Keywords: Uncertain systems
Abstract: Contract theory is the study of economic incentives when parties transact in the presence of private information. We augment classical contract theory to incorporate a role for learning from data, where the overa ll goal of the adaptive mechanism is to obtain desired statistical behavior. We consider applications of this framework to problems in federated learning, the delegation of data collection, and principal-agent regulatory mechanisms.

Keywords: Fault accommodation, Fault detection and identification, System reconfiguration
Abstract: In this work, we solve the problem of mitigating control authority degradation in real time. In particular, we focus on controlled nonlinear affine-in-control evolution equations with finite control input and finite- or infinite-dimensional state. We consider control input degradation parameterized by Lipschitz continuous maps. These degradation modes are encountered in practice due to actuator wear and tear, hard locks on actuator ranges due to over-excitation, as well as more general changes in the control allocation dynamics. In previous work, we have derived sufficient conditions for real-time identifiability of control authority degradation. In this work, we build on these results by introducing the concept of viabilizability, which deals with the existence of a viabilizing map. Viabilizing maps remap commanded control signals to viabilized signals that produce a minimally disturbed approximation of the commanded control signal after control authority degradation. We develop sufficient conditions on viabilizability for a class of control degradation modes, as well as error bounds on approximate viabilizing maps and methods for viabilizing fixed gain controllers.

Keywords: Fault detection and identification, Autonomous systems, Neural networks
Abstract: This paper presents a novel observer-based approach to detect and isolate faulty sensors in nonlinear systems. The proposed sensor fault detection and isolation (s-FDI) method applies to a general class of nonlinear systems. Our focus is on s-FDI for two types of faults: complete failure and sensor degradation. The key aspect of this approach lies in the utilization of a neural network-based Kazantzis-Kravaris/Luenberger (KKL) observer. The neural network is trained to learn the dynamics of the observer, enabling accurate output predictions of the system. Sensor faults are detected by comparing the actual output measurements with the predicted values. If the difference surpasses a theoretical threshold, a sensor fault is detected. To identify and isolate which sensor is faulty, we compare the numerical difference of each sensor measurement with an empirically derived threshold. We derive both theoretical and empirical thresholds for detection and isolation, respectively. Notably, the proposed approach is robust to measurement noise and system uncertainties. Its effectiveness is demonstrated through numerical simulations of sensor faults in a network of Kuramoto oscillators.

Keywords: Fault diagnosis, Filtering, Maritime
Abstract: Autonomous ships heavily depend on their sensor systems for safe and efficient operation. When these critical sensor systems are compromised by faults, the entire autonomous operation is put at risk. Detecting and accurately estimating the magnitude of such faults becomes imperative to ensure the reliability and safety of autonomous ships. In response to this challenge, this paper presents a robust methodology built upon adaptive Kalman filter with forgetting factor to estimate the magnitude of sensor faults. What sets our approach apart is the innovative perspective taken towards fault diagnosis. Instead of treating the fault as an additional state variable within the system, we directly estimate the fault magnitude based on the available measurements. Our approach is demonstrated through extensive simulations, showcasing the effectiveness and resilience of the proposed method. The results highlight its potential to significantly enhance the dependability of autonomous ships in the face of sensor faults, contributing to their continued success in a wide range of real-world applications.

Keywords: Fault detection and identification, Fluid flow systems, Optimization
Abstract: Stormwater networks are critical utility infrastructures designed to drain rainwater and nuisance flows, such as excess irrigation and groundwater seepage from urban communities. During this process, they can transport pollutants (e.g., pesticides, oils, and greases) to receiving waters such as rivers, bays and oceans. A recurring problem faced by these systems are dry weather flows (DWFs), where illicit discharges are introduced and propagated in the network during periods with no rain. Current techniques for monitoring DWFs consist of manual inspections and grab samples, which are costly and inefficient. However, with advances in sensing and communication, the Internet-of-Things (IoT) has enabled new opportunities for enhanced decision support and control. This paper proposes a quick and efficient physics-based backwards inference model to identify potential sources of pollutant discharges in DWFs, given time-series IoT observations and knowledge embedded in domain-expert simulations. Our approach leverages the underlying physics that drives flow propagation in stormwater systems, and optimizes multiple least-squares regressions to find potential DWF sources and their associated flows. We evaluate our backwards inference model on six real-world stormwater networks provided by domain experts, and show its efficacy in reconstructing anomalies.

Keywords: Fault detection and identification, Fault diagnosis, Robotics
Abstract: Strain wave gears (SWG) are employed in most robot joints, and hence monitoring their condition gets more important in robotic applications. The condition of the SWGs can, e.g., be observed by sensor signals of strain gauges that are mounted on the flex spline, the deformable part of the gear. In this paper, the potential of these sensor signals regarding the detection of faults in SWGs is evaluated. As a first important step towards fault detection in a real world application, synthetically generated sensor signals are considered that are derived from a simulation chain allowing the injection of different faults. In total, five distinct and practically relevant faults are considered and different algorithms are applied to classify them. In addition, robustness regarding disturbed synthetic data is investigated and the classifiers’ potential for out-of-distribution prediction is evaluated.

Keywords: Fault detection and identification, Fluid flow systems, Computational methods
Abstract: In this paper, we present a novel statistical convergence analysis for bilinear parameter estimators. We account for two variations of a two-stage separation technique introduced by Bai [1], where the variations differ in the second stage. It turns out for both estimators that the probability of a large error decreases as the inverse square root of the number of measurements. We numerically demonstrate the estimators' performance by solving a water leak localization problem involving bilinear parameter estimation.

Keywords: Optimal control, Computational methods
Abstract: This paper compares the impact of iterated and direct approaches to sensitivity computation in fixed step-size explicit singly diagonally-implicit Runge–Kutta (ESDIRK) methods when applied to optimal control problems (OCPs). We use the principle of internal numerical differentiation (IND) strictly for the iterated approach, i.e., reusing the iteration matrix factorizations, the number of Newton-type iterations, and Newton iterates, to compute the sensitivities. The direct method computes the sensitivities without using the Newton schemes. We compare the impact of the iterated and direct sensitivity computations in OCPs for the quadruple tank system. We benchmark the iterated and direct approaches with a base case. This base case is an OCP that applies an ESDIRK method with an exact Newton scheme and uses a direct approach for sensitivity computations. In these OCPs, we vary the number of integration steps between control intervals and we evaluate the performance based on the number of SQP and QPs iterations, KKT violations, and the total number of function evaluations, Jacobian updates, and iteration matrix factorizations. The results indicate that the iterated approach outperforms the direct approach but yields similar performance to the base case.

Keywords: Optimal control, Agents and autonomous systems, Military applications
Abstract: This paper focuses on the utilization of geometric tools in two-phase multiplayer reach-avoid games. In these games, a thief aims to enter a target region and subsequently reach a safe region while evading capture by guarders. We first obtain the solution to the game of kind in the game scenario where the safe region is a single point by using ellipses. Then we show that ellipses are also useful to solve the game of kind when two guarders cooperatively play against the thief. Furthermore, we study the dominance region for two-phase games with the help of ellipses. The construction method of the boundary of dominance regions is provided and illustrated with a numerical example.

Keywords: Optimal control, Computational methods, Robotics
Abstract: Nonlinear optimal control problems for trajectory planning with obstacle avoidance present several challenges. While general-purpose optimizers and dynamic programming methods struggle when adopted separately, their combination enabled by a penalty approach was found capable of handling highly nonlinear systems while overcoming the curse of dimensionality. Nevertheless, using dynamic programming with a fixed state space discretization limits the set of reachable solutions, hindering convergence or requiring enormous memory resources for uniformly spaced grids. In this work we solve this issue by incorporating an adaptive refinement of the state space grid, splitting cells where needed to better capture the problem structure while requiring less discretization points overall. Numerical results on a space manipulator demonstrate the improved robustness and efficiency of the combined method with respect to the single components.

Keywords: Optimal control, Robust control, Neural networks
Abstract: In this paper, we address the adversarial training of neural ODEs from a robust control perspective. This is an alternative to the classical training via empirical risk minimization, and it is widely used to enforce reliable outcomes for input perturbations. Neural ODEs allow the interpretation of deep neural networks as discretizations of control systems, unlocking powerful tools from control theory for the development and the understanding of machine learning. In this specific case, we formulate the adversarial training with perturbed data as a minimax optimal control problem, for which we derive first order optimality conditions in the form of Pontryagin's Maximum Principle. We provide a novel interpretation of robust training leading to an alternative weighted technique, which we test on a low-dimensional classification task.

Keywords: Optimal control, Robust control, UAV's
Abstract: We propose a novel robust nonlinear W-infinity optimal control method for dynamical systems with nonaffine control inputs. The nonlinear W-infinity control formulation extends the classic nonlinear H-infinity one, considering a weighted Sobolev norm of the cost variable. This approach assumes that the cost variable belongs to the weighted Sobolev space Wm,p,Gamma, ensuring continuity and differentiability up to degree m in a certain domain Omega. Consequently, in addition to the well-known features provided by the H-infinity approach in terms of disturbance attenuation, the closed-loop system benefits from the enhanced transient performance. Here, the robust nonlinear W-infinity optimal control problem is formulated via dynamic programming for increased-order systems, and a particular solution is proposed to the resulting Hamilton-Jacobi equation, along with the corresponding stability analysis. To validate the proposed method and its versatility, we provide numerical results for the control of a quadrotor. Additionally, leveraging the inherent L2-gain properties of our approach, we demonstrate that the resulting controller can achieve trajectory tracking with guaranteed asymptotic stability for the whole closed-loop system.

Keywords: Optimal control, Network analysis and control, Large-scale systems
Abstract: We present a method for optimal control with respect to a linear cost function for positive linear systems with coupled input constraints. We show that the Bellman equation giving the optimal cost function and resulting sparse state feedback for these systems can be stated explicitly, with the solution given by a linear program. Our framework admits a range of network routing problems with underlying linear dynamics. These dynamics can be used to model traditional graph-theoretical problems like shortest path as a special case, but can also capture more complex behaviors. We provide an asynchronous and distributed value iteration algorithm for obtaining the optimal cost function and control law.

Keywords: Optimal control, Stochastic control, Adaptive systems
Abstract: In this paper, we provide a theoretical framework that separates the control and learning tasks in a linear system. This separation allows us to combine offline model-based control with online learning approaches and thus circumvent current challenges in deriving optimal control strategies in applications where a large volume of data is added to the system gradually in real time and not altogether in advance. We provide an analytical example to illustrate the framework.

Keywords: Predictive control for linear systems, Behavioural systems
Abstract: Recently proposed data-driven predictive control schemes for LTI systems use non-parametric representations based on the image of a Hankel matrix of previously collected, persistently exciting, input-output data. Persistence of excitation necessitates that the data is sufficiently long and, hence, the computational complexity of the corresponding finite-horizon optimal control problem increases. In this paper, we propose an efficient data-driven predictive control (eDDPC) scheme which is both more sample efficient (requires less offline data) and computationally efficient (uses less decision variables) compared to existing schemes. This is done by leveraging an alternative data-based representation of the trajectories of LTI systems. We analytically and numerically compare the performance of this scheme to existing ones from the literature.

Keywords: Statistical learning, Uncertain systems, Optimal control
Abstract: As control engineering methods are applied to increasingly complex systems, data-driven approaches for system identification appear as a promising alternative to physics-based modeling. While the Bayesian approaches prevalent for safety-critical applications usually rely on the availability of state measurements, the states of a complex system are often not directly measurable. It may then be necessary to jointly estimate the dynamics and the latent state, making the quantification of uncertainties and the design of controllers with formal performance guarantees considerably more challenging. This paper proposes a novel method for the computation of an optimal input trajectory for unknown nonlinear systems with latent states based on a combination of particle Markov chain Monte Carlo methods and scenario theory. Probabilistic performance guarantees are derived for the resulting input trajectory, and an approach to validate the performance of arbitrary control laws is presented. The effectiveness of the proposed method is demonstrated in a numerical simulation.

Keywords: Identification, Randomized algorithms, Predictive control for linear systems
Abstract: The problem of data driven recursive computation of receding horizon LQR control through a randomized combination of online/current and historical/recorded data, is considered. It is assumed that large amounts of historical input-output data from a system, which is similar but not identical to the current system under consideration, is available. This (possibly large) data set is compressed through a novel randomized subspace algorithm to directly synthesize an initial solution of the standard LQR problem, which however is sub-optimal due to the inaccuracy of the historical model. The first instance of this input is used to actuate the current system and the corresponding instantaneous output is used to iteratively re-solve the LQR problem through a computationally inexpensive randomized rank-one update of the old compressed data. The first instance of the re-computed input is applied to the system at the next instant, output recorded and the entire procedure is repeated at each subsequent instant. As more current data becomes available, the algorithm learns automatically from the new data while simultaneously controlling the system in near optimal manner. The proposed algorithm is computationally inexpensive due to the initial and repeated compression of old and newly available data. Moreover, the simultaneous learning and control makes this algorithm particularly suited for adapting to unknown, poorly modeled and time varying systems without any explicit exploration stage. Simulations demonstrate the effectiveness of the proposed algorithm vs popular exploration/exploitation approaches to LQR control.

Keywords: Identification for control, Predictive control for linear systems, Stochastic control
Abstract: The practical implementation of Model Predictive Control (MPC) often presents challenges that remain unaddressed in theoretical formulations. Among these challenges, the tuning of the receding horizon cost becomes particularly intricate in the context of data-driven learning-based MPC, where models exhibit partial uncertainty. This paper introduces SelfMPC, a pioneering approach within a Gaussian process learning framework, illustrating that a tracking MPC cost can be formulated as the maximum likelihood estimation of the reference output. This formulation provides automatic cost shaping and effective regularization, eliminating the need for manual tuning efforts. Moreover, the proposed formulation provides a natural way to employ information from empirical experiments into the definition of the MPC optimization problem for unknown systems. Empirical validation against conventional weighting matrix selection methods confirms the effectiveness of the proposed approach.

Keywords: Fault detection and identification, Machine learning, Large-scale systems
Abstract: Graph-based state interpolation (GSI) is a state-of-the-art state reconstruction technique that operates over water distribution networks (WDN). This scheme retrieves the complete hydraulic state (represented by the nodal heads) from the network topology and pressure measurements gathered through a set of installed sensors. This process is coupled with a second stage to perform leak localization, comparing interpolated leak and leak-free states. This article presents a methodology to adapt GSI in order to learn from its off-line and on-line operation (i.e., gain knowledge about historical located leaks, as well as leaks appearing in the network) and using this information to improve its performance for future similar leak events. The methodology is tested over a well-known case study (Modena), showing promising results in terms of localization performance.

Keywords: Robotics, Linear time-varying systems, Observers for linear systems
Abstract: This paper addresses the problem of simultaneous estimation of the position, linear velocity and orientation of a rigid body using single bearing measurements. We introduce a Riccati observer-based estimator that fuses measurements from a 3-axis accelerometer, a 3-axis gyroscope, a single body-frame vector observation (e.g., magnetometer), and a single bearing-to-landmark measurement to obtain the full vehicle's state (position, velocity, orientation). The proposed observer guarantees global exponential convergence under some persistency of excitation (PE) condition on the vehicle's motion. Simulation results are presented to show the effectiveness of the proposed approach.

Keywords: Robotics, Optimization, Safety critical systems
Abstract: In the field of control engineering, the connection between Signal Temporal Logic (STL) and time-varying Control Barrier Functions (CBF) has attracted considerable attention. CBFs have demonstrated notable success in ensuring the safety of critical applications by imposing constraints on system states, while STL allows for precisely specifying spatio-temporal constraints on the behavior of robotic systems. Leveraging these methodologies, this paper addresses the safety-critical navigation problem, in Socially Responsible Navigation (SRN) context, presenting a CBF-based STL motion planning methodology. This methodology enables task completion at any time within a specified time interval considering a dynamic system subject to velocity constraints. The proposed approach involves real-time computation of a smooth CBF, with the computation of a dynamically adjusted parameter based on the available path space and the maximum allowable velocity. A simulation study is conducted to validate the methodology, ensuring safety in the presence of static and dynamic obstacles and demonstrating its compliance with spatio-temporal constraints under non-linear velocity constraints.

Keywords: Robotics, Predictive control for nonlinear systems
Abstract: This paper presents a three-layered hierarchical architecture to control manipulation tasks which involve in- teractions between the robot and workpieces. Learning from demonstration (LfD) is exploited to train dynamical movement primitives (DMPs) as fundamental building blocks for such tasks. Thus, position and wrench profiles are obtained from a kinesthetic demonstration, which makes the programming process intuitive for factory workers without expert knowledge. A model predictive path-following controller is used as the underlying control method, in order to make use of the explicit consideration of constraints within the controller formulation. Thereby, the progress parameter of the path-following control is used as the phase variable of the DMPs which results in a tight coupling of both methods. Finally, experimental results on a real robot system prove the effectiveness and real-time capable implementation of the approach.

Keywords: Robotics, Iterative learning control
Abstract: In this study, we introduce a control strategy that combines Proportional-Derivative (PD) control with Iterative Learning Control (ILC) to enhance legged robot velocity control with only the inverse kinematics and no additional system identification. This approach leverages the real-time feedback capabilities of PD control for gait tracking while incorporating ILC's learning abilities to eliminate inaccuracies from unmodeled dynamics iteratively and to reach desired velocities without residual errors. By uniting these techniques, the proposed method empowers legged robots to adapt and optimize their control behavior, achieving and maintaining desired walking velocities. Experimental results on the physical legged robot Go1 demonstrate the effectiveness of the proposed approach, highlighting its adaptability and reliability in real-world scenarios. This research represents a first step towards overcoming high computational effort and extensive data collection for quadruped robot velocity tracking through onboard learning.

