Keywords: Biological systems
Abstract: Living cells are complex, stochastic dynamical systems that achieve remarkable robustness and adaptability through feedback. Cybergenetics seeks to bring the principles of control theory into this domain, enabling the real-time regulation of cellular behavior while also revealing new challenges and opportunities for control. In this lecture, I will show how fundamental ideas from feedback control can be translated into genetic controllers that operate reliably within the noisy and nonlinear environment of living cells. I will present a universal internal model principle for biomolecular systems, showing that integral feedback and its generalizations underpin kinetics-independent robust tracking and disturbance rejection at the molecular scale. I will then show how the constraints of biological implementation, such as stochasticity and molecular discreteness, challenge classical assumptions and point toward new directions for control theory. These ideas will be illustrated through experimentally implemented genetic feedback controllers and emerging applications, including engineered cells that autonomously sense and respond to disease. Together, they establish a two-way bridge between control theory and biology, where feedback not only enables new capabilities in living systems but also expands the foundations of control toward a control theory of living systems.

Keywords: Adaptive control, Observers for linear systems, LMI's/BMI's/SOS's
Abstract: In this paper, we address the problem of data-driven stabilization of continuous-time multi-input multi-output (MIMO) linear time-invariant systems using the input–output data collected from an experiment. Building on recent results for data-driven output-feedback control based on non-minimal realizations, we propose an approach that can be applied to a broad class of continuous-time MIMO systems without requiring a uniform observability index. The key idea is to show that Kreisselmeier’s adaptive filter can be interpreted as an observer of a stabilizable non-minimal realization of the plant. Then, by postprocessing the input–output data with such a filter, we derive a linear matrix inequality that yields the feedback gain of a dynamic output-feedback stabilizer.

Keywords: Linear systems, Identification for control, Optimization
Abstract: In recent years, the so-called `direct data-driven control' has been a topic of intense research, and it is expected that it will become prominent in future complex dynamical systems control. Within this framework, regularization not only implicitly enforces system identification, but also plays a crucial role in ensuring reliable closed-loop behavior. To further enhance the performance of data-driven controllers, we propose a new regularization method for direct data-driven LQR control of unknown LTI systems, based on a regularized covariance parameterization. Unlike existing data-driven techniques, the proposed method remains effective in handling ill-conditioned cases, such as when the data matrix has a large condition number. Then, we demonstrate that our method is equivalent to the indirect certainty-equivalence LQR combined with Tikhonov regularization. Furthermore, we extend our method to the design of controllers for unknown nonlinear systems using Koopman linear embedding. Finally, the simulation results validate the effectiveness and advantages of the proposed regularization method.

Keywords: Linear systems, Uncertain systems, LMI's/BMI's/SOS's
Abstract: We present an approach to compute stabilizing controllers for continuous-time linear time-invariant systems directly from an input–output trajectory affected by process and measurement noise. The proposed output-feedback design combines (i) an observer of a non-minimal realization of the plant and (ii) a feedback law obtained from a linear matrix inequality (LMI) that depends solely on the available data. Under a suitable interval excitation condition and knowledge of a noise energy bound, the feasibility of the LMI is shown to be necessary and sufficient for stabilizing all non-minimal realizations consistent with the data. We further provide a condition for the feasibility of the LMI related to the signal-to-noise ratio, guidelines to compute the noise energy bound, and numerical simulations that illustrate the approach.

Keywords: Process control, Optimization, Transportation systems
Abstract: This paper examines how the Proportional- Integral-Derivative (PID) controller is evolving in a context where digitalisation, improved sensing, and data-driven engineering are becoming standard in industry. PID remains the main tool for everyday control, but new methods based on machine learning, reinforcement learning, and artificial intelligence are beginning to influence how tuning, supervision, and adaptation are carried out. The paper reviews why PID is still essential in practice, how data-driven techniques can enhance its tuning and emulation, and what challenges arise when introducing learning-based tools in real industrial environments. Particular attention is given to the balance between potential performance improvements and the practical requirements of industrial control, such as robustness, interpretability, and ease of maintenance. Rather than proposing specific solutions, the work aims to provide a structured overview of current developments and their practical implications, highlighting both opportunities and limitations. It also outlines possible ways to integrate AI with existing control architectures while preserving the robustness, transparency, and simplicity that make PID widely trusted in industrial applications. In this sense, the paper positions PID not as a legacy technology to be replaced, but as a central component that can be progressively extended within more complex, data-enabled control frameworks.

Keywords: Uncertain systems, Robust control, Optimal control
Abstract: This article explores a resilient, data-driven control framework for unknown linear time-invariant (LTI) systems under the threat of poisoned data. Specifically, we consider scenarios in which up to f out of T input-state samples may be manipulated by an adversary aiming to degrade performance or destabilize the system. We begin with a simple example demonstrating that even a single carefully crafted sample can destabilize a closed-loop system, underscoring the vulnerability of direct data-driven control methods and motivating the need for resilience. To address this challenge, we first propose a resilient control design strategy for noise-free settings, based on majority voting across all possible subsets of the input-state dataset with a certain cardinality. We then extend this approach to noisy and disturbed environments, showing that the geometric median of the mentioned subsets provides a resilient solution. Finally, we validate the effectiveness of our framework through numerical simulations.

Keywords: Linear parameter-varying systems, Identification for control, Nonlinear system identification
Abstract: This paper extends the use of Gated Recurrent Units (GRUs) within the Linear Parameter-Varying/Linear Fractional Transformation (LPV–LFT) framework for robust control, previously introduced in earlier work. A novel methodology is proposed for designing controllers for nonlinear plants through a strictly equivalent LPV–LFT formulation of a modified GRU model. In addition, H∞ gain-scheduled LPV controllers are developed to meet performance requirements on a nonlinear benchmark. The approach employs an Unscented Kalman Filter (UKF) for state estimation of the GRUs models, enabling real-time computation of varying parameters. The proposed framework bridges data-driven identification and robust LPV control, enabling systematic synthesis of controllers for nonlinear systems identified via GRUs architectures.

Keywords: Aerospace, Machine learning, Safety critical systems
Abstract: This paper presents a safety-guaranteed, runtime-efficient imitation learning framework for spacecraft close proximity control. We leverage Control Barrier Functions (CBFs) for safety certificates and Control Lyapunov Functions (CLFs) for stability as unified design principles across data generation, training, and deployment. First, a nonlinear Model Predictive Control (NMPC) expert enforces CBF constraints to provide safe reference trajectories. Second, we train a neural policy with a novel CBF-CLF-informed loss and DAGGER-like rollouts with curriculum weighting, promoting data-efficiency and reducing future safety filter interventions. Third, at deployment a lightweight one-step CBF-CLF quadratic program minimally adjusts the learned control input to satisfy hard safety constraints while encouraging stability. We validate the approach for ESA-compliant close proximity operations, including fly-around with a spherical keep-out zone and final approach inside a conical approach corridor, using the Basilisk high-fidelity simulator with nonlinear dynamics and perturbations. Numerical experiments indicate stable convergence to decision points and strict adherence to safety under the filter, with task performance comparable to the NMPC expert while significantly reducing online computation. A runtime analysis demonstrates real-time feasibility on a commercial off-the-shelf processor, supporting onboard deployment for safety-critical on-orbit servicing.

Keywords: Process control, Predictive control for nonlinear systems, Machine learning
Abstract: Learning-based surrogate models offer a powerful alternative to analytical models for model-based control of nonlinear dynamical systems with uncertain and context- dependent dynamics. A receding-horizon control framework is developed that exploits flow-matching models to generate state trajectories conditioned on the system’s initial state, uncertain parameters, and candidate input sequences. These models provide expressive, data-driven surrogate dynamics without requiring explicit analytical representations. To compute control inputs, Bayesian optimization minimizes a cost function evaluated on surrogate-generated trajectories, enabling efficient optimization despite the non-differentiable and computationally expensive nature of the generative model. The resulting inputs are applied in a receding-horizon fashion and re-optimized using updated state and parameter information, yielding an adaptive, learning-based control strategy. The effectiveness of the approach is demonstrated on a bioreactor system.

Keywords: Neural networks, Predictive control for nonlinear systems, Adaptive systems
Abstract: Neural Processes (NPs) can model distributions over functions, like Gaussian Processes, but with the great flexibility and evaluation speed of neural networks. They can be trained purely on data and without prior knowledge. Yet, their architecture allows for a certain degree of interpretability and modular use. We use an NP to model the dynamics in a nonlinear Model Predictive Control (MPC) scheme. After meta-training the NP on data from various system instances, deployment to an unseen system is highly efficient in terms of data and computation. To demonstrate the effectiveness of our NP-MPC scheme for real-time adaptive control, we deploy it to a physical two-axis pendulum experiment requiring short sampling times.

Keywords: Optimal control, Machine learning, Modeling
Abstract: Scaled model experiments are commonly used in various engineering fields to reduce experimentation costs and overcome constraints associated with full-scale systems. The relevance of such experiments relies on dimensional analysis and the principle of dynamic similarity. However, transferring controllers to full-scale systems often requires additional tuning. In this paper, we propose a method to enable a direct controller transfer using dimensionless model predictive control, tuned automatically for closed-loop performance. With this reformulation, the closed-loop behavior of an optimized controller transfers directly to a new, dynamically similar system. Additionally, the dimensionless formulation allows for the use of data from systems of different scales during parameter optimization. We demonstrate the method on a cartpole swing-up and a car racing problem, applying either reinforcement learning or Bayesian optimization for tuning the controller parameters. Software used to obtain the results in this paper is publicly available at url{https://github.com/josipkh/dimensionless-learning-mpc}

Keywords: Predictive control for nonlinear systems, Neural networks, Complex systems
Abstract: This paper presents a model predictive control (MPC) framework leveraging an ensemble of data-based models to optimally control complex systems under multiple operating conditions. A novel combination rule for ensemble models is proposed, based on the statistical Mahalanobis distance, enabling the ensemble weights to suitably vary across the prediction window based on the system input. In addition, a novel state observer for ensemble models is developed using moving horizon estimation (MHE). The effectiveness of the proposed methodology is demonstrated on a benchmark energy system operating under multiple conditions.

