehicle evolving in a formation dynamic structure are proposed. Our approach finds solutions to guarantee the satisfaction of STL tasks independent of the agents’ initial conditions.

Keywords: Multi-agent systems, Robotics, Decentralized control
Abstract: We propose a decentralized solution for high-level multi-agent task planning problems in environments considering communication network failure. In particular, we consider that robots can only sense each other and communicate within a limited radius, yet, they may need to collaborate to accomplish their tasks. These tasks are given in Metric Interval Temporal Logic (MITL), which is capable to capture complex task specifications involving explicit time constraints. To substitute for the lacking communication networks, we deploy an agile robot (e.g., drones) to transfer information between the heavy-duty robots while executing tasks. We propose an algorithm to decompose each MITL formula that is assigned to the corresponding heavy-duty robot into an independent task of that robot and an independent request for others. The agile robot systematically pursues heavy-duty robots to exchange requests. The heavy-duty robots use formal methods-based algorithms to compute path plans satisfying the independent promises and the received requests. While the robots’ plan computation is fully decentralized, the satisfaction of all tasks is guaranteed (if such plans are found). The proposed solution can be applied to practical applications where the communication network fails or is restricted, such as post-catastrophe search and rescue and wildlife surveillance.

Keywords: Optimisation, Distributed systems, Multi-agent systems
Abstract: In this paper, we consider a distributed stochastic optimization problem where the goal is to cooperatively minimize a non-stationary mean-risk functional. Such problem is an integral part of many important problems in wireless networks, transportation systems, sensor networks, and others. In particular, we focus on the reduction of computational effort needed to achieve a certain level of accuracy. Thus, we propose an improved Simultaneous Perturbation Stochastic Approximation-based consensus algorithm that achieves better accuracy in contrast to an existing solution over the same time horizon and provide its theoretical analysis. We also show the convergence to a bound for mean-squared errors of estimates. The simulation validates the new algorithm in a multi-sensor multi-target problem.