Keywords: Robotics, Machine learning, Neural networks
Abstract: This paper investigates the impact of training data with different quantities and compositions on the performance and robustness of a Neural Network (NN) controller for the wheel loader bucket filling task. Collecting training data for machine learning methods with a real-world Heavy Duty Mobile Machine (HDMM) is expensive, and therefore knowing how to collect the data and in what quantities will significantly reduce the data collection effort. We collected 2000 bucket fillings of non-homogeneous material, more specifically, a blasted rock pile with a kernel size of 0–400 mm. No previous study has reported such a challenging material composition. The collected data was divided into 6 datasets with sizes of 10, 20, 50, 100, 500, and 2000 bucket fillings. We use the Dynamic Time Warp (DTW) distance, k-medoids clustering, and the silhouette score, to create diverse and dissimilar datasets. Furthermore, one additional dataset was created with 10 bucket fillings, which are as similar as possible, resulting in 7 datasets in total. The datasets were used to synthesize 7 controllers that were then evaluated with a set of experiments to compare their performance to one another and the human operator. The results showed that the controller trained on similar bucket fillings was not robust and had poor performance, as expected. The experiment also showed that all the controllers trained on diverse data were robust enough to load the blasted rock material. However, the loaded material weight was less than the human operator, where the best controller loaded 9% less material weight, but 11% faster than the human operator.

Keywords: Robotics, Mechatronics, Stochastic filtering
Abstract: Pedestrian tracking using Light Detection and Ranging (LiDAR) is important to avoid pedestrians for autonomous vehicles. It is difficult, however, to measure the center position of the pedestrian directly because the point cloud data appears on the surface of the pedestrian facing the LiDAR. In addition, the arms and legs often wobble during locomotion, although the torso moves smoothly and wobbles less. Thus, in this study, to estimate the center position of the pedestrian, we approximate the torso as an ellipse model to represent the pedestrian's pose. Then, the point cloud on arms and legs that largely fluctuates is eliminated by Random Sample Consensus (RANSAC) by regarding them as outliers for the ellipse model. In addition, we propose a novel method that combines Moving Horizon Estimation (MHE) with maximum likelihood estimation sampling consensus (MLESAC) to consider the motion model of the pedestrian to prevent fitting failures. Experiments show the advantages of the proposed method.

Keywords: Stability of nonlinear systems, Delay systems, Communication networks
Abstract: Passive intermodulation interference in wireless systems is caused by downlink radio transmissions that are nonlinearly mixed and frequency shifted in close conductive objects. This interference may jam uplink receivers. The paper proposes mitigation by logarithmic feedback control using downlink transmission power actuation. The nonlinear single downlink feedback loop is proved to be globally stable provided that a loop-gain-loop-delay stability condition holds. The multi downlink problem is then addressed by derivation of a degree greedy multiple-input-single-output controller that is proved to be optimal in a static sense. Simulations using measured data traffic illustrate the performance of the control systems.

Keywords: Stability of nonlinear systems, Lyapunov methods, Cellular dynamics
Abstract: A common tool in system theory for formulating control laws that achieve local asymptotic stability are Control Lyapunov functions (CLFs), while Control Barrier functions (CBFs) are typically employed to enforce safety constraints. Combining these two types of functions is of interest, because it leads to stabilizing controllers with safety guarantees. A common approach to merge CLFs and CBFs is to solve an optimization problem where both CLF and CBF inequalities are imposed as constraints. In this paper, we show via an example from the literature that this approach can lead to undesirable behavior (i.e., slow convergence and oscillating inputs). Then, we propose a novel cost function that penalizes the deviation from Sontag's formula by using a state-dependent weighting matrix. We show that by minimizing the developed cost function subject to a CBF constraint, local asymptotic stability is obtained with an explicit domain of attraction, without using a CLF constraint. To deal with vanishing properties of the weight matrix as the state approaches the equilibrium, we introduce a hybrid continuous control law that recovers Sontag's formula locally. The effectiveness of the developed hybrid stabilizing control law based on CLFs and CBFs is illustrated in stabilization of a 3D tumor model, subject to physiological constraints (i.e., all states must be positive), which yields useful insights into optimal cancer treatment design.

Keywords: Nonlinear system theory, Lyapunov methods, Stability of nonlinear systems
Abstract: This paper proposes a novel methodology to design a gain-scheduled state-feedback control law for nonlinear systems subjected to input saturation constraints and bounded external disturbances. The overall goal is disturbance rejection understood as determining the smallest possible inescapable set starting from zero initial conditions. The control design is addressed via iterative algorithms based on Linear Matrix Inequalities, which are obtained from the application of H1 star norm together with small gain argumentations. Numerical simulations are provided to show the effectiveness of the proposed method.

Keywords: Nonlinear system theory, Lyapunov methods
Abstract: This paper applies the physics-informed neural network approach to approximate globally-defined Lyapunov functions and stabilizing controls for homogeneous dynamical systems. The advantage of this class of systems is that all analysis and design can be made locally on a suitably defined unit sphere, which corresponds perfectly to the applicability conditions of neural networks.

Keywords: Stability of nonlinear systems, Lyapunov methods, Optimization
Abstract: For complex nonlinear systems, it is challenging to design algorithms that are fast, scalable, and give an accurate approximation of the stability region. This paper proposes a sampling-based approach to address these challenges. By extending the parametrization of quadratic Lyapunov functions with the system dynamics and formulating an ell_1 optimization to maximize the invariant set over a grid of the state space, we arrive at a computationally efficient algorithm that estimates the domain of attraction (DOA) of nonlinear systems accurately by using only linear programming. The scalability of the Lyapunov function synthesis is further improved by combining the algorithm with ADMM-based parallelization. To resolve the inherent approximative nature of grid-based techniques, a small-scale nonlinear optimization is proposed. The performance of the algorithm is evaluated and compared to state-of-the-art solutions on several numerical examples.

Keywords: Stability of nonlinear systems, Algebraic/geometric methods
Abstract: We characterize when a compact, invariant, asymptotically stable attractor on a locally compact Hausdorff space is a strong deformation retract of its domain of attraction.

Keywords: Energy systems, Maritime, Optimal control
Abstract: Electrified propulsion systems, such as fuel cells (FCs) and batteries, are a promising solution to decarbonize the shipping sector. In this paper, we have conducted a comprehensive analysis of two months' worth of real-world container ship power demand data. From this analysis, we propose a novel multi-time scale Energy Management System (EMS) approach for a hybrid FC/battery propulsion system. This approach enables the individual control of each FC stack while factoring in battery and FC degradation losses and fuel consumption costs. By exploring different time scales, we have assessed the trade-offs between time complexity and system optimality, which has led us to devise an efficient strategy for the energy management of FC/battery hybrid ships.

Keywords: Energy systems, Observers for linear systems, Modeling
Abstract: This study presents a novel thermal model for cylindrical lithium-ion batteries using ten Ordinary Differential Equations (ODEs). The model covers all battery components and focuses on a simplified assembly with eight parts rolled up on a central mandrel and gaps filled with liquid electrolyte. One key input to the thermal model is the thermal power generated in the battery, calculated using a simplified electrochemical Single Particle Model (SPM) with two Partial Differential Equations (PDEs). This leads to a coupled system of 20 ODEs, where 10 ODEs represent the electrochemical model based on Pade approximation method, and 10 ODEs represent the thermal model. Temperature profiles within the battery are estimated using the Luenberger observer, with feasible sensor placement strategies discussed. Simulation results validate the model’s accuracy by demonstrating temperature consistency between the thermal model and the Luenberger observer.

Keywords: Energy systems, Identification
Abstract: The open circuit voltage (OCV) is a chemistry-dependent curve that characterizes the steady-state behaviour of a lithium-ion battery cell (LIB), namely the link between the voltage accross the cell and its state of charge (SOC), when no load is connected. It is a fundamental ingredient to estimate the SOC of LIBs with any technique, ranging from a simple look-up table to a Kalman filter based on an equivalent circuit model or an electrochemical model. However, an accurate determination of such a steady-state curve for a commercial LIB requires extensive standalone experimental campaigns that are time-consuming. This work, compares two different methods, Karhunen-Loève transform (or principal component analysis) and Gaussian process regressions, to model the zero-current behaviour of 18650 LIBs (Sony VTC6 and Samsung 30Q, 3000mAh) considering OCV-SOC charge and discharge experimental curves at C-rates between C/50 and C/5. The zero-current extrapolation given by these methods can reduce the experimental time up to 85% when compared to an average OCV characterized at C/100.

Keywords: Machine learning, Fault estimation, Fault diagnosis
Abstract: Advanced machine learning (ML) models have been developed for battery lifetime prediction in different use cases at all stages of a battery's life. As the first step to enable the transferability of ML models for battery lifetime prediction across multiple use cases, a scenario-aware machine learning pipeline is proposed, in which two feature engineering methods that have been able to generate input features with outstanding predictive power are used to learn the best ML model for battery lifetime prediction in a chosen usage scenario. The experimental results show that the histogram-based feature engineering method is able to generate input features with predictive power generalized across two usage scenarios (i.e., identical cycling and protocol cycling). Thus, to enable transferability of ML models for battery lifetime prediction across different scenarios, and even battery chemistries, this histogram-based feature engineering method will be further investigated together with online fine-tuning strategies.

Keywords: Aerospace, Electrical power systems, Optimal control
Abstract: The transition towards electric aircraft is particularly challenging due to the relatively low specific energy densities of electrical energy storage systems. If electric aircraft are to be realised, flight paths must be optimised to take advantage of the unique features of electric propulsion systems. In this paper, the problem of determining the optimal flight trajectories for aircraft powered by a combination of a fuel cell stack and lithium-ion battery pack is considered. Particular emphasis is given to the role of the battery pack's temperature and in-flight charging requirements on the results. Solutions that minimise the fuel consumption and the flight time are first considered. For both cases, the optimal solution was observed to discharge the battery during the climb with only minimal in-flight recharging of the battery by the fuel cell. A scenario that requires a fast climb and descent and the battery to be charged upon arrival was identified and shown to lead to an oscillatory profile for optimal in-flight charging. These results demonstrate the potential of solving optimal control problems to generate tailored electric aircraft trajectories.

Keywords: Electrical power systems, Energy systems, Optimization algorithms
Abstract: This work presents a study on the optimal control of an electric vehicle (EV) charging station, with a dual objective of minimizing operational expenses and accommodating EV owner preferences. The control frame work employs Model Predictive Control (MPC) to efficiently distribute power among the stations components, encompassing photovoltaic panel (PV), and energy storage system (ESS) battery, the grid connection, and connected EVs. Our MPC scheme takes into account EV owner requirements, with owners providing information about their charging needs (required energy) and departure times when plugging in their vehicles. With this extra information that helps energy management of the system, we exploited the integration of Vehicle-to-Grid (V2G) technology, enabling bidirectional power flow to and from grid. This feature allows EVs to supply electricity to the grid during periods of high demand or when solar energy generation is insufficient. The proposed MPC-based method is validated through simulation and compared with heuristic method that disregards owner information and tends to charge EV at maximum power rate available. This comparative analysis serves to evaluate the efficacy of the proposed approach in terms of cost savings, good energy management, and owner satisfaction.

Keywords: Aerospace, Computational methods, UAV's
Abstract: The computation of symbolic controllers for nonlinear plants is typically computationally expensive due to the well-known curse-of-dimensionality. In fact, those controllers must be computed before operating the closed loop. This note presents a method to modify symbolic controllers while they are operating the closed loop to avoid spontaneously inserted state obstacles. In addition, we utilize methods of plan recognition in combination with our new algorithm for providing a technique of decentralized runtime assurance for efficient task allocation and mission guidance in a multi-UAV setting. Promising results and the applicability of the found method is demonstrated by simulation and experiments with real physical systems.

Keywords: Aerospace, UAV's, Biological systems
Abstract: Dynamic soaring is a remarkable flight strategy employed by soaring birds like albatrosses to harness energy from the atmospheric wind gradient. This strategy is so efficient that soaring birds can sustain flight for very long distances without almost flapping their wings. This phenomenon has intrigued researchers across multiple disciplines including biology, physics, and applied mathematics. For aerospace and control engineering researchers, mimicking dynamic soaring means new technologies that contribute to a more sustainable aviation industry. Significant work has been done in the literature to mimic dynamic soaring using optimal control frameworks. However, these approaches have limitations as they are non-real-time, model-dependent, and computationally expensive. Very recently, the authors of this paper introduced a novel autonomous, real-time and model-free approach for mimicking dynamic soaring utilizing Extremum Seeking Control (ESC) methods. However, the ESC structures used in said emerging approach are sensitive to the curvature of the input-output map of the system. Therefore, in this paper, we propose a Newton-based ESC structure for dynamic soaring that is independent of the input-output map's curvature. This provides a twofold contribution: (1) further solidification that the dynamic soaring problem can be treated as a natural ESC system; and (2) a framework that captures dynamic soaring independent of the input-output map's curvature, which can be particularly useful in cases where the system model is unknown. We verify our real-time results via simulations and comparison with non-real-time powerful optimal control solvers.

Keywords: Aerospace, Cooperative control, Sliding mode control
Abstract: A cooperative guidance strategy is developed to force multiple interceptors to intercept a target simultaneously. The guidance law works to minimize the time-to-go difference between neighboring interceptors while still keeping the interceptors on track for interception. The guidance law is derived using sliding mode control, with one sliding surface for every pair of neighboring interceptors to remove time-to-go difference and one global sliding surface to make sure at least one interceptor is heading towards the target, guaranteeing the others as well. A time-to-go approximation scheme for a stationary target is used during the derivation. A two-dimensional nonlinear simulation of the relative kinematics is run for cases of both two interceptors and more, in which the guidance law is shown to successfully cause simultaneous interception between multiple interceptors starting from different initial conditions.

Keywords: Aerospace, Fault detection and identification
Abstract: In this paper we address the problem of detecting anomalies in the reaction wheel assemblies (RWAs) of a satellite. These anomalies can alert of an impending failure in a RWA, and effective detection would allow to take preventive action. To this end, we propose a novel algorithm that detects and categorizes anomalies in the friction profile of a RWA, where the profile relates spin rate to measured friction torque. The algorithm, developed in a probabilistic framework, runs in real time and has tunable false positive rate as a parameter. The performance of the proposed method is thoroughly tested in a number of numerical experiments, with different anomalies of varying severity.

Keywords: Aerospace, Lyapunov methods
Abstract: Space debris removal and on-orbit servicing missions require docking spacecraft guidance that not only enables reaching the target spacecraft but also avoids collisions with appendages protruding from it. In this paper, two modified guidance laws are proposed for spacecraft docking on non-cooperative tumbling targets, with collision avoidance, using Lyapunov Vector Fields. In the literature a two stage thrust constrained guidance strategy has been proposed, with the first stage used for a station keeping approach trajectory and the second for contraction towards the docking port. Our proposed strategy follows a similar approach, but additionally ensures that the chaser spacecraft enters and stays on a safe approach plane during both stages, so that collisions with protruding appendages of the target spacecraft can be avoided. This is done by introducing a third component of motion in the target’s body frame away from the initial approach plane, towards a plane free from intersections with undesirable appendages. It is shown that with these modified guidance laws, asymptotic convergence is guaranteed to the respective goal locations of each stage. A maneuver design example along with numerical simulations is presented to demonstrate the effectiveness of the proposed guidance strategy on a real mission.

Keywords: Aerospace, Transportation systems, Safety critical systems
Abstract: Aircraft anti-skid systems are key to maintaining directional control during ground handling and must balance performance and robustness over a wide operational envelope. In particular, the impact of longitudinal speed on the braking dynamics induces an important coupling between the longitudinal and vertical dynamics due to the aerodynamic effects. In this paper, longitudinal slip-based and wheel speed deceleration-based anti-skid controllers are designed based on control-oriented models of the braking dynamics for an aircraft with a tricycle landing gear configuration. A gain-scheduling strategy is devised to achieve high performance and maintain stability during landing maneuvers. The stability of the resulting closed-loop system affected by parametric variability and discretization effects is later verified in the framework of Linear Parameter-Varying systems by formulating a set of efficient Linear Matrix Inequalities. The resulting anti-skid designs are successfully evaluated in a validated multibody simulator for a target aircraft.

Keywords: Distributed control, Energy systems, Optimization
Abstract: The demand-side control of district heating networks is notoriously challenging due to the large number of connected users and the high number of states to be considered. To overcome these challenges, this paper presents a hierarchical optimization scheme using the flexibility in heating demand provided by the users to improve the performance of the network. This hierarchical scheme relies on a low level controller to calculate the costs for a subsystem over a given set of potential pressure drops for that subsystem. The high level controller then uses these calculated costs to determine the optimal set of pressure drops for every subgraph of the partitioned network. The proposed hierarchical optimization scheme is demonstrated on a representative 20 user district heating network, resulting in a 67% reduction in bypass mass flow while ensuring all network users stay within 2 degree C of their desired nominal temperatures.

Keywords: Distributed control, Decentralized control, Energy systems
Abstract: A scalable decentralised-distributed secondary control architecture for frequency restoration and power sharing control in MicroGrid (MG) clusters, is proposed in this paper. The proposed scheme allows to reduce the communication cost of classical distributed control schemes, thus possibly reducing privacy and security concerns, while guaranteeing scalability properties, by adopting a leaky integral controller. Based on the droop control mechanism, the model for Alternating Current (AC) MG clusters is considered. A leader-follower-based decentralised-distributed secondary control is then designed for frequency restoration and active power sharing among and within clusters, respectively. The parameters selection for the proposed controller is also analysed and simulations show the effectiveness of the proposed decentralised-distributed control method.

Keywords: Distributed control, Nonlinear system theory, Optimization
Abstract: This paper investigates the distributed feedback optimization problem of nonlinear multi-agent systems. In such systems, each agent can measure the relative outputs between itself and its neighbors but lacks access to their absolute states and internal controller states. By combining distributed optimization and singular perturbation methods, a novel distributed controller design is presented, that relies solely on each agent’s real-time gradient values of its local objective function and its relative output measurements to neighboring agents. The boundedness of the closed-loop signals and the convergence of the agent outputs to the minimizer of the total cost are proved rigorously. A numerical example is conducted to validate the effectiveness of the proposed approach.