Keywords: Optimization algorithms, Predictive control for nonlinear systems, Machine learning
Abstract: We propose a computationally efficient rollout-then-optimize method to improve a learned control policy at deployment time. A learned policy provides a nominal trajectory, which is refined online by a single Newton step implemented via a Riccati recursion within a model predictive control (MPC) scheme. This refinement combines model knowledge with the learned policy at minimal additional computational cost. We establish bounds on the approximation error of the learned policy relative to the MPC policy and show that one Newton step reduces the suboptimality of the learned rollout quadratically in the policy approximation error. The proposed controller is validated in simulation on a constrained trajectory-tracking task for a quadcopter with nonlinear dynamics. Results highlight that the Newton step significantly improves the learned policy, achieving performance close to a fully converged MPC solution while requiring roughly half of the computational time. The code is available at https://github.com/aghezz1/rl-riccati.

Keywords: Distributed control, Network analysis and control, Servo control
Abstract: We propose a distributed two-degrees-of-freedom (2DOF) architecture for driving autonomous, possibly heterogeneous, agents to agreement. The scheme mirrors classical servo structures, separating local feedback from network filtering. This separation enables independent network-filter design for prescribed noise attenuation and allows controller heterogeneity to reject local disturbances, including disturbances exciting unstable agreement poles -- which is known to be impossible via standard diffusive couplings. The potential of the framework is illustrated via two numerical examples.

Keywords: Agents and autonomous systems, Algebraic/geometric methods
Abstract: This paper investigates the synthesis of controllers for displacement-based formation control in the presence of bounded disturbances, specifically focusing on uncertainties originating from measurement noise. While the literature frequently addresses such problems using stochastic frameworks, this work proposes a deterministic methodology grounded in set-theoretic concepts. By leveraging the principles of set invariance, we adapt the theory of ultimate boundedness to the specific dynamics of displacement-based formations. This approach provides a rigorous method for analyzing the system's behavior under persistent disturbances. Furthermore, this set-theoretic framework allows for the optimized selection of the proposed control law parameters to guarantee pre-specified performance bounds. The efficacy of the synthesized controller is demonstrated in the challenging application of maintaining tight formations in a multi-obstacles environment.

Keywords: Distributed control, Optimization, Automotive
Abstract: Connected and automated vehicles (CAVs) can improve efficiency and safety at interaction zones such as intersections, yet coordinating mixed traffic with both CAVs and human-driven vehicles (HVs) in real time remains challenging. The computational burden of centralised optimisation and the nonconvex constraints arising from multiple conflict points limit the scalability of existing approaches. Distributed Model Predictive Control (DMPC) methods ease this burden but struggle to accommodate HV behaviour and maintain feasible coordination. This paper proposes a distributed coordination framework based on the Alternating Direction Method of Multipliers (ADMM). The problem is formulated as an estimated time-of-arrival (ETA) consensus optimisation, where each CAV solves a local MPC problem while a lightweight coordinator enforces ETA consistency across conflict points. HVs are incorporated through reservation constraints, enabling safe and efficient mixed-traffic interactions. Experiments on simulated intersections show that residuals decrease substantially within a few iterations, and traffic throughput improves by roughly 25% compared with a baseline DMPC method. The computational analysis further demonstrates real-time feasibility, achieving update times below 0.1 s with 15 vehicles. These results suggest that the proposed framework provides a scalable and practical solution for realtime coordination in mixed traffic.

Keywords: Distributed control, UAV's, Lyapunov methods
Abstract: This paper addresses the challenging problem of designing distributed rigidity-based formation control strategies for aerial robots with a spherical event camera as the primary on-board sensor. In fact, since accurate distance measurements are seldom available in aerial robotics, bearing-based control algorithms have recently emerged as a valid alternative, offering increased robustness. However, prior works have mainly focused on simplified robot dynamics (single integrators), thus limiting their practical applicability. In this study, distributed formation control laws are developed for double-integrator robots and quadrotors with full nonlinear dynamics, and their stability properties are formally established via Lyapunov arguments. Simulation experiments demonstrate successful convergence under different initial conditions and formation shapes, and robust performance in the presence of sensor noise.

Keywords: Distributed control, Agents and autonomous systems, UAV's
Abstract: This work presents a distributed control strategy for three-dimensional rigid formations in multi-agent systems (MAS) based on the incorporation of Lagrange multipliers into the control law. The approach ensures convergence towards predefined formations on differentiable surfaces, maintaining the scalability of the system. Two different implementations are analyzed, and the results show that, under mild assumptions, the proposed controllers are locally exponentially stable. Experimental validation is carried out in a virtual environment with drones as agents. The results show a significant reduction in overshoot and control signal magnitude compared to existing approaches, while maintaining comparable settling times and consistent behavior in configurations of up to twenty agents.

Keywords: Distributed control, Cooperative autonomous systems
Abstract: This paper studies distributed safe control for multi-agent systems subject to a coupling control barrier function (CBF) constraint. A key difficulty arises when an agent loses local control authority over the active coupling constraint, in which case local regularity may fail and distributed multiplier coordination may become ill-conditioned. We propose a continuous-time distributed architecture that separates feasibility from optimality. A zero-sum residual-budget decomposition preserves the global CBF inequality by construction, while relevance-weighted budget and multiplier coordination reduce the influence of agents with vanishing local CBF sensitivity. Under local solvability of the distributed subproblems, the resulting control input satisfies the global CBF inequality for all time. In the static case, equilibria are consistent with the KKT conditions of the centralized CBF-QP. For slowly time-varying data, a singular-perturbation analysis suggests the tracking-error scaling under time-scale separation. Numerical results show close agreement with the centralized reference and reduced oscillation relative to an unweighted baseline.

Keywords: Optimization, Optimization algorithms, Optimal control
Abstract: Proximal methods such as the Alternating Direction Method of Multipliers (ADMM) are effective at solving constrained quadratic programs (QPs). To tackle infeasible QPs, slack variables are often introduced to ensure feasibility, which changes the structure of the problem, increases its size, and slows down numerical resolution. In this letter, we propose a simple ADMM scheme to tackle QPs with slack variables without increasing the size of the original problem. The only modification is a slightly different projection in the z-update, while the rest of the algorithm remains standard. We prove that the method is equivalent to applying ADMM to the QP with additional slack variables, even though slack variables are not added. Numerical experiments show speedups of the approach.

Keywords: Optimization, Optimal control
Abstract: We study a continuous-time dynamical system which arises as the limit of a broad class of nonlinearly preconditioned gradient methods. Under mild assumptions, we establish existence of global solutions and derive Lyapunov-based convergence guarantees. For convex costs, we prove a sublinear decay in a geometry induced by some reference function, and under a generalized gradient-dominance condition we obtain exponential convergence. We further uncover a duality connection with mirror descent, and use it to establish that the flow of interest solves an infinite-horizon optimal-control problem of which the value function is the Bregman divergence generated by the cost. These results clarify the structure and optimization behavior of nonlinearly preconditioned gradient flows and connect them to known continuous-time models in non-Euclidean optimization.

Keywords: Optimization, Optimization algorithms, Optimal control
Abstract: We propose an acceleration scheme for first-order methods (FOMs) for convex quadratic programs (QPs) that is analogous to Anderson acceleration and the Generalized Minimal Residual algorithm for linear systems. We motivate our proposed method from the observation that FOMs applied to QPs typically consist of piecewise-affine operators. We describe our Krylov subspace acceleration scheme, contrasting it with existing Anderson acceleration schemes and showing that it largely avoids the latter's well-known ill-conditioning issues in regions of slow convergence. We demonstrate the performance of our scheme relative to Anderson acceleration using standard collections of problems from model predictive control and statistical learning applications. We show that our method is faster than Anderson acceleration across the board in terms of iteration count, and in many cases in computation time, particularly for optimal control and for problems solved to high accuracy.

Keywords: Optimization algorithms, Network analysis and control, LMI's/BMI's/SOS's
Abstract: We develop a system-theoretic framework for the structured analysis of distributed optimization algorithms with decomposable cost functions. We model such algorithms as a network of interacting dynamical systems and derive tests for convergence based on incremental dissipativity and contraction theory. This approach yields a step-by-step analysis pipeline suitable for any network structure, with conditions expressed as linear matrix inequalities. In addition, a numerical comparison with traditional analysis methods is presented, in the context of distributed gradient descent.

Keywords: Stochastic systems, Optimization, Optimal control
Abstract: We consider the problem of worst-case Conditional Value at Risk (CVaR) for quadratic loss functions with ambiguity sets determined by the first and the second moments of random vectors. Despite its popularity in recent times, analytical solutions to the worst-case CVaR problem are known only in a very restricted subclass of the problem, making it difficult to solve problems involving worst-case CVaR beyond its computation. In this article, we extend the subclass of worst-case CVaR problems that admits analytical solutions. For a quadratic loss with a definite quadratic term, we reduce the problem to the maximization of a quadratic polynomial over the unit ball with respect to Euclidean norm. Using the same reduction, we derive analytical solutions when the quadratic term is determined by a scaling of the identity matrix. We also derive analytical solutions when the quadratic loss function does not have a linear term.

Keywords: Linear systems, Optimization algorithms, Observers for linear systems
Abstract: Feedback-based optimization (FBO) provides a simple control framework for regulating a stable dynamical system to the solution of a constrained optimization problem in the presence of exogenous disturbances, and does so without full knowledge of the plant dynamics. However, closed-loop stability requires the controller to operate on a sufficiently slower timescale than the plant, significantly constraining achievable closed-loop performance. Motivated by this trade-off, we propose an estimator-based modification of FBO which leverages dynamic plant model information to eliminate the time-scale separation requirement of traditional FBO. Under this design, the convergence rate of the closed-loop system is limited only by the dominant eigenvalue of the open-loop system. We extend the approach to the case of design based on only an approximate plant model when the original system is singularly perturbed. The results are illustrated via an application to fast power system frequency control using inverter-based resources.