Keywords: Distributed control, Cooperative control, Agents and autonomous systems
Abstract: This paper deals with the potential benefits of coalition-based modelling and control of building zones heating and cooling to enhance the overall energy efficiency of the building. The paper applies model predictive control for tracking of setpoint temperature in building zones. To tackle the complexity of the model, the approach utilizes semiphysical independent thermal models of building zones and models thermal connections between adjacent zones based on resistive-capacitive analogy. Such coupled models are formed into coalitions and are used in model predictive control of zone temperature. Realistic two 5-day simulations are conducted to compare the performance of the coalitional control. Results indicate that the coalitional control approach reduces energy consumption by up to 14.96% in comparison to decentralized approach of independent model predictive control of each zone while providing consumption increase of 1.38% compared to a centralized floor-based model predictive control. The paper provides evaluation of coalitional control performance based on trade-off between energy savings, user comfort, and computational complexity.

Keywords: Distributed control, Cooperative control, Robotics
Abstract: In prioritized planning for vehicles, vehicles plan trajectories in parallel or in sequence. In parallel prioritized planning, the computation time remains approximately constant with an increasing number of vehicles, but it is difficult to guarantee collision-free trajectories. Although sequential prioritized planning can guarantee collision-free trajectories, the computation time increases with the number of sequentially computing vehicles, which we call computation levels. This number is determined by the directed coupling graph which results from the coupling and prioritization of vehicles. This work's contribution is twofold. First, we guarantee safe trajectories in parallel planning through reachability analysis. Although these trajectories are collision-free, they tend to be conservative. Second, we address this conservativeness by planning with a subset of vehicles in sequence. We formulate the problem of selecting this subset as a graph partitioning problem, in which we limit the size of the resulting subgraphs. Consequently, we can choose the number of computation levels independently from the directed coupling graph, and thus are able to limit the computation time in prioritized planning. In our simulations, we reduce the number of computation levels to approximately 64% compared to sequential prioritized planning while maintaining the solution quality.

Keywords: Distributed control, Predictive control for nonlinear systems, Agents and autonomous systems
Abstract: We consider the synthesis problem of a multi-agent system under signal temporal logic (STL) specifications representing bounded-time tasks that need to be satisfied recurrently over an infinite horizon. Motivated by the limited approaches to handling recurring STL systematically, we tackle the infinite-horizon control problem with a receding horizon scheme equipped with additional STL constraints that introduce minimal complexity and a backward-reachability-based terminal condition that is straightforward to construct and ensures recursive feasibility. Subsequently, we decompose the global receding horizon optimization problem into agent-level programs the objectives of which are to minimize local cost functions subject to local and joint STL constraints. We propose a scheduling policy that allows individual agents to sequentially optimize their control actions while maintaining recursive feasibility. This results in a distributed strategy that can operate online as a model predictive controller. Last, we illustrate the effectiveness of our method via a multi-agent system example assigned a surveillance task.

Keywords: Game theoretical methods, Agents and autonomous systems, Stability of nonlinear systems
Abstract: The replicator equation in evolutionary game theory describes the change in a population's behaviors over time given suitable incentives. It arises when individuals make decisions using a simple learning process - imitation. A recent emerging framework builds upon this standard model by incorporating game-environment feedback, in which the population's actions affect a shared environment, and in turn, the changing environment shapes incentives for future behaviors. In this paper, we investigate game-environment feedback when individuals instead use a boundedly rational learning rule known as logit learning. We characterize the resulting system's complete set of fixed points and their local stability properties, and how the level of rationality determines overall environmental outcomes in comparison to imitative learning rules. We identify a large parameter space for which logit learning exhibits a wide range of dynamics as the rationality parameter is increased from low to high. Notably, we identify a bifurcation point at which the system exhibits stable limit cycles. When the population is highly rational, the limit cycle collapses and a tragedy of the commons becomes stable.

Keywords: Game theoretical methods, Machine learning, Agents networks
Abstract: We consider finite games where the agents only share their beliefs on the possible equilibrium configuration. Specifically, the agents experience the strategies of their opponents only as realized parameters, thereby updating and sharing beliefs on the possible configurations iteratively. We show that combining non-bayes updates with best-response dynamics allows the agents to learn the Nash equilibrium, i.e., the belief distribution over the set of parameters has a peak on the true configuration. Convergence results of the learning mechanism are provided in two cases: the agents learn the equilibrium configuration as a whole, or the agents learn those strategies of the opponents that constitute such an equilibrium.

Keywords: Game theoretical methods, Optimization algorithms, Communication networks
Abstract: This work proposes a novel distributed approach for computing a Nash equilibrium in convex games with restricted strongly monotone pseudo-gradients. By leveraging the idea of the centralized operator extrapolation method presented in [4] to solve variational inequalities, we develop the algorithm converging to Nash equilibria in games, where players have no access to the full information but are able to communicate with neighbors over some communication graph. The convergence rate is demonstrated to be geometric and improves the rates obtained by the previously presented procedures seeking Nash equilibria in the class of games under consideration.

Keywords: Game theoretical methods, Emerging control applications
Abstract: Using information consensus and replicator dynamics (RD), this article presents a distributed algorithm for designing control schemes in urban drainage systems (UDSs). It demonstrates the stability of a closed-loop model with RD in UDSs using passivity arguments for single subsystems. As central models for UDSs, we present two distinct topologies and conduct passivity-based analysis to design appropriate payoff mechanisms. We further extend this to a decentralized scenario in which subsystems within a UDS share information and increase capacity at specific sites in response to intense rainfall. This algorithm with distributed consensus assistance seeks to improve system performance. Several simulations are presented to illustrate the benefits of this method.

Keywords: Game theoretical methods, Markov processes, Communication networks
Abstract: In complex large-scale systems, distributed solutions are often preferred over centralized ones due to their computational feasibility when dealing with control and state estimation algorithms. One prominent distributed state observer is the Interlaced Kalman Filter (IKF), which employs a set of parallel Kalman filters, each dedicated to estimating a specific subset of the overall state space. These filters then exchange information with their neighboring filters to collectively infer the current state of the entire system. However, in practical scenarios, the traditional filters synchronization, when at each iteration filters exchange their most recent estimations, is often unattainable due to technological constraints or different sensor working rates.

To address this problem, the Multirate Interlaced Kalman Filter is developed to allow two stations to synchronize only at fixed time intervals. When synchronization is difficult to achieve, these stations possess outdated information about a specific system portion. The innovative concept behind the Multirate Kalman filter is to exploit outdated information for state estimation by artificially increasing its uncertainty, represented by the covariance matrix.

This paper presents a dual-purpose approach: firstly, we scrutinize the structure of the Interlaced Kalman Filter to minimize information exchange among the filters, thereby reducing the computational demands of implementation. Secondly, we leverage the stochastic stability lemma to monitor the real-time performance of the filter, providing an error bound for assessment. We initially analyze the linear case study and then extend our findings to the non-linear one. Numerical simulations supports the effectiveness of our proposed approach.

Keywords: Game theoretical methods, Autonomous robots, Agents and autonomous systems
Abstract: This paper addresses the problem of multi-robot coordinated navigation in target localization missions. We employ a potential game with an added penalization term that is dynamically updated to improve the performance of the multi-robot system by decreasing the number of movements needed to localize the targets and therefore, save time and energy. Then, we give conditions on this penalization map to guarantee low probabilities of revisiting explored zones on consecutive turns. By employing binary log-linear learning (BLLL) we solve the game for different simulated scenarios and compare them to recently developed strategies. Afterwards, we implement a decentralized controller on a robot simulator and illustrate the penalization map on a physical robot in a simple scenario.

Keywords: Distributed parameter systems, Fault detection and identification, Fault estimation
Abstract: We propose a new method for leak detection and localization in water pipes based on a mathematical model that describes the flow dynamics by two coupled linear first order hyperbolic partial differential equations. Using the modulating function approach, a system of auxiliary PDEs is derived and solved in order to obtain appropriate modulating functions. This allows estimating the leak size and the leak position, resorting to algebraic I/O equations only. For this purpose, no spatial discretization of the PDE model is needed. The theoretical results are validated with experimental data from a water pipe prototype and the performance of the proposed approach is evaluated in comparison to an existing late lumping model-based leak detection system.

Keywords: Distributed parameter systems, Neural networks, Machine learning
Abstract: For the recently introduced deep learning-powered approach to PDE backstepping control, we present an advancement applicable across all the results developed thus far: approximating the control gain function only (a function of one variable), rather than the entire kernel function of the backstepping transformation (a function of two variables). We introduce this idea on a couple benchmark (unstable) PDEs, hyperbolic and parabolic. We alter the approach of quantifying the effect of the approximation error by replacing a backstepping transformation that employs the approximated kernel (suitable for adaptive control) by a transformation that employs the exact kernel (suitable for gain scheduling). A major simplification in the target system arises, with the perturbation due to the approximation shifting from the domain to the boundary condition. This results in a significant difference in the Lyapunov analysis, which nevertheless results in a guarantee of the stability being retained with the simplified approximation approach. The approach of approximating only the control gain function simplifies the operator being approximated and the training of its neural approximation, with an expected reduction in the neural network size. The price for the savings in approximation is paid through a somewhat more intricate Lyapunov analysis, in higher Sobolev spaces for some PDEs, as well as some restrictions on initial conditions that result from higher Sobolev spaces. It is essential to carefully consider the specific requirements and constraints of each problem to determine the most appropriate approach; indeed, recent works have demonstrated the successful application of both full-kernel and gain-only approaches in both adaptive control and gain scheduling contexts.

Keywords: Distributed parameter systems, Power plants, Modeling
Abstract: Heavy-duty gas engines are used in industrial applications as well as to supplement the supply of green energy to meet peak demand for electricity and heat. Compliance with increasingly stringent emissions regulations necessitates the use of exhaust gas aftertreatment systems. In this context, engines equipped with selective catalytic reduction (SCR) catalysts, where nitrogen oxides are reduced using urea, are of particular interest. The paper proposes a model-based approach for controlling the outlet NO_X concentration in such catalysts. A distributed-parameter model for both the thermal and the kinetic subsystems explicitly accounts for the transport phenomena within the catalyst. Feedforward control and observer-based state feedback are based on a lumped-parameter approximation of the model equations. Both the controller and the observer use a linear quadratic (LQ) approach to compute feedback and output injection gains, respectively. In order to achieve stationary accurate tracking despite model uncertainties, the observer is extended to estimate constant disturbances. Experimental validation on a real engine is performed to compare the proposed model-based controller (MBC) with a standard proportional–integral–derivative (PID) controller.

Keywords: Distributed parameter systems, Nonlinear system theory, Algebraic/geometric methods
Abstract: Exact time-dependent reachable sets for dynamic control systems are natural control theory extensions of the exact solutions of Cauchy problem for differential equations. The reachable sets are important in many applications, including optimal control, when analytical expressions are possible. The extensions are usually based on differential-geometric approaches that enable effective methods of analysis and design of nonlinear finite-dimensional control systems. Although a theoretical development for infinite-dimensional systems, including PDEs, requires significantly more advanced mathematics, the analytical technique is of a comparable level of complexity to the finite-dimensional case. The purpose of this paper is to present a set of examples illustrating techniques of obtaining exact analytical expressions of the reachable sets for important classes of linear and nonlinear PDEs: 1-st order, hyperbolic, and parabolic.

Keywords: Distributed parameter systems, Emerging control applications, Observers for linear systems
Abstract: This paper proposes a secure image encryption process using synchronization of chaotic hyperbolic partial differential equations (PDE) systems. Chaotic infinite-dimensional systems offer a promising alternative for secure communication since they allow for a larger key space, efficient computation, and more complex dynamics than finite-dimensional ones. Using an innovative invertible transform, we design a sensitive observer for an original chaotic system with increased encryption keys. It achieves finite-time stabilization of the error system. Therefore, using the observation data sent by the transmitter, the receiver can synchronize the chaotic observer PDE system. From the encrypted data and observer state, we can then reconstruct the original data. An analytical key sensitivity analysis is illustrated in simulations. Unlike classical approaches, the observer system must be the less robust: the more sensitive it is to a variation of the system's parameters, the more secure the cryptosystem will be. Different modulation strategies based on the synchronized chaotic states are proposed. Their robustness to basic crypto-attacks is assessed on a simple test case.

Keywords: Distributed parameter systems
Abstract: A generalized approach to the spatial discretization of the one-dimensional shallow water equations with moving boundary and an arbitrary cross-section is developed. Material-fixed coordinates are used to effectively cope with the moving boundary. The methodology involves the discretization of the Lagrangian on a material-fixed grid and the application of different quadrature schemes to derive finite-dimensional models. The proposed scheme explicitly considers mass conservation as an additional constraint, resulting in systems of semi-explicit differential-algebraic equations (DAEs). The particular structure of these DAEs depends on the chosen quadrature scheme and therefore requires slightly different methods for the numerical implementation. These methods are discussed on the basis of three examples, which are compared in simulation studies.

Keywords: Biological systems, Stochastic systems, Uncertain systems
Abstract: This work addresses a challenging agricultural control problem: to take into account environmental uncertainties (precipitation and solar radiance) in irrigation policies. To tackle these uncertainties, a Stochastic Model Predictive Control approach is implemented, wherein each type of uncertainty is addressed using distinct techniques tailored to effectively counteract it. Simulation experiments were conducted using real-world data spanning various types of days to validate the efficacy of the proposed approach. The results were benchmarked against other methods, showcasing the significant advantages of the proposed approach in terms of accuracy and robustness in agricultural irrigation control in the face of uncertainties. Therefore, this probabilistic approach also offers an effective solution to manage uncertainties and water resources, enhancing the productivity and sustainability of the sector.

Keywords: Optimal control, Neural networks, Biological systems
Abstract: Automatic control of greenhouse crop production is of great interest owing to the increasing energy and labor costs. In this work, we use two-level control, where the upper level generates suitable reference trajectories for states and control inputs based on day-ahead predictions. These references are tracked in the lower level using Nonlinear Model Predictive Control (NMPC). In order to apply NMPC, a model of the greenhouse dynamics is essential. However, the complex nature of the underlying model including discontinuities and nonlinearities results in intractable computational complexity and long sampling times. As a remedy, we employ NMPC as a data generator to learn the tracking control policy using deep neural networks. Then, the references are tracked using the trained Deep Neural Network (DNN) to reduce the computational burden. The efficiency of our approach under real-time disturbances is demonstrated by means of a simulation study.

Keywords: Neural networks, Machine learning, Biological systems
Abstract: Growing crops in Controlled Environment Agriculture (CEA) farms, such as greenhouses and vertical farms, can help in meeting the demands of urban centers and achieving climate goals. Recently, many advanced control techniques like Model Predictive Control (MPC) and its variants have been developed for energy-efficient operation and minimization of resource utilization. However, real-time implementation of these advanced strategies come along with certain computational hardware requirements, thus, increasing the operating costs. In this work, we propose to learn the MPC policy of a greenhouse control by means of a Deep Neural Network (DNN) in order to be implemented on a low-cost microcontroller. Additionally, we use a feedback law to reduce undesired quantization effects. The efficiency of our approach is exemplified by means of a simulation study for greenhouse control.

Keywords: Biological systems, Output feedback, Nonlinear system theory
Abstract: Plant–soil negative feedback (PSNF) is the rise in soil of negative conditions for plant performance induced by the plants themselves, limiting the full potential yield and thus representing a loss for the agricultural industry. It has been recently shown that detrimental effects the PSNF has on the growth of plant’s biomass can be mitigated by periodically intervening on the plant/soil system, for example by washing the soil. The periodic control inputs were computed by using an average model of the system and then applied in open-loop. In this paper we present two feedback control strategies, namely a PI and a MPC-based controllers, that, by adapting online the duty-cycle of the periodic control input, guarantee precise regulation of the biomass yield and at the same time robustness to unavoidable modeling errors and perturbations acting on the system. The performance of the proposed control strategies is then validated by means of extensive numerical simulations.

Keywords: Process control, Biological systems, Observers for nonlinear systems
Abstract: We study the problem of online state estimation during wine fermentation. This problem becomes relevant when trying to control the alcoholic fermentation process with a control law relying on our capacity to estimate the full state, but with partial measurements of the system (which is the case in an industrial framework). We focus on studying observability properties of an alcoholic fermentation model in wine-making conditions. We implement an algorithm of state estimation based on optimality principles on expanding time windows (Full Information Estimator). In order to test the developed algorithms, we compare the results obtained on simulated data as well as experimental data obtained in wine fermentations performed at a laboratory scale.

Keywords: Predictive control for linear systems
Abstract: This work addresses the problem associated with the variability of the parameters involved in irrigation control systems for real crops. Factors such as soil compaction, climatic variability or phenological state of the crops, among others, significantly influence the dynamics of these systems, challenging the implementation of model-based controllers in real use cases. In this context, an Adaptive Model Predictive Control scheme is proposed, which makes it possible to update the model by employing a recursive system identification. A comparison with a conventional predictive controller employing a constant model is made. The study is based on models identified from data collected in a production farm in Seville, Spain. The validation of the proposed strategy and the comparison between the adaptive MPC and the conventional MPC are performed by means of simulations. The results demonstrate the potential applicability and effectiveness of Adaptive MPC in real farming conditions.

Keywords: Cooperative autonomous systems, Concensus control and estimation, Autonomous systems
Abstract: We propose a distributed formation control algorithm augmented with channel awareness. We consider autonomous underwater vehicles (AUVs) that are able to communicate over an acoustic link using a Time Division Multiple Access (TDMA) protocol, and to measure the Signal-to-Noise Ratio (SNR) of incoming messages. Based on the measured SNR and packet loss, we endow them with a distributed formation control scheme that accounts for the time-varying nature of the acoustic communication channel. This scheme allows a network of N AUVs to follow a pre-determined, twice-differentiable path while adapting their formation. The size of the formation is dynamically scaled by a formation adaptation mechanism to stabilize the estimated packet loss probability at a desired level. A distributed packet loss estimator is then built on top of the same average consensus routines used by the formation control algorithm, and thus comes with a minimal communication overhead. We test the algorithm by means of high-fidelity simulators, and verify its efficacy in making the network of agents retain formation-wide communication capabilities in a range of cases.

Keywords: Cooperative autonomous systems, Stochastic systems, Stability of linear systems
Abstract: We consider a platooning control problem where the communication channels between vehicles are subject to coloured additive noises. Due to the stochastic nature of these channels, our analysis delves into examining the convergence of both the mean and variance of the vehicle tracking errors. We study the convergence as both time and number of vehicles grow unbounded. Our results include necessary and sufficient conditions for convergence and reveal that the colour of the noise does not impact the convergence characteristics of the error statistics, although it affects the values of the tracking error variances. Our findings offer insights into string stabilization. Numerical examples illustrate our results.

Keywords: Agents and autonomous systems, Decentralized control, Markov processes
Abstract: Fleet coordination and formation flight for Unmanned Aerial Vehicles (UAVs) are challenging and important problems that have received significant attention in recent years. In this paper, we propose a decentralized approach based on a Markov Decision Process (MDP) to ensure the control and formation flight of UAVs. We present a methodology for planning trajectories online that enables UAVs to maintain formation geometry while avoiding no-fly zones and reaching a desired goal area. By leveraging the dynamic model of fixed-wing UAV within the MDP formalization, we can provide optimal reference values for the UAV low-level controller. Furthermore, by harnessing the capabilities of MDPs to handle uncertainty, we can consider the behavior of nearby UAVs taking advantage of the predicted state vectors shared among them. Therefore, by exploiting the UAV dynamic model, and the estimation of the possible trajectories of the other UAVs, we can ensure collision-free and feasible actions for the UAV in a decentralized way. This approach has been validated and tested through simulations involving various scenarios.

Keywords: Cooperative autonomous systems, Cooperative control, System reconfiguration
Abstract: This paper proposes a method to recover from the failure or loss of a subset of agents in a distance-based formation problem, where the system is initially deployed forming a virtual shield embedded in the 3D space. First, a distributed algorithm is proposed to restore the topology, which is a Delaunay triangulation. After that, the nodes execute a distance-based distributed control law that considers adaptive target distances. These values are computed in parallel by the nodes, which try to reach an agreement with some constraints, given by the desired shield shape. The updating policy is based on events. The results are illustrated through simulation examples.

Keywords: Automotive, Predictive control for linear systems, Optimal control
Abstract: Trajectory planning at low speed applications can involve a large variety of different scenarios, including structured and unstructured environments, pedestrians, cyclists, etc. In this context, real-time planning is crucial to the dexterity of autonomous vehicles. In this paper, a real-time trajectory planner based on model predictive control (MPC) is proposed. Moreover, a dynamic obstacle avoidance and narrow passages control are designed in a scalable way through continuous updates of the optimization constraints. The performance of the proposed methodology are evaluated with a V-cycle model-based approach, through both model-in-the-loop (MIL) simulations, and real prototype in-vehicle experiments.

Keywords: Automotive, Sensor and signal fusion, Neural networks
Abstract: Advanced driver assistance systems are critically dependent on reliable and accurate information regarding a vehicles' driving state. For estimation of unknown quantities, model-based and learning-based methods exist, but both suffer from individual limitations. On the one hand, model-based estimation performance is often limited by the models' accuracy. On the other hand, learning-based estimators usually do not perform well in “unknown” conditions (bad generalization), which is particularly critical for semitrailers as their payload changes significantly in operation. To the best of the authors' knowledge, this work is the first to analyze the capability of state-of-the-art estimators for semitrailers to generalize across “unknown” loading states. Moreover, a novel hybrid Extended Kalman Filter (H-EKF) that takes advantage of accurate Artificial Neural Network (ANN) estimates while preserving reliable generalization capability is presented. It estimates the articulation angle between truck and semitrailer, lateral tire forces and the truck steering angle utilizing sensor data of a standard semitrailer only. An experimental comparison based on a full-scale truck-semitrailer combination indicates the superiority of the H-EKF compared to a state-of-the-art extended Kalman filter and an ANN estimator.

Keywords: Statistical learning, Adaptive control, Stochastic control
Abstract: The introductory tutorial has three objectives: (i) Provide the background and terminology needed to understand the subsequent talks. This includes fundamentals of Markov decision processes with special emphasis on linear quadratic systems, highlighting the conceptual challenges in obtaining a solution when the system model is not known. (ii) Provide an overview of fundamental ideas in machine learning/artificial intelligence, and how they translate to control settings. (iii) Provide an overview of the topics of the tutorial and help the audience understand how they fit in the larger context. The introductory tutorial will conclude with guidelines on important topics that are not covered in this tutorial.

Keywords: Statistical learning, Adaptive control, Predictive control for linear systems
Abstract: Fundamental performance limits, such as the Bode sensitivity integral, have long been a key component of the controls curriculum. In learning-based control, such fundamental performance limits typically take the form of information-theoretic lower bounds. In this talk, we give an overview of some of the key techniques used to prove such lower bounds, drawing on tools from estimation and information theory. We demonstrate how these tools can be applied to prove fundamental limits to learning the linear quadratic regulator and show how these allow us to relate the hardness of learning to control to classical controllability and observability notions.

Keywords: Statistical learning, Adaptive control, Predictive control for linear systems
Abstract: In perception-based control for complex systems, e.g., robotics, one practical question is: ``what makes a good state(-space) to accomplish certain control tasks?'' To better understand the empirical successes in addressing this question, we study the problem of learning state representations to control an unknown partially observable system. We pursue a ``direct latent model learning'' approach, where a dynamic model in some latent state space is learned by predicting quantities ``directly'' related to planning without reconstructing the observations. In particular, we focus on an intuitive ``cost-driven'' state representation learning method for solving Linear Quadratic Gaussian (LQG) control, one of the most fundamental partially observable control problems. As our main results, we establish finite-sample guarantees of finding a near-optimal state representation function and a near-optimal controller using the directly learned latent model. Our results underscore the value of predicting multi-step costs, an idea that is key to our theory, and notably also an idea that is known to be empirically valuable for learning state representations.

Keywords: Statistical learning, Adaptive control, Predictive control for linear systems
Abstract: Optimal control is an attractive framework for addressing complex control tasks, as it offers the promise of determining the best possible action given the dynamic and other constraints of the system. With some notable exceptions, however, optimal control problems are notoriously difficult to solve even for moderate size systems, let alone in the absence of models where decisions have to be made directly based on information collected directly from the system. Such problems are often addressed through reinforcement learning, integrating data sequentially into an estimate of the optimal policy, or the optimal Q-function from which the optimal policy can be extracted. It is known that many optimal control problems encoded as dynamic programs can equivalently be characterised through the solution of linear programs, where the decision variable is the unknown Q-function and the dynamics and stage cost are encoded through constraints. Replacing the (often infinite) linear program by a simpler (finite) counterpart leads to methods for approximating the solution of the original optimal control problem, in the spirit of Approximate Dynamic Programming. The data collected from the system is used in the constraints of the linear program, allowing us to solve the dynamic program approximately without relying on explicit knowledge of the dynamics or the cost function, offering an alternative to classical reinforcement learning. In this talk we will outline such an approach to approximate optimal control, how randomised optimisation that can be leveraged to derive bounds on the errors incurred, and methods for making this approach practically applicable.

Keywords: Statistical learning, Safety critical systems, Stochastic control
Abstract: This talk presents a theoretical framework for Inverse Reinforcement Learning (IRL) in constrained Markov decision processes (CMDPs). This problem is relevant in the case in which the expert policy must satisfy certain safety constraints. We extend prior results on reward identifiability and generalizability to both the constrained setting and a more general class of regularizations. In the CMDP setting, we characterize the set of rewards for which a given expert is optimal. And we show that generalizability to new transition laws and safety constraints is only possible if the expert’s reward is recovered up to a constant. Furthermore, we derive a finite sample guarantee for the suboptimality of the learned rewards. The proof techniques, based on convex analysis, unify and generalize past work. We conclude with simulation and experimental results, and with open directions.

Keywords: Control education
Abstract: Most of us think, that we act and decide very rationally. But we all have so-called unconscious biases. What are unconscious biases? What role do they play in our daily lives as researchers? How do they influence us, when we work together, integrate new team members, or in leadership situations? By understanding how biases can impact our interactions with others, we become more equipped to make robust decisions and objective judgments. The following aspects form the foundation of the input session: the definition and impact of unconscious biases, different types of unconscious biases and how to avoid or mitigate them. After the input session, we will work on several cases and discuss, how to improve meetings, interview situations or appraisal talks.

Keywords: Control education
Abstract: The world of control systems conferences can be a maze of presentations and networking events, especially for newcomers. Don't worry, the future of our field depends on you, the next generation of researchers! That's why the IEEE Control Systems Society (CSS) created NextCOM, a committee of young researchers to make the conference experience for students and early-career professionals like you amazing. At the ECC this year, NextCom is hosting a session at ECC on June 26th, 12:15-13:00, to help you navigate the conference with ease. Learn valuable tips on: Research presentation tactics: Get the most out of all those talks! Networking like a pro: Make connections that will boost your career. Plus, we'll have a fun controls-themed charades game to break the ice! NextCOM is here to make your ECC experience enriching and enjoyable.

Keywords: Emerging control theory, Predictive control for nonlinear systems
Abstract: In this talk I will offer a personal view on the latest developments in the area of so-called data-driven predictive control. Particular emphasis shall be given to the highly debated issue of direct vs. indirect methods as well as on the role models play, explicitly or implicitly, in control design, thus providing a connection with system identification. In doing so, I will establish a probabilistic Bayesian framework for data-driven predictive control that encompasses most existing approaches such as DeePC and Subspace Predictive Control as special cases. I shall argue that the probabilistic system-theoretic viewpoint is the cornerstone for enabling on-line data-driven closed-loop predictive control.

Keywords: Biological systems, Uncertain systems
Abstract: Natural systems across biology, ecology and epidemiology can be seen as dynamical networks, composed of multiple subsystems that interact according to an interconnection topology. Despite their large scale and complexity, most systems in nature exhibit extraordinary robustness: fundamental properties and qualitative behaviours that are crucial for survival are preserved even in the presence of huge parameter variations and environmental fluctuations. We discuss systems-and-control approaches to unravel the key features of biological systems and find the roots of their amazing robustness, by identifying properties and emerging behaviours that exclusively depend on the system structure, regardless of parameter values. We also consider epidemiological systems along with control approaches to cope with uncertain parameter values and optimally curb the contagion.

Keywords: Adaptive control, Autonomous robots, Optimal control
Abstract: This paper proposes an adaptive trajectory tracking control algorithm for autonomous ground vehicles. The nonlinear vehicle dynamics are decoupled into two subsystems corresponding to the longitudinal and lateral motions. Each subsystem is augmented with a Gaussian Process to compensate for modeling errors and external disturbances. Based on the augmented subsystems, adaptive control algorithms are synthesized. To give a mathematically correct performance measure, the induced L2-gain of the nonlinear closed-loop system is computed. The efficiency of the learning-based control method is demonstrated on a high-fidelity physical simulator using a digital twin model of the 1/10 scale F1TENTH vehicle platform.

Keywords: Adaptive control, Autonomous systems, Autonomous robots
Abstract: In this paper we validate, including experimentally, the effectiveness of a recent theoretical developments made by our group on control-affine Extremum Seeking Control (ESC) systems. In particular, our validation is concerned with the problem of source seeking by a mobile robot to the unknown source of a scalar signal (e.g., light). Our recent theoretical results made it possible to estimate the gradient of the unknown objective function (i.e., the scalar signal) incorporated in the ESC and use such information to apply an adaptation law which attenuates the oscillations of the ESC system while converging to the extremum (i.e., source). Based on our previous results, we propose here an amended design of the simple single-integrator control-affine structure known in ESC literature and show that it can functions effectively to achieve a model-free, real-time source seeking of light with attenuated oscillations using only local measurements of the light intensity. Results imply that the proposed design has significant potential as it also demonstrated much better convergence rate. We hope this paper encourages expansion of the proposed design in other fields, problems and experiments.

Keywords: Adaptive control, Linear systems, LMI's/BMI's/SOS's
Abstract: Iterative learning control emerged from the problem area of increasing the accuracy of finite-duration repetitive operations performed by robots, often termed trials. The ILC laws use past trial information to adjust the current trial's control signal. Most often, only data from the previous trial is used. A higher-order ILC law uses information from several previous trials. Recently, interest in these laws has increased in the literature with, in particular, robotic additive manufacturing problems. This paper develops a new higher-order ILC design for discrete linear uncertain systems that makes greater use of information generated over previous trials. An example using a model developed from measured frequency response data from a laboratory testbed illustrates the new design.

Keywords: Adaptive control, Lyapunov methods
Abstract: In this contribution we extend the classical update law of the certainty equivalence based (indirect) model reference adaptive control (MRAC) with an additional feedthrough error term. We use a linear state-space representation of the parameter estimation error in order to construct a Lyapunov function. Without unstructured uncertainties, we can show asymptotic stability of the closed loop. Even for relatively high adaptation gains the algorithm can reduce the oscillations of the state variables. In case of unstructured input uncertainties, the additional term reduces the model following error drastically without increasing the control signal for low frequency unstructured uncertainties. For the simplified pure linear case we analyze the effect of the additional proportional error feedthrough in the adaptation law showing its similarity to a disturbance observer. As a simulation example the wing-rock dynamics are reconsidered.

Keywords: Adaptive control, Optimal control, Predictive control for nonlinear systems
Abstract: In this paper, we contribute the Parallel Ensemble foreCAst coNtrol (PECAN) algorithm to enhance multi-objective water systems control through the integration of Ensemble Forecast and data-driven control techniques. This integration allows evolving parallel system simulations for each forecast ensemble member to maximize the benefit provided by the probabilistic forecasts. To avoid potential overfitting and ensure the generalization capabilities of the designed solutions, we also implement a Blocked K-Fold cross-validation. Testing on the Lake Como water reservoir system shows that PECAN improves the controller performance by 8.2% with respect to traditional methods relying solely on forecast ensemble averages and by 26.8% over approaches that do not use any forecast. These results highlight the benefits of ensemble-based techniques for controlling water systems under highly variable hydroclimatic conditions.

Keywords: Emerging control applications, Adaptive control, Feedback linearization
Abstract: Online advertising is typically implemented via real-time bidding, and advertising campaigns are then defined as extremely high-dimensional optimization problems. To solve these problems in light of large scale and significant uncertainties, the optimization problems are modularized in a way that makes feedback control a critical component of the solution. The control problem, however, is challenging due to plant uncertainties, nonlinearities, time-variance, and noise. An Oracle would define the control signal in terms of bid price adjustments only; however, we propose the introduction of a companion throttling control signal that creates a useful plant linearity. In this paper, the control problem is redefined in such a way that the linearity is exploited for improved feedback control. A dual lever control algorithm is designed and evaluated in simulations, with promising results.

Keywords: Optimal control, Linear systems, Stochastic systems
Abstract: In this study, we introduce numerical methods for discretizing continuous-time linear-quadratic optimal control problems (LQ-OCPs). The discretization of continuous-time LQ-OCPs is formulated into differential equation systems, and we can obtain the discrete equivalent by solving these systems. We present the ordinary differential equation (ODE), matrix exponential, and a novel step-doubling method for the discretization of LQ-OCPs. Utilizing Euler-Maruyama discretization with a fine step, we reformulate the costs of continuous-time stochastic LQ-OCPs into a quadratic form, and show that the stochastic cost follows the Chi-square distribution. In the numerical experiment, we test and compare the proposed numerical methods. The results ensure that the discrete-time LQ-OCP derived using the proposed numerical methods is equivalent to the original problem.

Keywords: Optimal control, Optimization algorithms, Optimization
Abstract: This work proposes an efficient treatment of continuous-time optimal control problems with long horizons and nonlinear least-squares costs. In particular, we present the Gauss-Newton Runge-Kutta (GNRK) integrator which provides a high-order cost integration. Crucially, the Hessian of the cost terms required within an SQP-type algorithm is approximated with a Gauss-Newton Hessian. Moreover, L2 penalty formulations for constraints are shown to be particularly effective for optimization with GNRK. An efficient implementation of GNRK is provided in the open-source software framework acados. We demonstrate the effectiveness of the proposed approach and its implementation on an illustrative example showing a reduction of relative suboptimality by a factor greater than 10 while increasing the runtime by only 10%.

Keywords: Optimal control, Output feedback, Optimization
Abstract: Pontryagin's Maximum Principle is one of two famous results towards characterization of solutions of optimal control problems. Due to a shift in the desire from maximizing gain or profit to minimizing costs it is nowadays more often and henceforth in this publication referred to as Minimum Principle. The theory behind it utilizes variational calculus and provides necessary conditions. Many versions of the minimum principle exist. Among them variations exist that consider constraints, sufficient conditions, whereas other realize time-discrete formulations of the problem. Nevertheless, so far a generalization towards output-feedback systems is not found in the literature. The goal of this publication is to extend the existing theory via including an output function and variables. Additionally, general equality and inequality constraints as well as terminal constraints and sufficient conditions will be incorporated. The original Pontryagin's Minimum Principle can then be seen as a special case of the derived criteria.

Keywords: Optimal control, Delay systems, Linear systems
Abstract: In this paper, we investigate the design of optimal spatially distributed controllers for a reaction-diffusion process evolving over the real line under the assumption that the controller receives state measurements from different spatial locations with non-negligible delays. For the class of proportional spatially invariant state feedback controllers, such a design problem leads to a feedback gain optimization for a spatially distributed delay system. As is well known, the problem becomes decoupled in the spatial frequency domain. We show that the spatial locality of optimal feedback gains is affected by the diffusion and reaction coefficients together with the amount of communication delay, which causes a sharp flattening of the control kernel. In the expensive control regime, we are able to solve for the optimal controller analytically, using which we offer some practical design guidelines.

Keywords: Optimal control, Delay systems, Stochastic systems
Abstract: For multi-variable static quadratic map, we present a time-delay approach to gradient-based extremum seeking (ES) with measurement noise, and provide a mean-square exponential ultimate boundedness (MSEUB) analysis. We consider the uncertain map where the Hessian matrix H has a nominal known part and norm-bounded uncertainty, the extremum point belongs to a known ball, and the extremum value to a known interval. By applying a time-delay approach to the resulting stochastic ES system, we arrive at the neutral type time-delay system with stochastic perturbations. We further present the latter system as a retarded one and employ the variation of constants formula for the MSEUB analysis. Under the assumption that the upper bound of the 6th moment of the estimation error is a known arbitrarily large constant L, explicit condition in terms of simple scalar inequality depending on the bound L, tuning parameters and the intensity of measurement noise is established to guarantee the MSEUB analysis of the ES control systems. Example from the literature illustrates the efficiency of the new approach.