Keywords: Robotics, Switched systems, Nonlinear system theory
Abstract: Robotic-assisted surgery has rapidly evolved, with existing control approaches often combined in hybrid systems governed by finite-state machines to manage transitions between modes. However, many current frameworks lack formal guarantees of stability/passivity for the overall controlled system, posing potential safety risks in human–robot interaction on both the surgeon and patient sides. This work presents a hybrid control framework that enables seamless alternation between autonomous, disconnection, and teleoperation modes while ensuring overall system stability/passivity. The proposed architecture integrates continuous robot dynamics with discrete mode transitions through principled switching policies and stability analysis, offering a safe and reliable solution for complex surgical tasks. Human-in-the-loop simulations demonstrate the framework’s capability to achieve both responsive teleoperation and accurate autonomous execution.

Keywords: Robotics, Safety critical systems, Optimization
Abstract: Recently, manipulators have started to transition towards operating in shared spaces with humans. However, it remains challenging to ensure both safety and high task performance in these shared spaces. To address this, we developed TRIK (Trajectory Refinement with iKinQP), which leverages an Inverse Kinematics Quadratic Programming (iKinQP) controller to refine task-oriented controls to be safe. TRIK’s intervention to ensure system safety is based on our novel ISO 15066 velocity and buffer scaling approach combined with an extension of iKinQP. Our extended iKinQP uses control barrier functions to enforce safety and refine based on continuous observations of a human. We experimentally validate, in simulation, TRIK’s ability to enforce safety while maintaining high performance with a realistic simulation study where a robot and human perform separate tasks in a shared space.

Keywords: Safety critical systems, Emerging control applications, Robotics
Abstract: Flexible payloads such as beams, cables, and deformable tools introduce complex dynamics that challenge safe manipulation with conventional robotic controllers. Excessive deformation can lead to instability, structural damage, or task failure. This paper presents a control barrier function (CBF)–based safety filter for robotic systems handling flexible payloads. The method guarantees constraint satisfaction on deformation limits while maintaining compatibility with existing nominal controllers. A high-order CBF formulation is derived for a general control-affine model of flexible robotic systems, and a quadratic program (QP) is employed to enforce real-time safety constraints. The approach is implemented on a planar flexible beam model, where deformation is limited through angle-based geometric constraints. Experiments using MuJoCo as the payload simulation environment demonstrate that the proposed CBF-based safety filter effectively maintains deformation within predefined limits, outperforming unconstrained control in terms of safety and stability without introducing significant control effort.

Keywords: Uncertain systems, Robotics
Abstract: This paper introduces the notion of stochastic simulation-gap function, which formally quantifies the gap between an approximate mathematical model and a high-fidelity stochastic simulator. Since controllers designed for the mathematical model may fail in practice due to unmodeled gaps, the stochastic simulation-gap function enables the simulator to be interpreted as the nominal model with bounded state- and input-dependent disturbances. We propose a data-driven approach and establish a formal guarantee on the quantification of this gap. Leveraging the stochastic simulation-gap function, we design a controller for the mathematical model that ensures the desired specification is satisfied in the high-fidelity simulator with high confidence, taking a step toward bridging the sim-to-real gap. We demonstrate the effectiveness of the proposed method using a TurtleBot model and a pendulum system in stochastic simulators.

Keywords: Robotics
Abstract: In this paper, a novel control scheme is proposed interconnecting a Dynamic Movement Primitives (DMP) system with a modified low impedance controlled robot to achieve compliance to unexpected contact events and high tracking accuracy under the presence of model and task uncertain- ties. The proposed control scheme is formulated in the joint space and theoretically shown to accurately track the desired trajectory under the presence of uncertainties. Simulations of 2-DOF manipulator model carrying an unknown load under the proposed scheme demonstrate its high tracking accuracy as compared to the low impedance controlled robot while exhibiting highly compliant reactions to external contact forces.

Keywords: Robotics, Optimal control, Predictive control for nonlinear systems
Abstract: Whole-Body Collision Avoidance (WBCA) is a fundamental requirement for articulated robotic systems operating in unstructured environments, particularly for redundant manipulators where both self-collision and environmental collision avoidance must be achieved efficiently. Existing methodologies often treat body-level collision avoidance and trajectory generation as separate problems, which limits their effectiveness. This paper introduces a unified framework for whole-body collision avoidance in articulated robots by integrating the Velocity Obstacles paradigm with Nonlinear Model Predictive Control and the Signed Minimum Distance metric, which quantifies the distance of a given velocity vector from the nearest VO boundary. The proposed framework has been preliminarily validated in a simulation of Franka Emika Panda robotic arm operating in an environment with moving spherical obstacles. Results demonstrate that the approach enables the robot to reach target positions while maintaining whole-body safety and respecting dynamic constraints, highlighting the potential of VO-based predictive control for real-time whole-body collision avoidance in redundant manipulators.

Keywords: Lyapunov methods, Stability of nonlinear systems, Computer aided control design
Abstract: We present verifiable conditions for synthesizing a single smooth Lyapunov function that certifies both asymptotic stability and safety under bounded controls. These sufficient conditions ensure the strict compatibility of a control barrier function (CBF) and a control Lyapunov function (CLF) on the exact safe set certified by the barrier. An explicit smooth control Lyapunov-barrier function (CLBF) is then constructed via a patching formula that is provably correct by design. Two examples illustrate the computational procedure, showing that the proposed approach is less conservative than sum-of-squares (SOS)-based compatible CBF-CLF designs.

Keywords: Lyapunov methods, Computational methods, Stability of nonlinear systems
Abstract: A Lyapunov function can be used to find the domain of attraction of an equilibrium through its sublevel sets. In this paper, we propose a new method to compute a Lyapunov function by approximating the solution of Zubov's equation using meshless collocation. The advantage of approximating the solution of Zubov's equation compared to other equations characterising a Lyapunov function, is that for this particular Lyapunov function the domain of attraction is given by its sublevel set of level one, and thus we can obtain better approximations of the domain of attraction.

Keywords: Lyapunov methods, Computational methods, Nonlinear system theory
Abstract: Continuous piecewise affine (CPA) Lyapunov function synthesis is one method to perform Lyapunov stability analysis for nonlinear systems. This method first generates a mesh over the region of interest in the system's state space and then solves a linear program (LP), which enforces constraints on each vertex of the mesh, to synthesize a Lyapunov function. Finer meshes broaden the class of Lyapunov function candidates, but CPA function synthesis is more computationally expensive for finer meshes -- particularly so in higher dimensional systems. This paper explores methods to mesh the region of interest more efficiently so that a Lyapunov function can be synthesized using less computational effort. Three methods are explored -- adaptive meshing, meshing using knowledge of the system model, and a combination of the two. Numerical examples for two and three dimensional nonlinear dynamical systems are used to compare the efficacy of the three methods.

Keywords: Robust control, Optimal control, Lyapunov methods
Abstract: Solving optimal control problems to determine a stabilizing controller involves a significant computational effort. Time-varying optimal control provides a remedy by designing a tracking system, given as an ordinary differential equation, to track the solution of the optimal control problem. To improve the applicability of the method, measurement errors are considered in this paper and it is described how these errors influence a control Lyapunov function-based decay condition. As a result of these investigations, input-affine constraints that meet the standard formulation and that describe the set of admissible controls are obtained. The paper also derives a requirement on the necessary measurement accuracy as well as a triggering condition for taking a new measurement. The main theorem combines these results into a robustly stabilizing control algorithm, meaning that all closed-loop trajectories starting in a vicinity around the true state converge to zero. Additionally, the tracking system ensures that the optimal control is tracked at the end of each sampling period. The effectiveness of this approach is demonstrated using a train acceleration model and the well-known predator-prey model.

Keywords: Lyapunov methods, Computational methods, Stability of nonlinear systems
Abstract: Verifying stability and safety guarantees for nonlinear systems has received considerable attention in recent years. This property serves as a fundamental building block for specifying more complex system behaviors and control objectives. However, estimating the domain of attraction under safety constraints and constructing a Lyapunov–barrier function remain challenging tasks for nonlinear systems. To address this problem, we propose a Zubov PDE formulation with a Dirichlet boundary condition for autonomous nonlinear systems and show that a physics-informed neural network (PINN) solution, once formally verified, can serve as a Lyapunov–barrier function that jointly certifies stability and safety. This approach extends existing converse Lyapunov–barrier theorems by introducing a PDE-based framework with boundary conditions defined on the safe set, yielding a near-optimal certified underapproximation of the true safe domain of attraction.

Keywords: Lyapunov methods, Constrained control, Optimization algorithms
Abstract: This paper presents an algorithm to optimize a given stabilizing controller by using control Lyapunov functions. The approach consists of two steps. Using linear programming, it first determines a Lyapunov function for the closed-loop system that is obtained by using the initial controller. Second, it determines a controller that is optimal w.r.t.~a given objective function, which tightens the input constraints and penalizes large variations of the control. A case study considers a predator-prey model with inputs and shows how the algorithm improves a given controller in the described manner.

Keywords: Markov processes, Manufacturing processes, Statistical learning
Abstract: Several industrial cyclic processes may comprise multiple non-identical process steps and possibly with non-standard interdependencies. Standard Hidden Markov Models (and variations) may be limited in addressing such finite-step cyclic processes. This paper advances a two-dimensional Generalized Input–Output Hidden Markov Model (IOHMM) framework for monitoring finite-step cyclic processes previously introduced by the author. Unlike classical HMMs or Markov jump linear systems that rely on a single latent chain, the model employs a two-dimensional Markov structure across step and cycle iterations, enabling a stochastic representation of mode evolution in cyclic processes. The present work focuses on practical implementation aspects, including an Expectation-Maximization (EM) based learning scheme with stochastic perturbations, as well as context-conditioned transition and emission models that incorporate continuous control variables (as often the case in industrial processes). The approach is demonstrated on a two-step electrostatic particle-transfer process, and a benchmark analysis is presented relatively to standard machine-learning models.

Keywords: Sensor and signal fusion, Machine learning, Model validation
Abstract: This paper validates a simple multi-fidelity (MF) modeling framework for industrial process monitoring. Our approach combines high-frequency, low-fidelity (LF) sensor data with less frequent, high-fidelity (HF) laboratory analyses to improve prediction accuracy of the monitored output y while maintaining interpretability. Existing MF methods often demonstrate high performance but remain complex and difficult to implement on industrial hardware. We propose a transparent structure that can be generalized across industrial systems with similar data workflow. We verify the methodology on the Tennessee Eastman Process (TEP) benchmark.

Keywords: Safety critical systems, Uncertain systems, Statistical learning
Abstract: Designing effective score functions in Conformal Prediction (CP) for time-series data remains challenging due to conservativeness and/or computational inefficiency. We propose Optimal Selection Conformal Prediction (OSCP), which parameterizes the score function via offset terms. To determine these parameters, we formulate a mixed-integer linear program (MILP) that minimizes an empirical proxy of the region size. We further reformulate this optimization problem into a smaller form (fewer constraints) to improve computational efficiency. We provide theoretical guarantees on both validity and CP-efficiency of OSCP. Numerical experiments demonstrate that OSCP reduces uncertainty-set size and has much lower computational requirements compared to the state-of-the-art method.

Keywords: Distributed parameter systems, Identification for control, Model validation
Abstract: A key step in control system design is testing closed-loop stability in a high-fidelity simulator, but this is only reliable if the simulator’s frequency response function (FRF) matches that of the real plant. This paper presents a frequency-domain validation of a lumped-parameter hydraulic transmission-line model of the pressure dynamics relevant for Managed Pressure Drilling (MPD) of deepwater wells. On a real ultra-deepwater drillship, we apply phase-optimized multisines in open- and closed-loop settings, obtaining high-resolution FRF estimates of both plant and loop gain within minutes. The results show dynamics and stability margins that closely align with simulations for wells reaching 22000 ft (6705 m) deep.

Keywords: Stochastic systems, Algebraic/geometric methods
Abstract: This paper presents a global, coordinate-free formulation of the Fokker–Planck equation on Riemannian manifolds. In the Stratonovich formulation, the infinitesimal generator is expressed intrinsically through Lie derivatives, and its adjoint is derived via the divergence theorem, yielding a concise geometric form of the Fokker–Planck equation. In the It^{o} formulation, a diffusion tensor field is introduced to generalize the Euclidean diffusion matrix, and a tensor-field-based analysis establishes an intrinsic double-divergence representation of the Fokker–Planck equation. The proposed framework provides a globally valid and geometrically consistent interpretation of diffusion and probability transport on Riemannian manifolds, supported by compact and intuitive proofs.

Keywords: Robotics, Modeling, Machine learning
Abstract: Predicting object displacement during tool–soil–object interaction is important for robotic excavation and manipulation tasks, yet existing methods are either computationally expensive or lack physical interpretability. This paper presents a novel three-level physics-informed design that systematically integrates domain knowledge through features, architecture, and loss function to model tool–soil–object interaction including granular dynamics. Trained on discrete element method simulations, our model demonstrates improved generalization to unseen object masses and lengths compared to black-box baselines, achieving lower prediction errors in extrapolative evaluation cases with improvements of up to ≈30% for extreme out-of-distribution scenarios.

Keywords: Identification for control, Energy systems, Modeling
Abstract: This paper presents a control-oriented interconnected thermal model for lithium-ion (Li-ion) battery packs and a corresponding identification framework for estimating coupled thermal parameters. The battery pack is modeled as an electro-thermal network that captures heat-transfer interactions among cells, neighboring units, and cooling channels. The resulting dynamics are expressed as a semi-linear system, where intercell heat conduction is encoded through the graph Laplacian and the control input enters nonlinearly. To identify the control coefficient matrix, a multi-excitation perturbation experiment is introduced that isolates its estimation from that of the system matrix. A reduced-order thermal model is then developed to characterize the dominant group-level thermal behavior of the pack. This reduced model is identified using temporally collected data from a single experiment, enabling efficient characterization of large-scale battery packs. Simulation studies on a 3S3P pack configuration is also presented to validate the accuracy and effectiveness of the proposed modeling and identification framework.

Keywords: Nonlinear system identification, Modeling, Energy systems
Abstract: This paper proposes a measure to quantify the informativity of an identification dataset used for obtaining models for lithium-ion batteries. The proposed informativity measure is inspired by space-filling input design methods and quantifies how well a dataset fills a user-defined region of interest based on the intended model application. To validate the utility of the proposed informativity measure, five different experimental datasets obtained for a 2.85-Ah NMC battery are investigated using a recently proposed linear parameter-varying framework for modelling lithium-ion batteries. It will be shown that the proposed informativity measure for the investigated datasets exhibits an inverse relation with the simulation error values for the battery models obtained using these datasets.

Keywords: Energy systems, Safety critical systems
Abstract: Lithium-ion batteries (LIBs) are widely utilized in electric transportation, energy storage systems, medical devices, and consumer electronics due to their high energy density, low cost, and long lifespan. However, LIBs can experience temperature-related issues, such as performance degradation and thermal runaway. To address these challenges, studies have developed battery thermal models, core temperature monitoring methods, and active cooling controllers, with the main goal of maintaining the core temperature within a safe operating range. However, there are situations where such a safe temperature range may be breached. For instance, during a multi-cell thermal runaway event, the surface temperature of a cell can become extremely high, especially if neighboring cells are already undergoing thermal runaway. In such cases, the system may be placed outside of the safe set, requiring a more advanced multi-layered thermal management controller that can mitigate thermal runaway even when the temperatures exceed the safe limits. To the best of our knowledge, such a multi-layered safe controller has not been developed for LIB core temperature management. This paper addresses this gap by utilizing a graceful safety control framework. The framework places a primary safety layer at the maximum safe temperature, and a secondary failsafe layer at the thermal runaway onset temperature. Then, the controller guarantees that, even if the core temperature violates the primary safety layer, it never breaches the secondary failsafe layer. Simulation studies demonstrate that the proposed controller can maintain a safe core temperature even under extremely high surface temperatures, improving the safety of the battery system.

Keywords: Energy systems, Modeling, Filtering
Abstract: This paper addresses state-of-charge (SOC) estimation for lithium iron phosphate (LFP) batteries, whose relatively flat open-circuit-voltage (OCV)–SOC characteristic leads to poor observability. A residual bias compensation dual extended Kalman filter (RBC-DEKF) is proposed, in which one EKF estimates the electrochemical states of a control-oriented parameter-grouped single-particle model with thermal effects, while the other estimates a residual bias to correct the voltage observation in real time. By decoupling bias estimation from state estimation, the proposed structure avoids the covariance coupling introduced by bias-augmented single-filter schemes. Validation is conducted on an LFP cell from a public dataset under three representative operating conditions: US06 at 0◦C, DST at 25◦C, and FUDS at 50◦C. Compared with a conventional EKF using the same model and identical state-filter settings, the proposed method reduces the average SOC RMSE from 3.75% to 0.20% and the voltage RMSE from 32.8 mV to 0.8 mV. The improvement is most evident in the mid-SOC region, demonstrating that residual bias compensation can significantly enhance physics-based SOC estimation for LFP batteries across a wide temperature rang

Keywords: Energy systems, Observers for nonlinear systems, Identification
Abstract: State of charge (SOC) is one of the most important states that has to be estimated in a battery management system (BMS). SOC represents the available battery capacity and is crucial for avoiding overcharge and overdischarge. The primary obstacle for maintaining accurate SOC estimation over time is aging. Models developed for the fresh cells lose their accuracy with aging, which lowers the precision of the model-based SOC estimation. One critical application of SOC estimation in BMS is its use in SOC-based cell balancing. The decrease in estimation accuracy caused by battery aging adversely affects the performance and reliability of such balancing methods. This work aims at quantifying the impact of aging on the performance of SOC-based balancing for a realistic case study.

Keywords: Optimization algorithms, Energy systems, Machine learning
Abstract: This paper studies the problem of maximizing revenue from a grid-scale battery energy storage system, accounting for uncertain future electricity prices and the effect of degradation on battery lifetime. We formulate this task as an online resource allocation problem. We propose an algorithm, based on online mirror descent, that is no-regret in the stochastic i.i.d. setting and attains finite asymptotic competitive ratio in the adversarial setting (robustness). When untrusted advice about the opportunity cost of degradation is available, we propose a learning-augmented algorithm that performs well when the advice is accurate (consistency) while still retaining robustness properties when the advice is poor.

Keywords: Distributed parameter systems

Keywords: Distributed parameter systems
Abstract: Algebraic Riccati equations have been widely studied, as they play a central role in the analysis and synthesis of optimal control laws for linear systems, both in finite and infinite-dimensional settings. In this paper, we present a comparison theorem for the maximal solutions of algebraic Riccati equations associated to two distinct continuous-time infinite-dimensional control systems. It is a quantitative tool that may be considered for comparing sensor designs in distributed parameter systems for all initial states. The theorem provides conditions under which the corresponding Riccati solutions can be ordered for strongly stabilizable systems. Our result generalizes a well-known comparison principle previously available only for exponentially stabilizable infinite-dimensional systems by extending its validity to strongly stabilizable systems. This extension enlarges the range of applications, to infinite-dimensional systems where exponential stability cannot be guaranteed but strong stabilizability remains achievable.

Keywords: Adaptive control, Distributed parameter systems, Optimization
Abstract: This paper addresses the design and analysis of an extremum-seeking (ES) controller for scalar static maps in the context of infinite-dimensional dynamics governed by complex-valued partial differential equations (PDEs) of Schrödinger type. The system is actuated at one boundary, and the map input is defined as a real-valued quadratic functional corresponding to the squared norm of the complex state at the uncontrolled boundary. An isomorphism between the complex Hilbert space and its two-dimensional real-valued representation is established to enable the use of the standard multivariable Newton-based ES method. To compensate for the PDE actuation dynamics, a boundary control strategy based on a two-step backstepping procedure is employed. With a perturbation-based estimate of the Hessian’s inverse, the local exponential stability to a small neighborhood of the unknown extremum point is proved. A numerical example illustrates the effectiveness of the proposed extremum-seeking methodology.