Keywords: Optimal control, Stochastic control, Identification
Abstract: In this paper, we propose a new algorithm to solve the Inverse Stochastic Optimal Control (ISOC) problem of the linear-quadratic sensorimotor (LQS) control model. The LQS model represents the current state-of-the-art in describing goal-directed human movements. The ISOC problem aims at determining the cost function and noise scaling matrices of the LQS model from measurement data since both parameter types influence the statistical moments predicted by the model and are unknown in practice. We prove global convergence for our new algorithm and at a numerical example, validate the theoretical assumptions of our method. By comprehensive simulations, the influence of the tuning parameters of our algorithm on convergence behavior and computation time is analyzed. The new algorithm computes ISOC solutions nearly 33 times faster than the single previously existing ISOC algorithm.

Keywords: Large-scale systems, Network analysis and control, Safety critical systems
Abstract: This paper develops a compositional framework for formal safety verification of an interconnected network comprised of a countably infinite number of discrete-time nonlinear subsystems with unknown mathematical models. Our proposed scheme involved subdividing the infinite network problem into individual subsystems, wherein the safety concept is modelled through a robust optimization program (ROP) via a notion of local-barrier certificates (L-BC). To address the difficulties associated with solving the ROP directly, primarily due to the absence of a mathematical model, we gather finite data from subsystem trajectories and leverage them to provide a scenario optimization program (SOP). We proceed with solving the resulted SOP and construct a local-barrier certificate for each unknown subsystem with a guarantee of correctness. Finally, in accordance with some small-gain conditions, we construct a global-barrier certificate (G-BC) derived from individual local certificates of subsystems, thus guaranteeing the safety of the infinite network within infinite time horizons. The practicality of our compositional findings becomes evident through a vehicle platooning scenario, characterized by a countably infinite number of vehicles with a single leader and an unlimited number of followers.

Keywords: Observers for linear systems, Linear systems, Modeling
Abstract: In this paper we propose a data-driven approach to the design of reduced-order unknown-input observers (rUIOs). We first recall the model-based solution, by assuming a problem set-up slightly different from those traditionally adopted in the literature, in order to be able to easily adapt it to the data-driven scenario. Necessary and sufficient conditions for the existence of a reduced-order unknown-input observer, whose matrices can be derived from a sufficiently rich set of collected historical data, are first derived and then proved to be equivalent to the ones obtained in the model-based framework. Finally, a numerical example is presented, to validate the effectiveness of the proposed scheme.

Keywords: V&V of control algorithms, Machine learning, Statistical learning
Abstract: Verification of uncertain, complex dynamical systems is crucial in the modern-day world. An increasingly common method to verify complex logic specifications for dynamical systems involves symbolic abstractions: simpler, finite-state models whose behaviour mimics the one of the systems of interest. By sampling trajectories of the concrete, unknown system and via robust analysis, we build a data-driven abstraction, related to the underlying model through a probabilistic behavioural inclusion relation. As the distribution from which the trajectories are drawn is unknown, we adopt two distinct distribution-free theories, namely scenario optimization and conformal prediction. We compare and discuss the differences between the two approaches in terms of the type of guarantees that they are able to provide. Furthermore, via experimental benchmarks we outline the efficiency of the two methods with respect to the number of samples available and the tightness of the guarantees.

Keywords: Stochastic systems, Safety critical systems, Markov processes
Abstract: Finite-state abstractions are widely studied for the automated synthesis of correct-by-construction controllers for stochastic dynamical systems. However, existing abstraction methods often lead to prohibitively large finite-state models. To address this issue, we propose a novel abstraction scheme for stochastic linear systems that exploits the system's stability to obtain significantly smaller abstract models. As a unique feature, we first stabilize the open-loop dynamics using a linear feedback gain. We then use a model-based approach to abstract a known part of the stabilized dynamics while using a data-driven method to account for the stochastic uncertainty. We formalize abstractions as Markov decision processes (MDPs) with intervals of transition probabilities. By stabilizing the dynamics, we can further constrain the control input modeled in the abstraction, which leads to smaller abstract models while retaining the correctness of controllers. Moreover, when the stabilizing feedback controller is aligned with the property of interest, then a good trade-off is achieved between the reduction in the abstraction size and the performance loss. The experiments show that our approach can reduce the size of the graph of abstractions by up to 90% with negligible performance loss.

Keywords: V&V of control algorithms, Safety critical systems, Neural networks
Abstract: Control Barrier Functions (CBFs) that provide formal safety guarantees have been widely used for safety-critical systems. However, it is non-trivial to design a CBF. Utilizing neural networks as CBFs has shown great success, but it necessitates their certification as CBFs. In this work, we leverage bound propagation techniques and the Branch-and-Bound scheme to efficiently verify that a neural network satisfies the conditions to be a CBF over the continuous state space. To accelerate training, we further present a framework that embeds the verification scheme into the training loop to synthesize and verify a neural CBF simultaneously. In particular, we employ the verification scheme to identify partitions of the state space that are not guaranteed to satisfy the CBF conditions and expand the training dataset by incorporating additional data from these partitions. The neural network is then optimized using the augmented dataset to meet the CBF conditions. We show that for a non-linear control-affine system, our framework can efficiently certify a neural network as a CBF and render a larger safe set than state-of-the-art neural CBF works. We further employ our learned neural CBF to derive a safe controller to illustrate the practical use of our framework.

Keywords: Safety critical systems, Autonomous systems, Linear systems
Abstract: In this paper, we introduce a hybrid zonotope-based approach for formally verifying the behavior of autonomous systems operating under Linear Temporal Logic (LTL) specifications. In particular, we formally verify the LTL formula by constructing temporal logic trees (TLT)s via backward reachability analysis (BRA). In previous works, TLTs are predominantly constructed with either highly general and computationally intensive level set-based BRA or simplistic and computationally efficient polytope-based BRA. In this work, we instead propose the construction of TLTs using hybrid zonotope-based BRA. By using hybrid zonotopes, we show that we are able to formally verify LTL specifications in a computationally efficient manner while still being able to represent complex geometries that are often present when deploying autonomous systems, such as non-convex, disjoint sets. Moreover, we evaluate our approach on a parking example, providing preliminary indications of how hybrid zonotopes facilitate computationally efficient formal verification of LTL specifications in environments that naturally lead to non-convex, disjoint geometries.

Keywords: Robotics, Sliding mode control, Uncertain systems
Abstract: In this paper, we propose a strategy for performing human-robot ergonomic handover in the case of partial knowledge of the robot dynamics and in absence of assumptions about the shape or mass of the object passed. In particular, we propose a strategy for generating the reference for the manipulator and a Deep Neural Network based Integral Sliding Mode control with Adaptive discontinuous gain, in the paper referred to as DNN-AISM. The DNN weights are adapted on-line according to update laws derived directly from the theoretical analysis, without relying on previously collected data. The proposal is experimentally assessed relying on a Franka Emika Panda robot and on an Xsens MTw IMU sensor, producing highly satisfactory results.

Keywords: Robotics, Computational methods, Biological systems
Abstract: Tendon-based anthropomorphic robotic hand implementations generally lack the ability to measure the joint angles, as the encoder installed for directly measuring the joint angle may compromise the dexterity of the hand. In this paper, we present a computational approach to estimate the joint positions of the hand using the measured tendon displacements and tensions. First, we introduce an efficient framework for the kinematic description of an anthropomorphic hand based on Denavit-Hartenberg's description. Then, we use a simplified tendon model to derive a system of nonlinear equations for the joint positions, which is finally solved by a gradient-based optimization solver. We used the model to control the hand along commanded gestures by feeding back the estimated joint angles through the moment arm Jacobian. The effectiveness and limitations of this method is illustrated in MuJoCo simulation environment on the Anatomically Correct Biomechatronic Hand having a 5 and 6 degrees of freedom kinematic models for the long fingers and the thumb, respectively.

Keywords: Robotics, Game theoretical methods, Agents and autonomous systems
Abstract: This paper considers a target attack defense scenario where the attacker aims to reach the target while avoiding the defender, and the defender wants to protect the target by intercepting the attacker. While most of the prior works assume the target to be known to the defender, in our settings the defender has access to only the current states of the attacker at any instant, and is not aware of the target coordinates. A novel approach is proposed in this work to leverage past data of the attacker states to construct a future trajectory of the attacker. The defender deploys a model predictive control scheme to minimize the discrepancy between its own future trajectory and the predicted future trajectory of the attacker. Simulation results show that use of estimated future trajectories helps in more effective protection of the target compared to when only current state of the attacker is used. The effectiveness of the proposed approach is also highlighted in the presence of obstacles.

Keywords: Robotics, Sensor and signal fusion, Hybrid systems
Abstract: In the field of autonomous legged robotics, accurate state estimation is crucial for control and planning. While traditional methods suffice for fully-actuated platforms, under-actuated systems face challenges due to sensory limitations and uncertainties. This paper presents a novel methodology for state estimation and phase prediction, integrating a torque-actuated spring-mass model with limited sensors using a multiple-hypotheses extended Kalman filter. Within this estimation framework, the optimal estimate is determined at each iteration by evaluating the likelihood functions associated with two distinct phase hypotheses, either stance or flight. We evaluate different sensor and motion model combinations, showing that our method achieves precise state and phase estimation even without advanced sensors for compliant and under-actuated platforms.

Keywords: Robotics, Modeling, Optimization algorithms
Abstract: The quadruped robots represent a rapidly growing field in the industrial world due to their various applications in supporting everyday and hazardous human activities. Despite significant research contributions, gait optimization remains a challenging task. Typically, gait patterns are inspired by mimicking nature but fall short of replicating the flexibility and articulated adaptability observed in animals. This paper introduces a new method for optimizing gait and designing quadruped robots offline. This approach takes into account the dynamics of the entire system, including leg masses, by employing genetic algorithms and nonlinear programming techniques. The results of this new technique enable the exploration of gait patterns that can accommodate the rigidity of robots while minimizing the energy cost of locomotion.

Keywords: Robotics, Optimization
Abstract: In variable impedance control the desired impedance parameters of a robot manipulator are adjusted in real-time. It has been observed that time varying impedance parameters can lead to a non-passive interaction and even destabilize the system. In this paper we consider a compliance controller, where the desired impedance is specified via a time-varying stiffness and damping. We propose to compute the controller gains from an additional online optimization instead of applying the desired time-varying stiffness and damping references directly. In this way the effect of the time-varying impedance parameters is evaluated over a finite time horizon. Together with a terminal constraint on the state this controller formulation aims to avoid a destabilization of the system due to the impedance variation. The proposed method is validated for free movement in two different simulations as well as for physical interaction on a Franka Research 3.

Keywords: Stability of nonlinear systems, Autonomous systems
Abstract: Path-following control is an alternative technique to trajectory tracking, with superior performance characteristics when applied to a class of nonholonomic systems, especially relevant in the case of unmanned vehicles. In this work, we study a variant of a well-known guidance logic path-following control and analyze its stability. In particular, we show that the studied method not only maintains the stabilizing properties in a neighborhood of the path, but significantly enlarges the domain of attraction of the control law, both when following a straight or a curved path.

Keywords: Nonlinear system theory, Complex systems, Nonlinear system identification
Abstract: The Koopman operator framework is a promising direction of analysis and synthesis of systems with nonlinear dynamics based on (linear) Koopman operators. In this paper, we address the resolvent of a Koopman operator for a nonlinear autonomous discrete-time system, which we call the Koopman resolvent, and its identification problem. First, we show that for the nonlinear system with a scalar-valued output, the z-transform of the output is represented by the action of Koopman resolvent. Second, we describe an identification method of the Koopman resolvent directly from time-series data of the output, in which we estimate parameters of the resolvent as well as poles and residues of the z-transform of the output. By combining the so-called frequency-domain Prony method with the Vandermonde-Cauchy form in the Dynamic Mode Decomposition (DMD), we propose the method which we call the frequency-domain DMD, in which all the unknowns can be estimated in the frequency domain.

Keywords: Nonlinear system theory, Network analysis and control
Abstract: New results on continuous time nonlinear consensus under varying topology are presented. The results are proved utilizing non Lyapunov based methods, i.e., the Hilbert metric, showing the possibility of further investigation of Hilbert metric for consensus and synchronization problems.

Keywords: Nonlinear system theory, Robust control, Stability of nonlinear systems
Abstract: The class of parameter-varying generalized Persidskii systems is introduced. For models within this class, characterized by nonlinearities satisfying the passivity property, the conditions for (integral) input-to-state stability are proposed. These conditions are established using both, parameter-dependent and parameter-independent, Lyapunov functions. To formulate these conditions, parameterized matrix inequalities are used, which can be reduced into linear ones under additional assumptions concerning the model's dependence on scheduling variables. The efficiency of these stability conditions is illustrated through a numerical example.

Keywords: Nonlinear system theory, Signal processing, Emerging control theory
Abstract: This paper investigates the issue of extended Kalman filtering (EKF) for graphical nonlinear systems. For the signals that have a nonlinear relationship with the graph Laplacian matrix, we introduce a variance-constrained (VC)-EKF, leveraging the graph Fourier transform (GFT), aimed at enhancing the filtering performance. In this paper, the GFT is applied for system variables and the updated estimate is designed with a diagonal gain matrix for the transformed system, then the higher-order terms are introduced into the predicted error and the updated error, and the diagonal gain matrix can be acquired through the solution of two Riccati-like equations. The advantages of the GFT-VC-EKF are that, the gain matrix represented by a diagonal matrix enables the node signal to be updated independently, thus reducing the iterative cumulative error, and the introduction of higher-order terms makes it possible to compensate for the linearization error caused by the fact that the EKF containing only first-order Taylor expansion term. A simulation example on a power system proves the superiority of the proposed filter.

Keywords: Stability of nonlinear systems, Nonlinear system identification, Network analysis and control
Abstract: This paper characterizes a new parametrization of nonlinear networked incrementally L2-bounded operators in discrete time. The distinctive novelty is that our parametrization is free - that is, a sparse large-scale operator with bounded incremental L2 gain is obtained for any choice of the real values of our parameters. This property allows one to freely search over optimal parameters via unconstrained gradient descent, enabling direct applications in large-scale optimal control and system identification. Further, we can embed prior knowledge about the interconnection topology and stability properties of the system directly into the large-scale distributed operator we design. Our approach is extremely general in that it can seamlessly encapsulate and interconnect state-of-the-art Neural Network (NN) parametrizations of stable dynamical systems. To demonstrate the effectiveness of this approach, we provide a simulation example showcasing the identification of a networked nonlinear system. The results underscore the superiority of our free parametrizations over standard NN-based identification methods where a prior over the system topology and local stability properties are not enforced.

Keywords: Emerging control applications, Fluid flow systems, Output feedback
Abstract: This work addresses the problem of controlling the local aerodynamic lift of a wind turbine blade taking into account disturbances caused by turbulent perturbations at the blade scale. This work deals with the study of a model-free based control algorithm implemented in the high fidelity simulated environment ISIS-CFD, where the controller acts at the level of the blade section to track the lift to a desired reference. Numerical experiments have been conducted in order to highlight some properties of the aerodynamic closed-loop system under several operating conditions.

Keywords: Energy systems, Stability of linear systems, Identification
Abstract: This contribution analyses the influence of a passive inerter-based network on the 5MW NREL FOWT with a barge-type foundation when the system is subjected to the specific problem of self-induced oscillation. The concept of implementation of an inerter-based network combined with a standard tuned mass damper (TMD) in the nacelle is here presented with the objective of demonstrating the effectiveness of the introduced network in satisfactorily reducing the amplitude of the self-induced oscillations in comparison to the case of the FOWT system fitted with only a TMD. Major improvements in the overall structural stability were found with a suppression rate ranging from 49% to up to 86%, in the wind velocities of interest, when the phenomenon may occur. Moreover, it was shown that the inerter installed in the nacelle reduces the overall natural frequency of the system by over 56%.

Keywords: Optimal control, Process control, Energy systems
Abstract: In order to mitigate periodic blade loads in wind turbines, recent research has analyzed different Individual Pitch Control (IPC) approaches, which typically use the multi-blade coordinate (MBC) transformation. Some of these studies show that the introduction of an additional tuning parameter in the MBC, namely the azimuth offset, helps to decouple the nonrotating axes in the MBC transformation and enhances the IPC performance. However, these improvements have been studied without considering the increased control effort performed by the pitch signal, which is the main negative side effect of the IPC. This work addresses this trade-off between pitch signal effort and blade fatigue reduction for IPC applied to a wind turbine operating in the full load region. Here, two IPC schemes, with and without additional azimuth offset, are designed and applied to a 15 MW monopile offshore wind turbine simulated with OpenFAST software. The optimal tuning of the IPC parameters is performed by means of a multi-objective optimization solved by genetic algorithms. The optimization procedure minimizes two objective functions related to pitch signal effort and blade fatigue load. The resulting Pareto fronts show a range of optimal solutions for each IPC scheme. The selected optimal solution for IPC with azimuth offset compared to the optimal solution for IPC without offset achieves improvements of more than 10% in blade load reduction maintaining similar pitch signal effort.

Keywords: Modeling, Aerospace
Abstract: MegAWES is a reference design and simulation framework for ground-generation, fixed-wing airborne wind energy systems with a nominal power output of 3 MW. In the original design, the winch size of MegAWES was based on a smaller system and needed to be scaled up. However, there were no available methods to select an appropriate size for the winch. Additionally, while it has been hypothesized that the size of the winch has a significant effect on the dynamics of the overall system for ground-generation concepts, this effect has not been quantified. In this work, we first analyze the effects of the winch size on the system dynamics using a linearized model. Second, we present a method to find the upper bound for the size of the winch based on a selected maximum tether force overshoot during nominal operation. Third, we apply this method to find an upper bound for the winch size for the MegAWES reference design. Using the nonlinear MegAWES simulation framework, we validated this upper bound. At the upper bound, the system accurately tracked the reference tether force without overshoot and when exceeding our upper bound, the tether force response was oscillatory and overshot its ideal value.

Keywords: UAV's, Energy systems
Abstract: We consider a motorized aircraft tethered to a central anchorage point in a configuration similar to a control line model airplane. For this system, we address the problem of automatic take-off and landing (ATOL) with a circular path, {whose} center and radius are defined by the anchorage point and the tether length, respectively. We propose a hierarchical control architecture for ATOL and discuss the controllers designed for each control layer and for each of the flight phases. Simulation results are reported, showing the viability of the approach, but also showing the limitations on the maximum altitude attainable with a fixed-tether length. The tethered aircraft and the proposed ATOL control architecture are to be used in an Airborne Wind Energy System.

Keywords: Energy systems
Abstract: This study investigates deep offshore, pumping Airborne Wind Energy systems, focusing on the kite-platform interaction. The considered system includes a 360 m2 soft-wing kite, connected by a tether to a winch installed on a 10-meter-deep spar with four mooring lines. Wind power is converted into electricity with a feedback controlled periodic trajectory of the kite and corresponding reeling motion of the tether. An analysis of the mutual influence between the platform and the kite dynamics, with different wave regimes, reveals a rather small sensitivity of the flight pattern to the platform oscillations; on the other hand, the frequency of tether force oscillations can be close to the platform resonance peaks, resulting in possible increased fatigue loads and damage of the floating and submerged components. A control design procedure is then proposed to avoid this problem, acting on the kite path planner. Simulation results confirm the effectiveness of the approach.

Keywords: Transportation systems, Optimization algorithms, Traffic control
Abstract: We propose an optimization method to place elec- tric vehicle charging stations to minimize total travel time, thereby minimizing additional congestion and detours caused by the chargers. For a tractable optimization scheme, we frame the drivers’ route choices as a congestion game that allows us to find equilibrium flows for each candidate set of locations. Our contribution has two primary components. First, we refine the modeling of driver cost functions to account for charging needs as well as travel time, and introduce different agent types based on their unique valuations of charging benefits. Second, we address the exponential growth of the search space of charger locations with a greedy optimization approach. We demonstrate with numerical experiments that: (i) the congestion game formulation allows us to efficiently compute equilibrium flows for each candidate charger placement; (ii) the greedy approach can closely approximate the optimal selection.

Keywords: Automotive, Autonomous systems
Abstract: This paper proposes a technology for road-shape monitoring using crowdsourced vehicle data. The technology uses vehicle measurements and a dynamics model in a statistical estimation framework with a kernel model approximating the road shape. The rationale for considering the vehicle dynamics for road monitoring is that the same road yields different measurements/oscillations for different vehicle types. Next, this paper shows how to use such estimated road shape and vehicle dynamics for semi-active suspension control with the objective to improve passenger comfort. Results using the high-fidelity simulator CarSim show that the proposed technology (i) only needs a few vehicles for estimating the road shape, (ii) can improve passenger comfort by semi-active suspension control, and (iii) is robust to model mismatch indicating the applicability to a real-world scenario.

Keywords: Automotive, Autonomous systems, Cooperative autonomous systems
Abstract: In this paper, a novel non-linear distributed MPC coordination scheme for autonomous vehicles based on the Optimality Condition Decomposition (OCD) algorithm is proposed. As result, a local planner for each vehicle is obtained via the OCD application to the MPC coordination scheme is obtained for each vehicle, providing realistic paths to be executed in a collaborative manner. The proposed approach is able to determine paths that consider both individual goals and the environment so that agents can collaborate to safely navigate the studied environment. A case study in a ROS simulation environment is used to assess the validity of the proposed approach for real-time implementation.

Keywords: Automotive, Neural networks, Emerging control applications
Abstract: The use of predictive energy management systems can improve the efficiency of multi-energy storage vehicles. However, current systems have limitations, such as short prediction horizons, the requirement for input data that is not publicly available, or the training of the Neural Networks on the routes on which the prediction is made. To overcome these challenges, this paper introduces a novel method for long-horizon energy prediction, utilizing readily available data such as route geometry and traffic information. Our study compares Convolutional Neural Networks (CNNs), Gated Recurrent Units (GRUs), and Transformer Networks optimized using the Asynchronous Successive Halving Algorithm (ASHA). The models were evaluated in a simulated environment using the Simulation of Urban MObility (SUMO) and further tested on real-world driving data, demonstrating that we are able to predict the consumed energy over a 45km stretch of highway with a median RMSE of 0.018 kWh/km for practical application. The energy prediction developed in this study has the potential to enhance predictive energy management systems, thereby optimizing energy usage and contributing to CO2 emission reduction.

Keywords: Automotive, Observers for nonlinear systems, Optimization
Abstract: Environmental factors such as rain, snow, or ice significantly impact road friction, which correlates strongly with traffic safety. This work uses a sensitivity-based moving horizon estimator to gauge the maximum road friction coefficient online. The adopted friction estimation strategy combines the estimated rack force based on a linear steering system model with a nonlinear two-degree-of-freedom (DoF) single-track vehicle model. The parametric output sensitivity is monitored and integrated within the moving horizon estimation (MHE) framework to prevent an arbitrary estimate in the absence of sufficient excitation. The method is evaluated by simulations and experiments under various road conditions. The results validate the proposed strategy and demonstrate its capability to reliably estimate the maximum road friction coefficient.

Keywords: Automotive, Autonomous systems
Abstract: This paper addresses the joint vehicle-state and road-map estimation based on global navigation satellite system (GNSS), camera, steering-wheel sensing, wheel-speed sensors, and a prior map. Because prior maps, e.g., generated from mobile mapping systems, are updated infrequently and do not capture high-frequency events such as road construction, we include the map parameters in the estimation problem. Both GNSS and camera measurements, such as lane-mark measurements, have noise characteristics that vary in time. To adapt to the changing noise levels and hence improve positioning performance, we combine the sensor information in a noise-adaptive variational Bayes Kalman filter to jointly estimate the vehicle state, the parameter vector of the map, and the measurement noise. Simulation results indicate that the method can accurately adjust the measurement noise to the environmental conditions and thereby correct for errors in the prior map while providing accurate vehicle positioning.

Keywords: Distributed estimation over sensor nets, Concensus control and estimation, Large-scale systems
Abstract: Distributed state estimation is a relevant research topic due to its application opportunities in different fields, such as multi-robot cooperation and control of large-scale networked systems. In addition, event-triggering mechanisms have been studied in recent years to reduce communication between network nodes without significantly compromising the desired behavior. In this work, we contribute a distributed algorithm to estimate the state of a stochastic system under event-triggered communication. The proposal uses consensus on the state estimates, and it takes advantage of the asymptotic form of the well-known Kalman-Bucy filter so that only state information needs to be transmitted during the online execution. We provide guarantees of boundedness of the error covariance for the state estimates under event-triggered communication. Moreover, we show that the centralized optimal solution can be recovered when the event threshold is decreased, which is an improvement with respect to existing event-triggered estimators in the stochastic context. Finally, we show via simulation that the proposal effectively reduces communication without sacrificing the quality of the estimates, and it improves performance with respect to previous approaches.

Keywords: Distributed estimation over sensor nets
Abstract: This paper studies the joint localization and clock synchronization problem in a time-of-arrival-based sensor network employing joint rigidity theory. First, we study the topological conditions and prior information requirements for uniquely determining the network position and clock parameters. Second, we propose a distributed algorithm for the joint localization and clock synchronization problem and discuss the approaches for algorithm initialization. Finally, we analyze the poor conditioning of the problem and propose two possible methods that result in better conditioning as well as maintaining the distributed manner.

Keywords: Concensus control and estimation, Distributed control, Cooperative control
Abstract: In this paper, we address the average consensus problem of multi-agent systems for possibly unbalanced and delay-prone networks with directional information flow. We propose a linear distributed algorithm (referred to as RPPAC) that handles asynchronous updates and time-varying heterogeneous information delays. Our proposed distributed algorithm utilizes a surplus-consensus mechanism and information regarding the number of incoming and outgoing links to guarantee state averaging, despite the imbalanced and delayed information flow in directional networks. The convergence of the RPPAC algorithm is examined using key properties of the backward product of time-varying matrices that correspond to different snapshots of the directional augmented network.

Keywords: Distributed control, Optimal control, Distributed cooperative control over networks
Abstract: Cost optimal control in water distribution networks is considered. Focus is on distribution networks with tree like structures, that can be separated into sections composed of an elevated reservoirs (water towers), a pumping stations, and a consumer district. Given the structure of the network, a centralized and a distributed economic model predictive controller is developed. The later is expected to improve scalability and easy commissioning of the controller setup. Both control methods are based on first principle models. The controllers performance and effectiveness are tested on a scaled water distribution network and their performances are compared. The tests shows similar performance of the two proposed controllers, and both controllers effectively reduce the operational cost when compared to controllers that disregard energy prices.

Keywords: Distributed control, Optimal control, Predictive control for linear systems
Abstract: In this work, we focus on the design of optimal controllers that must comply with an information structure. State-of-the-art approaches do so based on the H2 or Hinfty norm to minimize the expected or worst-case cost in the presence of stochastic or adversarial disturbances. Large-scale systems often experience a combination of stochastic and deterministic disruptions (e.g., sensor failures, environmental fluctuations) that spread across the system and are difficult to model precisely, leading to sub-optimal closed-loop behaviors. Hence, we propose improving performance for these scenarios by minimizing the regret with respect to an ideal policy that complies with less stringent sensor-information constraints. This endows our controller with the ability to approach the improved behavior of a more informed policy, which would detect and counteract heterogeneous and localized disturbances more promptly. Specifically, we derive convex relaxations of the resulting regret minimization problem that are compatible with any desired controller sparsity, while we reveal a renewed role of the Quadratic Invariance (QI) condition in designing informative benchmarks to measure regret. Last, we validate our proposed method through numerical simulations on controlling a multi-agent distributed system, comparing its performance with traditional H2 and Hinfty policies.

Keywords: Distributed estimation over sensor nets, Observers for nonlinear systems, Delay systems
Abstract: Distributed observer design is critical for large- scale systems to collectively estimate the system state via networked sensors. In this paper, we propose a novel distributed observer scheme for estimating the states of a class of nonlinear systems. Unknown and time-varying communication delays are considered due to the ubiquitous network latency when information is exchanged among observer nodes. Based on the Lyapunov stability criterion, a set of linear matrix inequalities (LMIs) are derived for the design of observer gains, which ensure asymptotic convergence of the state estimates to the true state trajectories in the presence of communication delays. Simulation results are given to validate the effectiveness of the proposed method and its advantage over a recent approach without considering communication delays.

Keywords: Game theoretical methods, Optimization algorithms, Stochastic systems
Abstract: We derive the rate of convergence to the strongly variationally stable Nash equilibrium in a convex game, for a zeroth-order learning algorithm. Though we do not assume strong monotonicity of the game, our rates for the one-point feedback, O(Nd/t^{1/2}), and for the two-point feedback, O({N^2d^2}/t), match the best known rates for strongly monotone games under zeroth-order information.

Keywords: Game theoretical methods, Agents networks, Markov processes
Abstract: We introduce a dynamic model for network participation and resource-sharing problems based on the principles of non-cooperative game theory. Within a social network, individuals must decide whether to join cooperative activities or share resources based on anticipated benefits versus incurred costs. We cast these problems as non-cooperative games and comprehensively characterize the Nash equilibria in these settings. Furthermore, we introduce Log-Linear Learning (LLL) as a potential decision strategy for the participants and analyze the long-term dynamics of this approach within the framework. We perform extensive simulations on random networks to validate our research findings empirically. These simulations provide compelling evidence that user engagement in network participation and sharing dilemmas within our proposed framework aligns with the well-established concepts of k-core and (r, s)-core within network structures.

Keywords: Game theoretical methods, Agents networks, Statistical learning
Abstract: We investigate the propagation of stubborn behavior in a network coordination game where players update their strategies using log-linear learning dynamics. A network is considered robust if the stubborn players cannot impact the stable behavior of the other players. We present a graph-theoretic framework for analyzing the robustness of various networks, establishing conditions wherein all network nodes switch to a stubborn behavior. Our framework leverages the notion of graph closed-knittedness, which measures the strength of external influence on a set of nodes. Using closed-knittedness and a closely related notion of graph plumpness, we derive necessary and sufficient conditions for network robustness to stubborn behavior. We validate our analytical results and bound through extensive Monte-Carlo simulations.

Keywords: Game theoretical methods, Distributed control, Optimization
Abstract: Under the umbrella of non-cooperative game theory, we formulate a transactive energy framework to model and control energy communities comprised of heterogeneous agents including (yet not limited to) prosumers, energy storage systems, and energy retailers. The underlying control task is defined as a generalized Nash equilibrium problem (GNEP), which must be solved in a distributed fashion. To solve the GNEP, we formulate a Gauss-Seidel-type alternating direction method of multipliers algorithm, which is guaranteed to converge under strongly monotone pseudo-gradient mappings. As such, we provide sufficient conditions on the private cost and energy pricing functions of the community members, so that the strong monotonicity of the overall pseudo-gradient is ensured. Finally, the proposed framework and the effectiveness of the solution method are illustrated through a numerical simulation.

Keywords: Optimization algorithms, Game theoretical methods, Delay systems
Abstract: In this paper, we propose a distributed algorithm to seek the Nash equilibrium of the non-cooperative game over time-varying networks with time-varying delays. The slack block parameters, updated adaptively at each iteration, are designed to predict the players' future actions based on their current and past information. Each player then selects the appropriate slack block parameter according to the delay information and updates its local parameters accordingly. This method guarantees the convergence of the proposed algorithm, which can be proved via a non-cooperative game over a generalized delay-free network.

Keywords: Game theoretical methods, Optimization algorithms, Energy systems
Abstract: Over the past decade, the continuous surge in cloud computing demand has intensified data center workloads, leading to significant carbon emissions and driving the need for improving their efficiency and sustainability. This paper focuses on the optimal allocation problem of batch compute loads with temporal and spatial flexibility across a global network of data centers. We propose a bilevel game-theoretic solution approach that captures the inherent hierarchical relationship between supervisory control objectives, such as carbon reduction and peak shaving, and operational objectives, such as priority-aware scheduling. Numerical simulations with real carbon intensity data demonstrate that the proposed approach successfully reduces carbon emissions while simultaneously ensuring operational reliability and priority-aware scheduling.

Keywords: Distributed parameter systems, Nonlinear system theory, Aerospace
Abstract: Reparable systems are systems that are characterized by their ability to undergo maintenance actions when failures occur. These systems are often described by transport equations, all coupled through an integro-differential equation. In this paper, we address the understudied aspect of the controllability of reparable systems. In particular, we focus on a two-state reparable system and our goal is to design a control strategy that enhances the system availability- the probability of being operational when needed. We establish bilinear controllability, demonstrating that appropriate control actions can manipulate system dynamics to achieve desired availability levels. We provide theoretical foundations and develop control strategies that leverage the bilinear structure of the equations.

Keywords: Distributed parameter systems, Stability of nonlinear systems, Stochastic systems
Abstract: In this paper we consider state-feedback global stabilization of stochastic semilinear 2D parabolic PDEs with nonlinear multiplicative noise, where the nonlinearities satisfy globally Lipschitz condition. We consider the Dirichlet actuation and design the controller with the shape functions in the form of eigenfunctions corresponding to the first comparatively unstable N eigenvalues. We suggest an appropriate change of variables leading to homogeneous boundary conditions and employ N-dimensional dynamic extension with the corresponding proportional-integral controller. By using a direct Lyapunov method and It^{o}'s formula for stochastic ODEs and PDEs, we provide mean-square L^2 exponential stability analysis of the full-order closed-loop system. We provide linear matrix inequality (LMI) conditions for finding N and the controller gain. We prove that the LMIs are always feasible provided the Lipschitz constants are small enough and N is large enough. Numerical examples demonstrate the efficiency of our method and show that the employment of the suggested dynamic extension allows for larger Lipschitz constants than the previously used dynamic extensions.

Keywords: Distributed parameter systems
Abstract: This work extends the theory of backstepping control of (m + n) hyperbolic PIDEs and m ODEs to blocks of isotachic states (i.e. where some states have the same transport speed). This particular yet physical and interesting case has not received much attention beyond a few remarks in the early hyperbolic design, and leads to a block backstepping design. Our motivation is the rapid stabilization of N-layer Timoshenko composite beams with anti-damping and anti-stiffness at the uncontrolled boundaries. The problem of stabilization for a two-layer composite beam has been previously studied by transforming the model into a 1-D hyperbolic PIDE-ODE form and then applying backstepping to this new system. In principle this approach is generalizable to any number of layers. However, when some of the layers have the same physical properties (as e.g. in lamination of repeated layers), the approach leads to isotachic hyperbolic PDEs. We use a Riemann transformation to transform the states of N-layer Timoshenko beams into a 1-D hyperbolic PIDE-ODE system. The block backstepping method is then applied to this model, obtaining closed-loop stability of the origin in the L2 sense. An arbitrarily rapid convergence rate can be obtained by adjusting control parameters. Finally, numerical simulations are presented corroborating the theoretical developments.

Keywords: Distributed parameter systems
Abstract: This work proposes mid-course trajectory reconfigurations for humans escaping hazardous environments in indoor surroundings as part of disaster management in civil infrastructures. Hazardous environments are interpreted as spatial fields such as carbon monoxide concentrations, and have accumulated effects on escaping humans within indoor environments. When the spatial field is known and available to an evacuee then a level-set based guidance can provide an optimal trajectory to an escape exit that corresponds to the smallest accumulated amount of hazardous material in the evacuee's lungs. However, when the spatial field is unknown to an evacuee, an integrated estimation and trajectory planning scheme is warranted. This paper combines the asymptotic embedding approach for state estimation of spatially distributed processes via mobile sensor with a modified level-set trajectory generation scheme. Incorporating realism due to computing and planning time, a trajectory cycle is decomposed into a planning stage in which a mobile agent (human) is immobile and uses the most recent state estimate to generate viable escape trajectories, and the travel stage in which the mobile agent is executing the trajectory computed during the planning stage. As new process state information is updated, the escape trajectories are recalculated thereby leading to continuous escape trajectory reconfiguration. In both stages of a given cycle the mobile agent is continuously estimating the spatially distributed process but only using the most recent snapshot of the spatial process estimate for trajectory recalculation. Extensive numerical studies are included to shed light on the detrimental effects of accumulated amounts of hazardous environments on the escape trajectories of humans during indoor evacuation.