Keywords: Distributed parameter systems, Optimal control, Computational methods
Abstract: A computational PDE-constrained optimization approach is proposed for optimal trajectory planning under uncertainty by means of an associated Schroedinger Bridge Problem (SBP). The proposed SBP formulation is interpreted as the mean-field limit associated with the energy-optimal evolution of a particle governed by a stochastic differential equation (SDE) with nonlinear drift and reflecting boundary conditions, constrained to prescribed initial and terminal densities. The resulting mean-field system consists of a nonlinear Fokker-Planck equation coupled with a Hamilton-Jacobi-Bellman equation, subject to two-point boundary conditions in time and Neumann boundary conditions in space. Through the Hopf-Cole transformation, this nonlinear system is recast as a pair of forward-backward advection-diffusion equations, which are amenable to efficient numerical solution via a standard finite element discretization. The weak formulation naturally enforces reflecting boundary conditions without requiring explicit particle-boundary collision detection, thus circumventing the computational difficulties inherent to particle-based methods in complex geometries. Numerical experiments on challenging 3D maze configurations demonstrate fast convergence, mass conservation, and validate the optimal controls computed through reflected SDE simulations.

Keywords: Delay systems, Reduced order modeling
Abstract: We consider a nonlinear single machine infinite bus system with delayed measurements operating in closed-loop at a given equilibrium point. The synchronous generator voltage is regulated using a delayed state-feedback controller with integral action, designed via partial pole placement relying on the multiplicity-induced-dominance property and realistic delay measurements. In this paper, we construct low-order linear time-delay models of the resulting closed-loop single machine infinite bus system. We propose a novel delay-aware Loewner matrices-based model reduction method that complements linearization-based approaches. The low-order time-delay models are constructed from input-output frequency-response data such that the measured frequency behavior is matched and a prescribed subset of poles and the measurement delay are enforced. We then indicate how the design parameters (poles, delay, and additional parameters) have been naturally selected. The method is illustrated using a numerical example.

Keywords: Stochastic systems, Distributed parameter systems, Optimal control
Abstract: In this paper, we design a risk-averse controller for an interconnected system composed of a linear Stochastic Differential Equation (SDE) actuated through a linear parabolic heat equation. These dynamics arise in various applications, such as coupled heat transfer systems and chemical reaction processes that are subject to disturbances. While existing optimal control methods for these systems focus on minimizing average performance, this risk-neutral perspective may allow rare but highly undesirable system behaviors. To account for such events, we instead minimize the cost within a coherent risk measure. Our approach reformulates the coupled dynamics as a stochastic PDE, approximates it by a finite-dimensional SDE system, and applies a gradient-based method to compute a risk-averse feedback controller. Numerical simulations show that the proposed controller substantially reduces the tail of the cost distribution, improving reliability with only a minor reduction in average performance.

Keywords: Neural networks, Biomedical systems, Medical signal processing
Abstract: Digital Twin (DT) technology in neuromuscular rehabilitation relies on the ability to generate physiologically realistic muscle activations that reflect true electrophysiological behavior. This work presents a data-driven generative framework capable of synthesizing surface electromyography (sEMG) contractions on demand through a Conditional Variational Autoencoder (cVAE). The model was trained on sEMG signals acquired from upper-limb muscles during controlled physiotherapy-inspired movements derived from the Fugl–Meyer Assessment (FMA). After normalization and augmentation, the cVAE learned a latent representation of muscle activation dynamics conditioned on the executed movement.The results show that the proposed architecture is capable of reproducing the temporal, spectral, and time–frequency characteristics of sEMG signals. The proposed framework can be interpreted as a first step toward a neuromuscular Digital Twin, enabling controllable generation of physiologically consistent muscle activity. Future developments will focus on extending the model to multi-channel configurations and high-density sEMG, as well as enabling subject-specific and real-time integration.

Keywords: Agents and autonomous systems, Biomedical systems, Control education
Abstract: Musicians face a high risk of playing-related musculoskeletal disorders, with violinists and clarinetists among the most affected due to instrument-specific biomechanical demands. Violinists sustain asymmetric postures that chronically overload the neck and shoulders, whereas clarinetists maintain prolonged wrist ulnar deviation and thumb-support strain. Current prevention (teacher observation, self-report) lacks objectivity, continuity and sensitivity to early physiologicalrisk markers. This paper presents a musculoskeletal digital twin frame- work tailored to these instrumental groups. The digital twin integrates surface electromyography and inertial measurement units with a personalized OpenSim musculoskeletal model to es- timate muscle activations, joint torques and cumulative fatigueduring real performance. Real-time outputs are compared to evidence-based, profile-adapted thresholds to trigger escalating alerts (Alert Level 1–3). Although conceptual, the framework relies on validated biomechanics and wearable-sensing methods, establishing a literature-grounded foundation for preventive musician health monitoring and pedagogy.

Keywords: Biomedical systems, Signal processing, Neural networks
Abstract: We present a data-driven surrogate model of aortic hemodynamics that synthesizes aortic pressure and flow waveforms from a compact set of physiological inputs using a conditional Variational Autoencoder (cVAE). The model is trained on single-beat aortic-root signals including time (t), aortic pressure (P ), flow (Q), and cross-sectional area (A), derived from a large population of virtual subjects. The genera- tion process is conditioned on physiological parameters, namely heart rate (HR), stroke volume (SV) scaling, pulse wave velocity (PWV), and arterial diameter scalings for elastic and muscular vessels. After phase normalization and standardization, the cVAE learns a low-dimensional latent representation of cen- tral hemodynamics, enabling the generation of physiologically consistent waveforms for arbitrary parameter combinations. On 3,325 test beats, the model achieves a median pressure root mean square error (RMSE) of 5.5 mmHg (interquartile range: 4.5–6.5 mmHg), with a normalized RMSE (NRMSE) of approximately 9–10% and correlation coefficient r ∼ 0.998. Flow reconstruction shows a median NRMSE of approxi- mately 1–2% with correlation r ∼ 0.999. The model also exhibits minimal bias in central systolic and diastolic blood pressure (cSBP/cDBP), with consistent performance across different HR·SV regimes. The proposed approach enables fast and controllable waveform generation for in-silico “what-if” analysis and data augmentation. Furthermore, it provides a foundation for integrating physics-informed constraints and multi-site signals, supporting the development of personalized and computationally efficient cardiovascular models.

Keywords: Neural networks, Biological systems, Machine learning
Abstract: We introduce the Compartmental Recurrent Neural Network (COMP-RNN), a novel method for modeling glucose-insulin dynamics in type 1 diabetes mellitus patients. By integrating physiological knowledge and topology into recurrent neural networks, the COMP-RNN significantly improves predictive accuracy and parameter efficiency compared to traditional models used in control. Simulated patient data validate its superior performance and demonstrate that the COMP-RNN’s internal states reflect key physiological patterns, proving its potential to improve artificial pancreas systems.

Keywords: Neural networks, Nonlinear system identification, Biological systems
Abstract: Estimating fractional parameters from epidemic data is challenging due to parameter coupling induced by non- local memory kernels and the non-convex optimization landscape of fractional inverse problems. We develop a physics-informed neural network (PINN) framework for estimating fractional parameters in SEIR epidemic models with Power Fractional Derivatives, where the Power Mittag-Leffler kernel introduces an additional power parameter p modulating memory decay. We establish existence, uniqueness, and stability results and derive a numerical discretization scheme. Multi-start ensemble optimization estimates fractional orders α, β, and p from synthetic data, achieving errors below 13% (α: 3.5%, β: 12.2%, p: 8.1%) with R² > 0.96. Systematic evaluation of 24 architectures reveals that shallow, wide networks outperform deep alternatives for fractional systems. This provides the first demonstration that power parameters in Mittag-Leffler epidemic models can be reliably estimated from time series data.

Keywords: Biomedical systems, Nonlinear system identification, Neural networks
Abstract: Understanding brain–heart interaction (BHI) requires models that capture how the central nervous system and the cardiovascular system co-regulate each other under internal and external stressors to preserve homeostasis and give rise to macroscopic physiological states such as sleep, arousal, or vigilance. At the core of this loop are interoceptive variables, defined as latent autonomic control signals that encode the body’s internal state and drive cardiac adjustments; however, these variables are not directly measurable. Recovering these hidden drives from peripheral cardiac signals requires taking into account nonlinear dynamics and physiological confounds, as well as limited measurement data. This work proposes a physics-informed neural estimator for simultaneous state estimation and reconstruction of unknown control inputs in a nonlinear, coupled model of autonomic cardiac regulation. The estimator enforces model-based constraints along with data-driven regularization, and embeds structural identifiability conditions derived from the system itself, yielding guarantees without prescribing a dynamical prior for the unknown input. Validation of stress-evoked cardiac recordings shows the accurate recovery of heart rate, along with estimates of state (being blood pressure) and the unknown control input (being the blood-pressure setpoint); hence, enabling physiology-consistent, left-invertible inference of interoceptive autonomic dynamics.

Keywords: Nonlinear system identification, Stability of nonlinear systems, Optimization algorithms
Abstract: This paper addresses the problem of identifying contractive Lur’e-type systems. Specifically, it proposes an identification framework that integrates linear prior knowledge with a kernel representation of the nonlinear feedback while systematically enforcing contractivity via Lipschitz constant design. The resulting algorithms provide models that are accurate in prediction, interpretable, and faithful to the contractive nature of the true system. Numerical experiments demonstrate that enforcing contractivity significantly improves parameter estimation and yields models that are both accurate and physically meaningful.

Keywords: Nonlinear system identification, Nonlinear system theory, Modeling
Abstract: This paper presents an online Sparse Identification of Nonlinear Dynamics (SINDy) algorithm, which allows for online system identification in response to both abrupt and slowly varying changes in the system dynamics. The approach extends the Extended Kalman Filter (EKF) - SINDy algorithm, inspired by concepts from the abrupt-SINDy method. The proposed Abrupt EKF-SINDy algorithm is validated on simulations of the Selkov and the Lorenz systems, which are nonlinear systems with oscillatory dynamics. The proposed algorithm can recover a model in cases that the EKF-SINDy algorithm and the abrupt-SINDy method find individually challenging. In the simulations, we get lower state estimation root mean square error and parameter estimation error using the proposed Abrupt EKF-SINDy algorithm compared to the standard EKF-SINDy algorithm.

Keywords: Identification, Identification for control, Machine learning
Abstract: We study the identification of continuous-time vector fields from irregularly sampled trajectories. We introduce spectral flow learning, which learns in a windowed flow space using a lag-linear label operator that aggregates lagged Koopman actions. We provide finite-sample, high-probability (FS-HP) guarantees for the class of variable-step linear multistep methods (vLMM). The FS-HP rates are constructed using spectral regularization with qualification-controlled filters for flow predictors under standard source and filter assumptions. A multistep observability inequality links flow error to vector-field error and yields two-term bounds that combine a statistical rate with an explicit discretization bias from vLMM theory. Simulations on a controlled mass--spring system corroborate the theory and clarifies conditioning, step–sample tradeoffs, and practical implications.

Keywords: Identification, Filtering, Linear time-varying systems
Abstract: This paper consider the estimation of unknown time-varying parameters in a linear regression equation. We consider dynamic regressor extension and mixing (DREM) for the parameter estimation and analyse how the time-varying parameter affects the parameter estimation error. Following this, we propose a cascaded DREM estimator, which has a smaller bound on the parameter estimation error compared with standard DREM. We illustrate the findings in a numerical example where the two presented DREM-based methods are compared with standard methods from the literature.

Keywords: Nonlinear system identification, Identification, Sampled data control
Abstract: We propose KoopCL, an online method for identifying Koopman-based lifted linear models using Concurrent Learning (CL) with a compressed memory of informative past lifted regressors. In contrast to Extended Dynamic Mode Decomposition (EDMD), which is primarily a batch identification method, KoopCL performs recursive weighted gradient updates that combine the instantaneous lifted prediction error with a history-dependent correction term constructed from accumulated information matrices. The memory is updated selectively, so that newly arriving data are incorporated only when they improve the information content of the stored lifted regressors. As a result, the method keeps a finite informative memory and requires only recursive matrix updates, which suits real-time implementation. We present the CL and Koopman formulations, describe their integration into a unified online algorithm, and validate the method on a nonlinear power control loop of a grid-following inverter (GFL). Numerical results show that KoopCL learns accurate lifted models online under finite excitation with a performance comparable to EDMD, and can also successfully adapt to system changes. These results support KoopCL as a practical online method for Koopman-based approximation when recursive model updating is required.

Keywords: Sensor and signal fusion, Filtering, Signal processing
Abstract: We propose a probabilistic semantic filtering framework in which parameters of a dynamical system are inferred and associated with a closed set of semantic classes in a map. We extend existing methods to a multi-parameter setting using a posterior that tightly couples semantics with the parameter likelihoods, and propose a filter to compute this posterior sequentially, subject to dynamics in the map's state. Using Bayesian moment matching, we show that the computational complexity of measurement updates scales linearly in the dimension of the parameter space. Finally, we demonstrate limitations of applying existing methods to a problem from the driving domain, and show that the proposed framework better captures time-varying parameter-to-semantic associations.

Keywords: Cooperative control, Distributed cooperative control over networks, Output feedback
Abstract: This paper addresses the longitudinal control problem for a platoon of networked vehicles using negative imaginary (NI) systems theory. By analyzing the longitudinal dynamic model of the vehicle, it is shown that, through state-feedback linearization, the system can be transformed into an output strictly negative imaginary (OSNI) form, which is a subclass of negative imaginary (NI) systems. A positive feedback interconnection structure between the networked OSNI vehicle systems and an output feedback controller is designed over an undirected connected graph to enable tracking of prescribed trajectories, maintaining uniform longitudinal inter-vehicular spacing, and achieving longitudinal velocity consensus across the platoon. In addition, to address platooning in the presence of external disturbances, an input-to-state string stability (ISSS) analysis is performed. This analysis shows that the effect of bounded disturbances is not amplified along the platoon, thereby supporting robust performance for the entire vehicle fleet. Simulation results are presented to validate the effectiveness of the proposed control strategy in achieving stable and disturbance-resilient platooning.

Keywords: Network analysis and control, Transportation systems
Abstract: Recent studies on transportation networks have shown that real-time route guidance can inadvertently induce congestion or oscillatory traffic patterns. Nevertheless, such technologies also offer a promising opportunity to manage traffic non-intrusively by shaping the information delivered to users, thereby mitigating congestion and enhancing network stability. A key step toward this goal is to identify information signals that ensure the existence of an equilibrium with desirable stability and convergence properties. This challenge is particularly relevant when traffic density and routing dynamics evolve concurrently, as increasingly occurs with digital signaling and real-time navigation technologies. To address this, we analyze a parallel-path transportation network with a single origin–destination pair, incorporating joint traffic density and logit-based routing dynamics that evolve at the same timescale. We characterize a class of density-dependent traffic information that guarantees a unique equilibrium in the free-flow regime, ensures its asymptotic stability, and keeps traffic densities within the free-flow region for all time. The theoretical results are complemented by a numerical case study demonstrating how the framework can inform the design of traffic information that reduces total travel time without compromising credibility.

Keywords: Autonomous systems, Optimization algorithms, Uncertain systems
Abstract: The Electric Vehicle Routing Problem (EVRP) is a multi-objective, NP-hard optimization problem focused on efficiently managing various constraints involved in routing single or multiple electric vehicles. Recently, several evolutionary algorithms have been applied to EVRP, particularly those that incorporate battery State of Health (SoH) as a key constraint in the optimization process. However, most existing optimization algorithms treat SoH deterministically and do not account for the uncertainty inherent in its estimation. As a result, these methods may yield solutions that do not accurately reflect real-world battery conditions. To address this limitation, we propose a hybrid genetic algorithm that incorporates the stochastic process distribution of battery SoH estimation into the optimization procedure. Results from a lithium-ion battery-based fleet scenario show that incorporating uncertainty into the optimization improves battery SoH prediction accuracy by up to 70% compared to baseline methods.

Keywords: Predictive control for linear systems, Agents and autonomous systems, Distributed control
Abstract: This paper addresses the problem of safely coordinating a platoon of autonomous vehicles during outdoor missions when the communication between each vehicle and its successor, as well as with remote distributed control units, is compromised by malicious agents. In particular, this operating scenario gives rise to the following methodological issues: capability of the distributed control algorithm to be feasible until the mission is accomplished and effectiveness of adequate countermeasures to contrast the anomalous dynamical system behavior resulting from unpredictable attack phenomena. Here, model predictive control techniques coupled with the so-called perturbation analysis will be exploited to comply with these objectives, while the concept of digital twin will be used to formally reveal anomalous actions on the data shared amongst the involved vehicles.

Keywords: Transportation systems, Cooperative control, Automotive
Abstract: This paper deals with the safe vehicles platoon control problem with safety space constraint guarantees in the presence of nonlinear and heterogeneous unmodeled dynamics, parameter uncertainties, external disturbing factors and state constraints. This challenging framework is modeled via the Multi-Agent Systems and a proper transformation allows to recast the constrained problem into an unconstrained one, on the basis of which a novel distributed adaptive backstepping-based control law is designed. The stability of the overall closed-loop vehicular network is proved by leveraging the Lyapunov theory and specific barrier functions. This combination allows to derive the adaptive mechanisms and, at the same time, prevents spacing constraint violations, whilst strengthening the safety due to the confinement of the spacing errors within a preset bound both in transient and steady-state phases. Numerical simulations confirm the effectiveness of the theoretical findings.

Keywords: Computer networks, Output regulation, Transportation systems
Abstract: Transmission Control Protocol (TCP) congestion control must address complex nonlinear dynamics and network delays, rendering traditional control methods less effective. Funnel control provides a model-free mechanism for imposing strict, pre-defined performance constraints on nonlinear systems; however, its application to actual TCP dynamics remains in the exploratory stage. This paper investigates the funnel control of a relative-degree two TCP model, with a specific focus on output-dependent delays - a delay structure that is both practically significant and theoretically challenging. We design a funnel control law tailored to this delay structure and establish corresponding theoretical results for the closed-loop system. Simulation results demonstrate the behavior of the controller and expose oscillatory phenomena near the funnel boundaries - an open challenge that motivates future work on providing a rigorous global stability proof for such complex, relative degree two systems.

Keywords: Optimal control, Predictive control for nonlinear systems, Machine learning
Abstract: This introductory talk establishes the foundational bridge between Model Predictive Control (MPC) and Reinforcement Learning (RL). Starting from a unified Markov Decision Process (MDP) formulation and consistent notation, we provide a high-level overview of both paradigms: examining RL through value functions and policies, and MPC via receding-horizon optimization. By highlighting recent success stories, we present a systematic comparison of their respective strengths in handling constraints, model mismatch, stochasticity, and online computational requirements. This presentation serves as the technical primer for the advanced synthesis and software topics covered in the remainder of the tutorial.

Keywords: Optimal control, Predictive control for nonlinear systems, Machine learning
Abstract: This talk provides a structured classification of hybrid Model Predictive Control (MPC) and Reinforcement Learning (RL) algorithms. We introduce the core learning and inference architectures—ranging from hierarchical to integrated structures—that define how neural networks interface with optimization layers. Using the actor-critic framework, we categorize these methods by the functional role of MPC: as an expert for guidance, a learnable critic, or the deployed policy itself. Finally, we contrast aligned learning, which maintains physical interpretability, with closed-loop learning, which optimizes for end-to-end performance. This overview offers a clear roadmap for navigating the diverse landscape of modern MPC-RL synthesis.

Keywords: Machine learning, Optimal control, Predictive control for nonlinear systems
Abstract: This talk establishes the theoretical foundations connecting Model Predictive Control and Reinforcement Learning approaches through the unifying lens of Markov Decision Processes (MDPs). We demonstrate how MPC can be formally viewed as an approximate solution to the underlying MDP, and identify the key paradigm shifts that enable their effective combination: moving from fitting dynamics models to optimizing closed-loop performance, and adopting a “holistic” parametrization where the entire MPC scheme - including cost functions, constraints, and terminal conditions - becomes learnable.