Keywords: Concensus control and estimation, Distributed cooperative control over networks, Distributed parameter systems
Abstract: This paper delves into the distributed estimator-based Dynamic Average Consensus (DAC) control problem within multi-agent systems (MASs) modeled by Partial Differential Equations (PDEs). The objective of DAC is for agents to converge to time-varying average profiles, referred to as dynamic average reference signals. Unlike prior research, this study will use a distributed estimator to recover the average reference signals for agents to track, which converges to the desired average reference in finite time. The reference signal with compact support employed in this study represents a more generalized signal type compared to previous works. Based on the distributed estimator, this research explores the DAC problem within in-domain PDE control. In-domain control is where input control acts within the governing equation. To assess the stability of the closed-loop system, we employ the Lyapunov technique for analysis. Finally, the proposed control designs’ effectiveness in each section of the paper is demonstrated through simulation examples.

Keywords: Observers for nonlinear systems, Sensor and mesh networks, Optimization algorithms
Abstract: The optimal placement and design of sensors is commonly encountered in industrial and applied problems, such as urban planning and the supervision of temperature and pressure in gas networks. In essence, sensors are considered optimally designed when they ensure the highest level of observation for the specific phenomenon in question. Typically, this design process is guided by specific objectives and is subject to constraints commonly defined by an appropriate partial differential equation (PDE), taking into account the underlying physics of the process. In the present work, we focus on two independent study cases: • The optimal shape design of a convex sensor. • The optimal placement of a finite number of sensors inside a given region. Here, we address the problem in a purely geometric setting, without involving a specific PDE model. We consider a simple and natural geometric criterion of performance, based on distance functions. But, as we shall see, tackling it will require to employ geometric analysis methods

Keywords: Biological systems, Identification, Computational methods
Abstract: In previous work, we proposed Auto-Regressive (AR) modelling on population trees for the stochastic transmission of individual-cell kinetic gene expression parameters at cell division. We addressed inference of the AR model parameters from individual gene expression profiles in a growing population, under the assumption of known parental relationships. In this paper, we explore the same inference problem in the case where only the generation that cells belong to is known, while parental relationships are unknown. First assuming that individual-cell parameters are measured directly with known degree of uncertainty, we develop a likelihood-based method that is applicable beyond the specific case of gene expression. Then, for data consisting of gene expression profiles, we extend the method into a pipeline for the identification of the AR model parameters via preliminary reconstruction of individual-cell parameters and their uncertainty. Performance of all methods is demonstrated via simulations inspired from real data.

Keywords: Biological systems, Modeling, Linear time-varying systems
Abstract: We address online estimation of microbial growth dynamics in bioreactors from measurements of a fluorescent reporter protein synthesized along with microbial growth. We consider an extended version of standard growth models that accounts for the dynamics of reporter synthesis. We develop state estimation from sampled, noisy measurements in the cases of known and unknown growth rate functions. Leveraging conservation laws and regularized estimation techniques, we reduce these nonlinear estimation problems to linear time-varying ones, and solve them via Kalman filtering. We establish convergence results in absence of noise and show performance on noisy data in simulation.

Keywords: Biological systems, Modeling, Nonlinear system theory
Abstract: The algal competition between blue-green algae (cyanobacteria) and green algae (chlorophytes) is a topic that is widely treated both from the biological point of view and from that of the ecological modelling. The spread of blue-green algae in fresh waters, in fact, represents a problem of great interest linked to the quality of water and to the production by these organisms of cyanotoxins particularly toxic not only for the flora and fauna but also for humans. Moreover, blue-green algal blooms are associated with high levels of water turbidity and, due to their inedibility, less ecosystem biodiversity. Several models have been proposed to test algal competition and to propose different control strategies on the ecosystem that can lead to the disappearance of blue-green algae with consequent dominance of green algae. However, these models relate to spatially homogeneous lakes. Here, a more realistic algal competition model in a heterogeneous lake is proposed, by introducing a vertical stratification in the water body and assuming that organisms and nutrients can migrate (with mixing rate D) from the lower to the upper compartment and vice-versa. The analysis of the model not only shows that small/medium mixing rate D can promote algal coexistence, but also provides useful suggestions for the control of water quality that differ to a large extent from what has been achieved in the case of lake homogeneity.

Keywords: Biomedical systems, Identification for control
Abstract: For individuals with diabetes, the intraperitoneal drug-delivery route may enable fully automated artificial pancreas technology. For such systems, the model predictive control (MPC) algorithm is favorable. However, MPC requires a reliable predictive model. In this work, we aim to design a trial protocol to collect data for identification of a bi-hormonal intraperitoneal prediction model. We apply model-based design of experiment (MBDoE) to determine the optimal input of meals, subcutaneous insulin injections, and subcutaneous glucagon injections. Based on parameters from two anesthetized pigs, we design experiments to identify parameters in awake animals. Our results demonstrate how MBDoE may be used as a planning tool when designing trial protocols. The approach may hold potential as a support tool for clinicians when personalizing control algorithms for human AP users.

Keywords: Biological systems, Predictive control for nonlinear systems, Stochastic control
Abstract: There has been significant interest in using advanced control strategies for medical treatments in recent years. This study proposes a two-fold approach to enhance drug dosing in cancer treatment. Firstly, a stochastic model predictive control (SMPC) is designed to address the uncertainties inherent in patient responses. Secondly, this SMPC is formulated as a sequential quadratic programming (SQP) MPC to manage the system's non-linearities. Therefore, this study proposes a stochastic SQP-MPC drug delivery framework to enhance patient outcomes and reduce side effects. The effectiveness of the proposed strategy is assessed via simulations and compared with other strategies.

Keywords: Biomedical systems, Autonomous systems, Machine learning
Abstract: Type 1 diabetes is one of the major concerns in current medical studies. Traditional clinical practice involves non-autonomous manual injection of insulin in the blood, while current research in the field of autonomous regulation of blood glucose concentration mostly focuses on model-based control techniques. This paper introduces a novel Reinforcement Learning-based controller for autonomous glycemic regulation in the treatment of type 1 diabetes, building on the Deep Deterministic Policy Gradient algorithm. The proposed control method is validated through in-vitro simulations on the Bergman glucoregulatory model, proving that it successfully preserves healthy values of blood glucose concentration, while overcoming both standard clinical practice and classical model-based control techniques in terms of both control effort and computational efficiency for real-time applications.

Keywords: Linear systems, Sampled data control, Stability of linear systems
Abstract: This paper focuses on designing an event-triggering mechanism aimed at reducing control updates while maintaining the stability of a saturated closed-loop system. It addresses the regional stabilization of linear systems under input saturation conditions from a data-driven perspective. To do so, we propose a systematic method to convert model-driven conditions into data-driven control design Linear Matrix Inequality (LMI) conditions, enabling the co-design of the event-triggering rule and the state feedback gain. The theoretical contribution is then applied to the control of a spacecraft rendezvous problem.

Keywords: Predictive control for linear systems, Computational methods
Abstract: In the field of model predictive control, Data-enabled Predictive Control (DeePC) offers direct predictive control, bypassing traditional modeling. However, challenges emerge with increased computational demand due to recursive data updates. This paper introduces a novel recursive updating algorithm for DeePC. It emphasizes the use of Singular Value Decomposition (SVD) for efficient low-dimensional transformations of DeePC in its general form, as well as a fast SVD update scheme. Importantly, our proposed algorithm is highly flexible due to its reliance on the general form of DeePC, which is demonstrated to encompass various data-driven methods that utilize Pseudoinverse and Hankel matrices. This is exemplified through a comparison to Subspace Predictive Control, where the algorithm achieves asymptotically consistent prediction for stochastic linear time-invariant systems. Our proposed methodologies' efficacy is validated through simulation studies.

Keywords: Predictive control for nonlinear systems, Nonlinear system identification, Computational methods
Abstract: This paper considers the extension of data-enabled predictive control (DeePC) to nonlinear systems via general basis functions. Firstly, we formulate a basis-functions DeePC behavioral predictor and identify necessary and sufficient conditions for equivalence with a corresponding basis-functions multi-step identified predictor. The derived conditions yield a dynamic regularization cost function that enables a well-posed (i.e., consistent with the multi-step identified predictor) basis-functions formulation of nonlinear DeePC. Secondly, we develop two alternative, computationally efficient basis-functions DeePC formulations that use a simpler, sparse regularization cost function and ridge regression, respectively. An insightful relation between Koopman DeePC and basis-functions DeePC is also presented. The effectiveness of the developed basis-functions DeePC formulations is shown on a benchmark nonlinear pendulum state-space model, for both noise-free and noisy data, while using only output measurements.

Keywords: Behavioural systems, Linear systems
Abstract: In this paper, we present a novel approach to combine data-driven non-parametric representations with model-based representations of dynamical systems. Based on a data-driven form of linear fractional transformations, we introduce a data-driven form of generalized plants. This form can be leveraged to accomplish performance characterizations, e.g., in the form of a mixed-sensitivity approach, and LMI-based conditions to verify finite-horizon dissipativity. In particular, we show how finite-horizon l2-gain under weighting filter-based general performance specifications can be verified for implemented controllers on systems for which only input-output data is available. The overall effectiveness of the proposed method is demonstrated by simulation examples.

Keywords: Linear systems, Emerging control applications, Sampled data control
Abstract: Many industrial processes, computing devices, and networks, such as traffic systems and power grids, exhibit complex dynamics. The model capturing the true dynamics may not be readily available for these systems. Due to advanced sensing technologies it is possible to collect large amounts real- time data. Leveraging this data offers an opportunity to design data-driven control without constructing an explicit model for the system. In this paper, we derive data-dependent matrices from an ensemble of input-state trajectories to parameterize the closed-loop Linear Time-Invariant (LTI) system, accounting for exogenous inputs. The stabilizing control law is designed in such a way that it gets updated based on events, where an event refers to the violation of certain performance conditions. The results proposed are implemented on a computing system, in particular demonstrating auto-scaling of web servers hosted on a private cloud.

Keywords: Linear systems, Identification for control, Sampled data control
Abstract: We study the problem of estimating the states of a linear system based on measured data. We investigate the problem in both deterministic and stochastic settings. In the deterministic case, we develop data-driven conditions under which we can reconstruct state trajectories uniquely. Also, we discuss the case in which we have some missing data in the given input/output measurements. In the stochastic case, we develop a Kalman filter-like algorithm to recursively estimate both states and outputs. Finally, we consider a multi-input multi-output system to elucidate the developed results.

Keywords: Control over networks, Safety critical systems, Control education
Abstract: The first block of the tutorial provides an illustrative introduction to the young but emerging field of encrypted control. General concepts of confidential (or privacy-preserving) controller evaluations are discussed, along with an overview of enabling technologies such as HE. Following this general introduction, the established (partial) homomorphic Paillier cryptosystem is presented in more detail and exemplarily applied to realize two simple encrypted controllers: encrypted state feedback and PI control. Each realization is accompanied by code from a custom Matlab toolbox developed by the tutors, which will be utilized throughout the tutorial session.

Keywords: Control over networks, Safety critical systems, Control education
Abstract: The second block of the tutorial moves towards encrypted control using more modern cryptosystems. Such cryptosystems often build on the so-called learning with errors (LWE) problem. We discuss this fundamental problem and show how it can be used to construct powerful HE schemes. We then present the LWE-based GSW scheme in more detail. In contrast to the partially homomorphic Paillier scheme, GSW offers encrypted additions and multiplications, i.e., (leveled) fully HE. At this point, we specify the concept of levels central to many advanced HE cryptosystems. We then apply GSW to our running controller examples and discuss the benefits resulting from the more flexible cryptosystem.

Keywords: Control over networks, Safety critical systems, Control education
Abstract: The third block of the tutorial finally deals with state-of-the-art HE. Related cryptosystems often build on a ring variant of the LWE problem. We introduce the corresponding ring LWE (RLWE) problem and subsequently present the CKKS cryptosystem in more detail. In particular, we show that, due to its special message space, CKKS allows for more efficient encodings, several other optimizations, and more functionalities. To illustrate the extended capabilities of CKKS, we apply it not only to the running examples but also to slightly more challenging controllers. Finally, we highlight some open problems regarding HE in encrypted control.

Keywords: Nonlinear system identification, Delay systems, Safety critical systems
Abstract: Early detection of delay attacks on feedback control systems can be achieved by recursive identification of delay and dynamics. The paper contributes with an analysis of the convergence of a multiple-model based algorithm for joint recursive identification of fractional delay and continuous time nonlinear state space dynamics. It is proved that the true parameter vector is in the set of global convergence points, while reasons are given why a standard local stability analysis fails. A numerical example illustrates these results.

Keywords: Nonlinear system identification, Identification
Abstract: This paper presents a concurrent estimation method for a polynomial kernel-based nonlinear observer canonical model with a generic structure for nonlinear systems.The algorithm alternates the separable least squares parameter estimation algorithm and extended Kalman filtering. It has the characteristic of fast convergence to a local optimum which depends strongly on the initial values of parameters. To improve its convergence to the global optimum, a genetic algorithmbased concurrent estimation (GA-CE) method is proposed to search for the initial parameter set that minimizes the output estimation error. This optimal concurrent estimation method converges to the optimum or near-optimum solution without any trial-and-error steps for initial values. Validations on the Silverbox and Bouc-Wen hysteresis benchmarks show that the GA-CE method is able to find excellent solutions yielding models with superior predictive performance.

Keywords: Nonlinear system identification, Stochastic systems, Statistical learning
Abstract: It is well known that ignoring the presence of stochastic disturbances in the identification of stochastic Wiener models leads to asymptotically biased estimators. On the other hand, optimal statistical identification, via likelihood-based methods, is sensitive to the assumptions on the data distribution and is usually based on relatively complex sequential Monte Carlo algorithms. We develop a simple recursive online estimation algorithm based on an output-error predictor, for the identification of continuous-time stochastic parametric Wiener models through stochastic approximation. The method is applicable to generic model parameterizations and, as demonstrated in the numerical simulation examples, it is robust with respect to the assumptions on the spectrum of the disturbance process.

Keywords: Nonlinear system identification, Optimal control, Neural networks
Abstract: This paper proposes a generalized algorithmic approach for learning linear model representations for nonlinear systems within the Koopman framework. We focus on schemes that rely on learning the nonlinear transformation functions using deep neural networks. Beyond achieving dynamical accuracy, our primary objective is to develop models capable of simulating nonlinear systems across multiple time steps in the linear space. An algorithm that is based on recursive least squares is proposed to address the optimization complexities inherent in learning such models. In addition, we leverage the learned linear representation to design a linear quadratic regulator to control the original nonlinear systems. The effectiveness of the proposed algorithm is demonstrated in two numerical examples.

Keywords: Nonlinear system identification, Statistical learning, Neural networks
Abstract: This paper addresses the monitoring and continual learning of data-based dynamical models. Throughout the lifespan of any process, many changes can occur. In an indirect control design framework, in order to maintain an effective control system, it is crucial to monitor the modelling performance and adapt the existing model to possible system variations while preserving previously acquired information. A comprehensive methodology is hence proposed to detect a system-model mismatch and its cause, and to update the model accordingly. The proposed idea consists in leveraging control charts constructed on operational data to spot an anomaly and to determine its cause (endogenous or exogenous). The procedure then provides an adaptation algorithm based on the type of change detected: if endogenous, the model is "partially" updated by means of a Moving Horizon Estimation (MHE) algorithm, if exogenous, the model is "incrementally" updated by means of a model uncertainty estimation algorithm. The proposed methodology is tested in simulation on a district heating system benchmark, showing promising results from the monitoring and continual learning perspective.

Keywords: Constrained control, Neural networks, Switched systems
Abstract: The paper on hand considers the optimal control of piecewise affine systems subject to polytopic constraints. While this problem can be addressed by receding horizon control, the approach is known to be computationally demanding. This paper considers the approximation of receding horizon control laws by deep artificial feed-forward neural networks. The concept of projecting inadmissible inputs onto regions derived from feasible sets is extended to the considered problem setup in order to achieve deterministic guarantees on feasibility and constraint satisfaction. Two approaches are proposed and illustrated in numerical examples.

Keywords: Constrained control, Output feedback, Stability of nonlinear systems
Abstract: This paper studies the global asymptotic stabilization of passive nonlinear systems with finite, countable control actions. We show that for nonlinear passive systems that are large-time norm observable and admit a finite control input set whose convex hull contains the origin, the origin can be globally asymptotically stabilized and locally exponentially stabilized by means of relaxed control and nearest-action control approaches. In particular, we improve on a recent result of practical stabilization via nearest-action control by utilizing switching controllers that can synthesize extra control actions from an existing control input set. Three switching methodologies are proposed to enlarge the control set and enable global asymptotic and local exponential stabilization. These three methodologies vary in the cardinality of the expanded control set. These methods are validated in numerical simulations where a comparison of the convergence rate is provided.

Keywords: Constrained control, Robotics, Autonomous robots
Abstract: In this article, we propose a body-aware local navigation strategy for asymmetric holonomic robots for collision-free navigation in narrow pathways with sharp turns. In such scenarios, a robot with non-circular or asymmetric footprint that is comparable to the dimension of the pathways collides with walls when tracking Voronoi paths or risk-aware paths. This problem is addressed in this article through a novel multi-control barrier functions (CBF) based control strategy that achieves the objective of safe collision-free maneuvering at sharp turns. The proposed method is significantly computationally light in comparison to approaches based on model predictive control and online occupancy-grid based free-space and collision detection. In the proposed approach, a minimal set of parameters that characterize a sharp turn and the robot footprint are used to define six control barrier functions that define safe and unsafe regions of operation for a robot. A quadratic programming based CBF safety filter is designed that takes a nominal goal-reaching control as input and returns a minimally-deviating output that enforces the control barrier constraints and renders the safe set forward invariant throughout the turning maneuver. The three kinematic control inputs of the holonomic robot are shared in a conflict-free manner among the six control barrier constraints. The proposed local navigation approach was thoroughly validated in multiple scenarios in a simulated environment, where a robot with asymmetric footprint achieves collision-free maneuvering along multiple sharp turns, while respecting the safety and actuation constraints.