Keywords: Optimal control, Optimization
Abstract: The efficient computation of parametric solution sensitivities is a key challenge in the integration of learning-enhanced methods with nonlinear model predictive control, as their availability is crucial for many learning algorithms. In this talk, we discuss the computation of solution sensitivities of general nonlinear programs using the implicit function theorem and smoothed optimality conditions as used in interior-point methods. Furthermore, we provide a detailed analysis for sensitivity computation within a sequential quadratic programming method which employs an interior point method for the quadratic subproblems. As a practical example, we present the efficient open-source implementation within the texttt{acados} framework, providing both forward and adjoint sensitivities for general optimal control problems, achieving speedups exceeding 3x over the state-of-the-art solvers texttt{mpc.pytorch} and texttt{cvxpygen}. The talk is based on the preprint arxiv.org/abs/2505.01353.

Keywords: Optimization, Predictive control for nonlinear systems, Machine learning
Abstract: While MPC-RL research is ongoing, available software tools for MPCRL methods are scarce. In this talk, we present the software package leap-c (Learning Predictive Control), which extends the domain of said tools. It leverages acados' capabilities to provide a fast, versatile and differentiable MPC layer as a PyTorch module. Aside of this core functionality, the package provides some examples and different instances of MPC-RL algorithms, like combining Soft Actor-Critic (SAC) with an MPC controller in a hierarchical structure. We discuss these algorithms in more detail, highlighting the different design choices that lead to them.

Keywords: Iterative learning control
Abstract: The explosive growth of machine learning and data-driven methodologies has revolutionized numerous fields. Yet, translating these successes to dynamical physical systems remains a significant challenge, hindered by the complexity, uncertainty, and safety-critical nature of such environments. In this talk, we present a unified framework that bridges this gap by introducing novel generative representations for reinforcement learning and control. On the critic side, we develop a structured representation of system dynamics that focuses on modeling how actions influence future state distributions. This transition-based perspective enables the design of nonlinear stochastic control and reinforcement learning algorithms that are efficient, safe, robust, and scalable, with provable guarantees. On the actor side, we represent stochastic feedback policies using diffusion-based generative models, treating control as a generative process. This approach leads to new methods for policy optimization, while providing a flexible and expressive framework for decision-making in dynamical systems. We further demonstrate how these representations help close the sim-to-real gap, improve data efficiency in imitation learning, and enable scalable computation of localized policies for large-scale nonlinear networked systems, with applications including robotics and energy systems.

Keywords: Emerging control theory
Abstract: Recent cyberattacks on critical infrastructure—most notably the Industroyer cyberattack—have exposed the vulnerability of modern energy systems to adversarial manipulation of control and monitoring components. While conventional control systems are designed to handle disturbances and faults, coordinated cyberattacks pose fundamentally different challenges that existing safety mechanisms only partially address. This has motivated the development of resilient control frameworks that ensure graceful performance degradation and recovery under attack. This lecture takes a control-theoretic perspective on modeling, detecting, and mitigating attacks in cyber-physical systems. We review representative attack scenarios and discuss system architectures, constraints, and the evolving attack surface. We then present model-based tools to characterize fundamental limits of detection and mitigation, and conclude with recent advances in data-driven approaches for improving security and resilience.

Keywords: Machine learning, Neural networks, Nonlinear system theory
Abstract: Metriplectic conditional flow matching (MCFM) learns dissipative dynamics without violating first principles. Neural surrogates often inject energy and destabilize long-horizon rollouts; MCFM instead builds the conservative–dissipative split into both the vector field and a structure-preserving sampler. MCFM trains via conditional flow matching on short transitions, avoiding long-rollout adjoints. In inference, a Strang-prox scheme alternates a symplectic update with a proximal metric step, ensuring discrete energy decay; an optional projection enforces strict decay when a trusted energy is available. We provide continuous- and discrete-time guarantees linking this parameterization and sampler to conservation, monotonic dissipation, and stable rollouts. On a controlled mechanical benchmark, MCFM yields phase portraits closer to ground truth and markedly fewer energy-increase and positive energy-rate events than an equally expressive unconstrained neural flow, while matching terminal distributional fit.

Keywords: Output regulation, Neural networks, Emerging control applications
Abstract: This work addresses the full-information output regulation problem for nonlinear systems, assuming the states of both the plant and the exosystem are known. In this setting, perfect tracking or rejection is achieved by constructing a zero-regulation-error manifold π(w) and a feedforward input c(w) that render such manifold invariant. The pair (π(w), c(w)) is characterized by the regulator equations, i.e., a system of PDEs with an algebraic constraint. We focus on accurately solving the regulator equations by introducing a physics-informed neural network (PINN) approach that directly approximates π(w) and c(w) by minimizing the residuals under boundary and feasibility conditions, without requiring precomputed trajectories or labeled data. The learned operator maps exosystem states to steady state plant states and inputs, enables real-time computation and, critically, generalizes across families of the exosystem with varying initial conditions and parameters. The framework is validated on a regulation task that synchronizes a helicopter’s vertical dynamics with a harmonically oscillating platform. The resulting PINN-based controller reconstructs the zero-error manifold with high fidelity and sustains regulation performance under exosystem variations, highlighting the potential of learning-enabled solvers for nonlinear output regulation. The proposed approach is broadly applicable to nonlinear systems that admit a solution to the output regulation problem.

Keywords: Cooperative control, Neural networks, Distributed cooperative control over networks
Abstract: Cooperative control of multi-agent systems with dynamic communication topologies presents a significant challenge in distributed automation and intelligent transportation. This study introduces a recurrent neural network (RNN) based approach for adaptive tuning of control matrices in cooperative systems. For each topology, a gated recurrent unit (GRU) based RNN is trained to compute control parameters based on the current global state. These trained networks are integrated into a coalitional control framework that adapts both the control strategy and the communication topology in real time, thereby improving system performance and reducing communication costs. A comparative analysis with a linear matrix inequality (LMI) based method demonstrates the efficiency of the proposed approach. Validation on a vehicle platooning scenario indicates the method's potential for application to nonlinear cooperative systems. This represents the first integration of GRU-based recurrent models within a coalitional control framework for cooperative systems with switching topologies.

Keywords: Neural networks, Machine learning
Abstract: Deep neural networks (DNNs) offer significant flexibility and robust performance. This makes them ideal for building not only system models but also advanced neural network controllers (NNCs). However, their high complexity and computational needs often limit their use. Various model compression strategies have been developed over the past few decades to address these issues. These strategies are effective for general DNNs but do not directly apply to NNCs. NNCs need both size reduction and the retention of key application-specific performance features. In structured pruning, which removes groups of related elements, standard importance metrics often fail to protect these critical characteristics. In this paper, we introduce a novel framework for calculating importance metrics in pruning groups. This framework not only shrinks the model size but also considers various application-specific constraints. To find the best pruning coefficient for each group, we evaluate two approaches. The first approach involves simple exploration through grid search. The second utilizes gradient descent optimization, aiming to balance compression and task performance. We test our method in two use cases: one on an MNIST autoencoder and the other on a Temporal Difference Model Predictive Control (TDMPC) agent. Results show that the method effectively maintains application-relevant performance while achieving a significant reduction in model size.

Keywords: Optimal control, Neural networks, Machine learning
Abstract: It is well known that deep learning, and in particular the training of ResNets and neural ODEs, can be formalized and analyzed from an optimal control perspective. In this work, we extend the dynamic systems and optimal control perspectives to fully connected neural networks with ReLU activations and no skip connections. By exploiting equivalence relations between ResNets and networks without skip connections, we show that the corresponding training problems exhibit the turnpike property under conditions analogous to those established for ResNets. We illustrate our findings with numerical experiments for the two-spirals dataset and MNIST.

Keywords: Stability of nonlinear systems, Optimal control, Neural networks
Abstract: The growing complexity of modern control tasks calls for controllers that can react online as objectives and disturbances change, while preserving closed-loop stability. Recent approaches for improving the performance of nonlinear systems while preserving closed-loop stability rely on time-invariant recurrent neural-network controllers, but offer no principled way to update the controller during operation. Most importantly, switching from one stabilizing policy to another can itself destabilize the closed-loop. We address this problem by introducing a stability-preserving update mechanism for nonlinear, neural-network-based controllers. Each controller is modeled as a causal operator with bounded ℓ_p-gain, and we derive gain-based conditions under which the controller may be updated online. These conditions yield two practical update schemes, time-scheduled and state-triggered, that guarantee the closed-loop remains ℓ_p-stable after any number of updates. Our analysis further shows that stability is decoupled from controller optimality, allowing approximate or early-stopped controller synthesis. We demonstrate the approach on nonlinear systems with time-varying objectives and disturbances, and show consistent performance improvements over static and naive online baselines while guaranteeing stability.

Keywords: Predictive control for nonlinear systems, Stochastic control, Energy systems
Abstract: This paper proposes a method to find probabilistic guarantees for a scenario-based stochastic economic model predictive control (SEMPC) scheme with an empirical expected shortfall (EES) constraint. The objective function includes the minimisation of an average economic cost, desirable from an operational perspective, while keeping the risk of high costs under control through satisfaction of the EES constraint. The pick-to-learn (P2L) method is employed to find guarantees for the solution of the proposed SEMPC problem. The suggested framework is applied to a water distribution network, and the results show that a balance between economic performance and probabilistic guarantees is achieved.

Keywords: Predictive control for nonlinear systems, Machine learning, Robotics
Abstract: Learning-based control techniques use data from past trajectories to control systems with uncertain dynamics. However, learning-based controllers are often computationally inefficient, limiting their practicality. To address this limitation, we propose a learning-based controller that exploits differential flatness, a property of many robotic systems. Recent research on using flatness for learning-based control either is limited in that it (i) ignores input constraints, (ii) applies only to single-input systems, or (iii) is tailored to specific platforms. In contrast, our approach uses a system extension and block-diagonal cost formulation to control general multi-input, nonlinear, affine systems. Furthermore, it satisfies input and half-space flat state constraints and guarantees probabilistic Lyapunov decrease using only two sequential convex optimizations. We show that our approach performs similarly to, but is multiple times more efficient than, a Gaussian process model predictive controller in simulation, and achieves competitive tracking in real hardware experiments.