Keywords: Constrained control, Stability of nonlinear systems, Control over networks
Abstract: This paper deals with the stability analysis of continuous-time Lur’e systems under dynamic periodic event-triggered saturating control. A looped-functional approach is applied to handle the continuous-time dynamics of the plant and the aperiodic control signal updates. Considering the emulation design, sector-based inequalities, and Lyapunov Theory arguments, sufficient conditions are derived to ensure local asymptotic stability of the origin of the closed-loop system. Then, a convex optimization problem is proposed to synthesize the event generator parameters aiming to reduce the number of events regarding the time-based implementation. The proposed approach is illustrated by a numerical example.

Keywords: Constrained control, Stability of nonlinear systems, Output feedback
Abstract: This paper deals with the design of a dynamic output feedback controller of the same order as the plant. The plant under consideration is a linear system subject to saturating input and exponentially stable in open loop. The design technique takes advantage of a non-quadratic Lyapunov function involving sign-indefinite quadratic forms, which allows exploiting additional degrees of freedom with respect to a classical quadratic form. The design conditions, combining adequate changes of variables and several sector conditions, are stated in the form of linear matrix inequalities ensuring global exponential stability of the closed-loop system, in addition to a guaranteed prescribed local exponential convergence rate, typically selected as faster than the open-loop plant exponential convergence rate.

Keywords: Transportation systems, Traffic control, Emerging control applications
Abstract: Eco-driving emerges as a cost-effective and efficient strategy to mitigate greenhouse gas emissions in urban transportation networks. Acknowledging the persuasive influence of incentives in shaping driver behavior, this paper presents the `eco-planner,' a digital platform devised to promote eco-driving practices in urban transportation. At the outset of their trips, users provide the platform with their trip details and travel time preferences, enabling the eco-planner to formulate personalized eco-driving recommendations and corresponding incentives, while adhering to its budgetary constraints. Upon trip completion, incentives are transferred to users who comply with the recommendations and effectively reduce their emissions. By comparing our proposed incentive mechanism with a baseline scheme that offers uniform incentives to all users, we demonstrate that our approach achieves superior emission reductions and increased user compliance with a smaller budget.

Keywords: Transportation systems, Traffic control, Optimization
Abstract: This paper presents a modeling and optimization framework to study congestion-aware ride-pooling Autonomous Mobility-on-Demand (AMoD) systems, whereby self-driving robotaxis are providing on-demand mobility, and users headed in the same direction share the same vehicle for part of their journey. Specifically, taking a mesoscopic time-invariant perspective and on the assumption of a large number of travel requests, we first cast the joint ride-pooling assignment and routing problem as a quadratic program that does not scale with the number of demands and can be solved with off-the-shelf convex solvers. Second, we compare the proposed approach with a significantly simpler decoupled formulation, whereby only the routing is performed in a congestion-aware fashion, whilst the ride-pooling assignment part is congestion-unaware. A case study of Sioux Falls reveals that such a simplification does not significantly alter the solution and that the decisive factor is indeed the congestion-aware routing. Finally, we solve the latter problem accounting for the presence of user-centered private vehicle users in a case study of Manhattan, NYC, characterizing the performance of the car-network as a function of AMoD penetration rate and percentage of pooled rides within it. Our results show that AMoD can significantly reduce congestion and travel times, but only if at least 40% of the users are willing to be pooled together. Otherwise, for higher AMoD penetration rates and low percentage of pooled rides, the effect of the additional rebalancing empty-vehicle trips can be even more detrimental than the benefits stemming from a centralized routing, worsening congestion and leading to an up to 15% higher average travel time.

Keywords: Game theoretical methods, Transportation systems
Abstract: A social planner who wishes to influence human decision-making must shape incentives to account for non-rational decision-making biases among the population to be influenced. However, the planner does not operate in a vacuum; societies are comprised of many heterogeneous economic actors whose self-interested behavior may hamper the social planner's ability to achieve their goals. For instance, in a transportation network which is subject to road tolls, a 3rd-party economic agent (whom we call the arbitrageur) may launch a service to provide users with information which helps them optimize their toll pricing and congestion experiences. What are the effects of such an arbitrageur on the social planner's incentive design problem? Do there still exist behavior-optimizing incentive schemes in the presence of an arbitrageur? In this work, our contributions are a formal model of this scenario and analytical derivations of game-theoretic equilibria for a simple transportation network.

Keywords: Transportation systems, Game theoretical methods, Optimization
Abstract: In this paper, we present a study of a mobility game with uncertainty in the decision-making of travelers and incorporate prospect theory to model travel behavior. We formulate a mobility game that models how travelers distribute their traffic flows in a transportation network with splittable traffic, utilizing the Bureau of Public Roads function to establish the relationship between traffic flow and travel time cost. Given the inherent non-linearities and complexity introduced by the uncertainties, we propose a smooth approximation function to estimate the prospect-theoretic cost functions. As part of our analysis, we characterize the best-fit parameters and derive an upper bound for the error. We then show the existence of an equilibrium and its its best-possible approximation.

Keywords: Transportation systems
Abstract: With the increasing popularity of ride-hailing services, new modes of transportation are having a significant impact on the overall performance of transportation networks. As a result, there is a need to ensure that both the various transportation alternatives and the spatial network resources are used efficiently. In this work, we analyze a network configuration where part of the urban transportation network is devoted to dedicated bus lanes. Apart from buses, we let pool ride-hailing trips use the dedicated bus lanes which, contingent upon the demand for the remaining modes, may result in faster trips for users opting for the pooling alternative. Under an aggregated modelling framework, we characterize the spatial configuration and the multi-modal demand split for which this strategy achieves a system optimum. For these specific scenarios, we compute the equilibrium when ride-hailing users can choose between solo and pool services, and we provide a pricing scheme for mitigating the gap between total user delays of the system optimum and user equilibrium solutions, when needed.

Keywords: Autonomous systems, Neural networks, Observers for nonlinear systems
Abstract: This study presents a novel approach for estimating lateral velocity, an important parameter for vehicle stability characterization. Aiming to resolve the problems of poor estimation accuracy caused by the insufficient modeling of traditional model-based methods and issues with sampled and delayed measurements, a sampled delay data neural network method for lateral velocity estimation is designed. Our approach incorporates a compensating injector to fill information gaps between samples, an extended compensation dynamic to reduce delays' impact, and a radial basis function neural network to mimic vehicle motions. Continuous weight updates ensure adaptability, and stability is demonstrated using the Lyapunov methodology. Experimental results confirm the effectiveness of our approach, providing promising insights to enhance lateral velocity estimation and improve control and stability in autonomous vehicle systems.

Keywords: Autonomous systems, Aerospace, Intelligent systems
Abstract: Dynamic obstacle avoidance is an essential function for Unmanned Aerial Vehicles (UAVs) to ensure the safe and reliable operations of drones in real-world environments. It allows drones to navigate and react to environmental changes in real time, preventing collisions and maintaining their flight paths. Dynamic obstacle avoidance also improves the success rate of the drone's mission by reducing the need for manual control. In this study, we propose a model predictive control (MPC) concept to generate high-level control commands for drones to avoid dynamic obstacles by integrating Gaussian process regression to forecast the motion of the moving obstacle based on noisy observations. Additionally, we also investigated the applicability of the Kalman filter as an alternative approach in this context. Our tests demonstrate promising results for multi-rotor drones in physics-based simulations.

Keywords: Autonomous systems, Automotive, Transportation systems
Abstract: Data annotation in autonomous vehicles is a critical step in the development of Deep Neural Network (DNN) based models or the performance evaluation of the perception system. This often takes the form of adding 3D bounding boxes on time-sequential and registered series of point-sets captured from active sensors like Light Detection and Ranging (LiDAR) and Radio Detection and Ranging (RADAR). When annotating multiple active sensors, there is a need to motion compensate and translate the points to a consistent coordinate frame and timestamp respectively. However, highly dynamic objects pose a unique challenge, as they can appear at different timestamps in each sensor’s data. Without knowing the speed of the objects, their position appears to be different in different sensor outputs. Thus, even after motion compensation, highly dynamic objects are not matched from multiple sensors in the same frame, and human annotators struggle to add unique bounding boxes that capture all objects. This article focuses on addressing this challenge, primarily within the context of Scania-collected datasets. The proposed solution takes a track of an annotated object as input and uses the Moving Horizon Estimation (MHE) to robustly estimate its speed. The estimated speed profile is utilized to correct the position of the annotated box and add boxes to object clusters missed by the original annotation.

Keywords: Autonomous systems, Automotive, Transportation systems
Abstract: The continuous improvements of sensors' accuracy, communication frameworks, and computing technologies, have pushed the horizon of autonomous driving to a new branch of mobility, which enhances safety, efficiency, and convenience of automotive transportation. A fundamental step to the success of these systems is the design of a robust, safe, and sample-efficient decision-making module. However, real-world applications to semi-structured or even unstructured environments, such as home zones, parking valets, and narrow passages, are very limited. This paper proposes Informed Hybrid A Star (InHAS), a new computationally lightweight path planning algorithm which efficiently provides optimal paths while taking into account vehicle dimensions and satisfying non-holonomic constraints. We validate the effectiveness of the proposed method both in simulation and in real-world application.

Keywords: Delay systems, Linear systems, Lyapunov methods
Abstract: In this paper, we present necessary and sufficient stability tests for two classes of linear delay systems: neutral-type linear time-delay systems and linear time-delay systems with distributed delays.In both cases, they are based on the Lyapunov-Krasovskii functionals with prescribed derivative expressed in terms of the delay Lyapunov matrix. The stability tests amount to verify the semi-positivity of a matrix resulting from the substitution of Legendre polynomial approximation of the functional argument. As illustrated in two examples, the low matrix dimension for establishing sufficiency makes the test numerically efficient.

Keywords: Modeling, Fuzzy systems, Neural networks
Abstract: Knowledge of complex systems is often imprecise and subject to frequent modifications. Consequently, creating a knowledge representation and inference model that can adapt to changes in information is crucial. In this paper, we integrate the functional link into the fuzzy Petri net and propose a generalized model called functional link fuzzy Petri net (FLFPN). This model retains the explanatory ability of fuzzy Petri nets while acquiring the powerful learning ability of functional link neural networks. Finally, since the forming of congestion is highly sensitive to traffic situations in peak hours, we use FLFPN to predict traffic situations of an expressway ramp, which shows a significant improvement in the prediction accuracy, as compared with the traditional FPN model.

Keywords: Computational methods, Randomized algorithms, Optimization algorithms
Abstract: We introduce new techniques to check whether a zonotope is contained in another zonotope. This fundamental problem in control theory has many applications, such as the verification of invariant sets, formal verification of controllers, and fault detection. Our first method uses a search-based vertex enumeration to quickly and efficiently check containment. We also propose two stochastic methods that are able to rapidly disprove or confirm containment with a certain probability. Furthermore, we generalize the first approach to the case where the circumbody is an ellipsotope and generalize the stochastic methods to the case where both the inbody and the circumbody are ellipsotopes. We conclude by comparing the efficiency of our algorithms to currently available ones.

Keywords: Computational methods, Large-scale systems, Nano systems
Abstract: LAMMPS, an acronym for Large-scale Atomic and Molecular Massively Parallel Simulator, is a widely used open-source tool for high-fidelity molecular dynamics (MD) simulations. In this paper, we take the initial steps towards using LAMMPS for synthesis and validation of feedback control in nanoscale manipulation. We begin by introducing the field of MD itself, discussing the specific challenges related to control synthesis, applications of nanoscale manipulation, and the intricacies of high-fidelity MD simulations. Then, we explain the main steps in modeling a molecular system in LAMMPS and provide an illustrative example. In the example, we consider a nanoscale flake of molybdenum disulfide manipulated with the tip of an atomic force microscope over an atomic surface. We designed a simple PID controller to slide the flake with the microscope tip into a desired position. To run LAMMPS simulations with closed-loop control, we utilized the official Python wrapper for LAMMPS, upon which we implemented additional functionalities. We share the code of the simulations freely with the research community through a public repository.

Keywords: Computational methods, Lyapunov methods, Stability of nonlinear systems
Abstract: The RBF method to compute Lyapunov functions for nonlinear systems uses generalized interpolation in Reproducing Kernel Hilbert spaces. We present two different implementation in C++. One that is computationally efficient and one that is memory efficient. The former uses standard functions of a numerical library and the latter directly calls LAPACK routines for packed matrices to perform in-place Cholesky factorization of the interpolation matrix. The memory efficient implementation only needs one-fourth of the memory needed when using a standard numerical library and thus makes it possible to use double the amount of collocation points. Both implementations are easily adapted to different generalized interpolation problems.

Keywords: Feedback linearization, Computational methods, UAV's
Abstract: This paper details development of a dynamic feedback linearization controller tailored to trajectory tracking of hybrid UAVs with a (tailsitter) flying wing topology. First, a differential flatness transform is presented using a simplified aerodynamic model with negligible lateral forces. The proposed controller derives from the flatness transform. For every state we can determine a collection of flat trajectories that correspond with that state. From that collection, we choose a flat trajectory that converges smoothly to the reference flat trajectory and apply the flat inverse dynamics to compute a suitable control input. To remedy the lack of an explicit inverse of the forward dynamics, we propose a dynamic inverse mapping approach which keeps the control algorithm computationally affordable. We evaluate and compare the control architecture with state-of-the-art cascade control in simulation.

Keywords: Network analysis and control, Optimal control, Modeling
Abstract: As future hydrogen networks will be strongly linked to the electricity system via electrolysers and hydrogen power plants, challenges will arise for their operation. A suitable response to phenomena, such as rapidly changing boundary conditions and unbalanced supply and demand, requires the implementation of operational concepts based on transient pipe models. The transient pipe flow can be described by the isothermal Euler equations, which we discretize using an explicit Method Of Characteristics. Based on this, we develop a nonlinear space-time discretized network model that incorporates various other components, including hydrogen storage facilities, active elements such as valves and compressor stations, as well as electrolyzers and fuel cells. This network model serves as the foundation for the development of a tailored economic model predictive control algorithm designed for fast timescales. The algorithm enables controlled pressure changes within specified bounds in response to changes in supply and demand while simultaneously minimizing fast pressure fluctuations in the pipelines. Through a detailed case study, we demonstrate the algorithm's proficiency in addressing these transient operation challenges.

Keywords: Optimal control, Network analysis and control
Abstract: The transformation of fossil fuel-based district heating grids (DHGs) to CO2-neutral DHGs requires the development of novel operating strategies. Model predictive control (MPC) is a promising approach, as knowledge about future heat demand and heat supply can be incorporated into the control, operating constraints can be ensured and the stability of the closed-loop system can be guaranteed. In this paper, we employ MPC for DHGs to control the system mass flows and injected heat flows. Following common practice, we derive terminal ingredients to stabilize given steady state temperatures and storage masses in the DHG. To apply MPC with terminal ingredients, it is crucial that the system under control is stabilizable. By exploiting the particular system structure, we give a sufficient condition for the stabilizability in terms of the grid topology and hence, for the applicability of the MPC scheme to DHGs. Furthermore, we demonstrate the practicability of the application of MPC to an exemplary DHG in a numerical case study.

Keywords: Energy systems, Network analysis and control, Stability of nonlinear systems
Abstract: An increase in local renewable gas production will necessitate gas outflows to higher pressure networks to achieve balancing in the future gas networks. We expand on traditional valve-based pressure regulation by introducing a grid-forming compressor unit. This unit incorporates a centrifugal compressor based on the Greitzer model and enables pressure regulation with bi-directional gas flows. Nonlinear controllers are proposed for the valves of the compressor unit, whereas a proportional-integral controller with damping injection is used for the centrifugal compressor. We provide scalar design inequalities for controller which ensure the controlled compressor unit is equilibrium-independent passive (EIP). By showing that the gas pipelines also EIP, we demonstrate the modular and topology-independent stability of the gas network equilibrium in which so-called prosumers may inject or extract gas at any node in the network. The controller and stability results are verified via simulations.

Keywords: Energy systems, Optimization, Predictive control for linear systems
Abstract: The integration of a large share of renewable generation poses growing challenges to the safe and reliable operation of modern energy systems. Utilizing large scale energy storage is an effective way to accommodate high penetration of renewable energy. Compared with other large-scale energy storage technologies, such as pumped storage hydropower and compressed air energy storage, pumped thermal energy storage (PTES) has significantly higher energy density and less space requirements. Furthermore, PTES is able to integrate multiple energy networks, e.g., electric and heating, which is an additional source of flexibility. In this paper, the use of a PTES to provide simultaneous peak shaving and voltage control in multi-energy systems (MES) is explored and an operation optimization framework is developed. In the proposed framework, the operating characteristics of the PTES are linearized, and the dynamics of the PTES are handled by a model predictive control (MPC) scheme. The overall optimization problem is formulated as a mixed-integer nonlinear problem. The performance of the proposed optimization framework is verified through realistic case studies and the potential benefits of the PTES to support more sustainable grid operation is demonstrated.

Keywords: Energy systems, Predictive control for nonlinear systems, Large-scale systems
Abstract: The inherently nonlinear, large-scale, and time-varying nature of district heating systems pose significant challenges from a control perspective. In this paper, we address these challenges by applying an economic MPC. Economic MPC is a dynamic real-time optimization method, enabling both optimal planning and stability of the closed-loop system. Our strategy constitutes several steps. First, we introduce a discrete-time modular framework for the district heating system, establishing its strict dissipativity with respect to a desired, potentially time-varying, equilibrium. We identify a set of meaningful objective functions for the district heating systems, preserving this property. Second, we show how strict dissipativity implies the turnpike property, which, in turn, guarantees approximate optimality, practical stability, and recursive feasibility for the EMPC closed-loop. Finally, we provide numerical simulations to demonstrate the effectiveness of our work.