Keywords: Predictive control for nonlinear systems, Optimization algorithms, Autonomous systems
Abstract: Applying model predictive control on embedded systems remains challenging due to the high computational cost of solving optimal control problems. To address this limitation, computationally efficient Gaussian process approximations of the implicit model predictive control law can be employed. However, for trajectory-tracking applications, the large amount of training data required for successful generalization across distinct reference trajectories poses a significant challenge. To improve data efficiency, we propose to transform the model into curvilinear coordinates around the reference trajectory. Secondly, we introduce a nominal feedforward component, allowing the Gaussian process to learn only the residual control input, making the approximation of a trajectory-tracking controller feasible. To underline the applicability of the approach, we deploy the controller on a Raspberry Pi in a small-scale vehicle and validate it experimentally. Compared to a model predictive control implementation using real-time iterations, the Gaussian process based approximation computes control inputs about five times faster while achieving similar closed-loop tracking performance.

Keywords: Switched systems, Predictive control for linear systems, Machine learning
Abstract: This paper discusses a risk-averse approximate MPC formulation for a class of Markov Jump Linear Systems with potentially varying transition probabilities. To safeguard against these so-called distribution shifts, we adopt a distributionally robust optimization perspective and robustify the stochastic optimal control problem in the space of probability distributions. Finally, to account for real-time feasibility, we use imitation learning to train a Neural Network policy that approximates the optimal control problem. By employing robust model predictive control mechanisms we can apply Chernoff's inequality to achieve probabilistic feasibility and stability guarantees under the approximate control policy.

Keywords: Optimal control, Machine learning, Predictive control for linear systems
Abstract: Model Predictive Control evolved as the state of the art paradigm for safety critical control tasks. Control-as-Inference approaches thereof model the constrained optimization problem as a probabilistic inference problem. The constraints have to be implemented into the inference model. A recently introduced physics-informed Gaussian Process method uses Control-as-Inference with a Gaussian likelihood for state constraint modeling, but lacks guarantees of open-loop constraint satisfaction. We mitigate the lack of guarantees via an additional sampling step using Hamiltonian Monte Carlo sampling in order to obtain safe rollouts of the open-loop dynamics which are then used to obtain an approximation of the truncated normal distribution which has full probability mass in the safe area. We provide formal guarantees of constraint satisfaction while maintaining the ODE structure of the Gaussian Process on a discretized grid. Moreover, we show that we are able to perform optimization of a quadratic cost function by closed form Gaussian Process computations only and introduce the Matérn kernel into the inference model.

Keywords: Predictive control for nonlinear systems, Behavioural systems, Autonomous robots
Abstract: Data-Driven Predictive Control (DPC) optimizes system behavior directly from measured trajectories without requiring an explicit model. However, its computational cost scales with dataset size, limiting real-time applicability to nonlinear robotic systems. For robotic tasks such as trajectory tracking and motion planning, real-time feasibility and numerical robustness are essential. Nonlinear DPC often relies on large datasets or learned nonlinear representations to ensure accuracy, both of which increase computational demand. We propose Contextual Sampling, a dynamic data selection strategy that adaptively selects the most relevant trajectories based on the initial output trajectory and future reference. By reducing dataset size while preserving representativeness, it improves computational efficiency. Experiments on a scaled autonomous vehicle and a quadrotor show that Contextual Sampling achieves comparable or better tracking than Random Sampling with fewer trajectories, improving online feasibility. Compared with Select-DPC [1], it achieves similar tracking accuracy at lower computational cost. Contextual Sampling and Select-DPC were developed independently and concurrently. In comparison with the full DPC formulation without sampling, Contextual Sampling attains comparable tracking performance while requiring less computation, highlighting the benefit of efficient data selection in data-driven predictive control.

Keywords: Distributed cooperative control over networks, Lyapunov methods, Cooperative autonomous systems
Abstract: In this paper, we present a novel distributive Lyapunov-based guidance law for the cooperative circumnavigation of a stationary target by a heterogeneous system of unicycle pursuers. We address the challenging scenario where the location of the target is known only to a subset of the pursuers, based on their sensing radii and initial positions. For the uninformed pursuers, an estimation method is integrated into the control design. This estimation method requires only the heading angle of a pursuer's out-neighbour and the inter-agent distance between them. Under specific communication graph conditions, controlling only the angular speeds, the proposed guidance law ensures that all pursuers converge to their pre-assigned orbital radii. The efficacy and versatility of the approach are demonstrated through numerical simulations, showcasing its ability to achieve both rigid and non-rigid body formations and further validated experimentally on the Turtlebot3 Burger robots.

Keywords: Optimal control of communication networks, Distributed control, Stochastic systems
Abstract: To bridge the gap between idealised communication models and the stochastic reality of networked systems, we introduce a framework for embedding asynchronous communication directly into algorithm dynamics using stochastic differential equations (SDE) driven by Poisson Jumps. We apply this communication-aware design to the continuous-time gradient flow, yielding a distributed algorithm where updates occur via sparse Poisson events. Our analysis establishes communication rate bounds for asymptotic stability and, crucially, a higher, yet sparse, rate that provably any desired exponential convergence performance slower than the nominal, centralized flow. These theoretical results, shown for unconstrained quadratic optimisation, are validated by a numerical simulation.

Keywords: Cooperative control, Sliding mode control, Robust control
Abstract: This paper investigates a distributed consensus control for a class of multi-agent systems (MAS) based on a new adaptive supertwisting (ASTW) control with auto-tuning gains. The main objective is to design a robust consensus controller for MASs without any knowledge of perturbation bounds, besides having an improved adaptability against disturbances with different dynamics, thanks to the auto-tuning feature. In this method, the proposed protocols and the parameters are adapted and tuned based on the estimated data from the system. Using Lyapunov's stability theory, the convergence of the sliding variables to a real second-order sliding mode (2-SM) is achieved in finite time, such that there is no overestimation of the gains. Some numerical simulations validate the robustness and effectiveness of the designed method.

Keywords: Distributed control, Control over networks, Communication networks
Abstract: In this paper, an approach for opacity enforcement is developed for distributed control systems. In distributed control systems, the agents are communicating their state information among each other to achieve a predefined control objective. The true state of the agents may be the secret to be protected from an adversary that eavesdrops the communication channels. Motivated by this, mask signals are added into the true state information transmitted by the agents over the network. An approach is given to design the mask signals to keep the error between the transmitted state estimate and the true state estimate above a prespecified lower bound and make the trajectory of the transmitted state estimate very different from the trajectory of the true state estimate. In order to keep the control performance of distributed control systems unchanged, the mask signals are selected in the right null space of the feedback gain matrices of the agents. The proposed opacity enforcement approach is illustrated by a simulation example.

Keywords: Distributed control, Optimal control of communication networks, Linear systems
Abstract: In this paper, we address the distributed Linear-Quadratic-Gaussian optimal control problem for discrete-time networked systems. The feedback controller consists of local control stations, each receiving partial measurement data from the plant and regulating a portion of the input signal. We propose a novel solution based on a reduced communication protocol, capable of achieving performance levels arbitrarily close to those of the optimal centralized approach, provided the number of consensus iterations is sufficiently large.

Keywords: Decentralized control, Distributed control, Distributed cooperative control over networks
Abstract: Multi-agent target tracking in the presence of nonlinear dynamics and agent heterogeneity, where state-space dimensions may differ, is a challenging problem that traditional graph Laplacian methods cannot easily address. This work leverages the framework of cellular sheaves, a mathematical generalization of graph theory, to natively model such heterogeneous systems. While existing coordination sheaf frameworks focus on cooperative problems like consensus, this work extends them to the non-cooperative target-tracking problem. The tracking of multiple, unknown targets is formulated as a harmonic extension problem on a cellular sheaf, accommodating nonlinear dynamics and external disturbances for all agents. A decentralized control law is developed using the sheaf Laplacian, and a corresponding Lyapunov-based stability analysis is provided to guarantee tracking error convergence, with results validated by simulation.

Keywords: Markov processes, Identification, Optimization algorithms
Abstract: We address the problem of identifying the dynamical law governing the evolution of a population of indistinguishable particles, when only aggregate distributions at successive times are observed. Assuming a Markovian evolution on a discrete state space, the task reduces to estimating the underlying transition probability matrix from distributional data. We formulate this inverse problem within the framework of entropic optimal transport, as a joint optimization over the transition matrix and the transport plans connecting successive distributions. This formulation results in a convex optimization problem, and we propose an efficient iterative algorithm based on the entropic proximal method. We illustrate the accuracy and convergence of the method in two numerical setups, considering estimation from independent snapshots and estimation from a time series of aggregate observations, respectively.

Keywords: Robust control, Uncertain systems, Optimization algorithms
Abstract: We present a variational free-energy formulation for distributionally robust decision-making with ambiguity in the generative model. The formulation, related to a broad range of learning and control frameworks, yields a minimax optimal control problem where maximization is over an uncertainty set that represents ambiguities. We prove that computing the optimal policy requires solving a non-convex minimization problem and propose an algorithm with convergence guarantees to find the solution. The effectiveness of our results is illustrated via simulations on a pendulum swing-up problem.

Keywords: Game theoretical methods, Agents and autonomous systems, Uncertain systems
Abstract: In many multiagent settings, such as electric vehicle charging and traffic routing, agents must make decisions in the face of uncertain behavior exhibited by others. Often, this uncertainty arises from multiple sources, such as incomplete information, limited computation, or bounded rationality, ultimately impacting the aggregate behavior. To tackle this challenge, we follow recent work on strategically robust game theory and postulate that agents seek protection directly against deviations around the emergent behavior, as opposed to explicitly modeling all sources of uncertainty. Specifically, we propose that each agent protects itself against the worst-case aggregate behavior within an optimal-transport-based ambiguity set centered at the emergent aggregate population behavior. This leads to a novel equilibrium concept, called strategically robust Wardrop equilibrium, that enables one to interpolate between standard Wardrop equilibria (no robustness) and security strategies (maximum robustness). In the setting of convex aggregative games, we establish the existence of a pure strategically robust Wardrop equilibrium and provide tractable computational tools for computing it. Through an application in electric vehicle charging, we demonstrate that strategically robust Wardrop equilibria lead to better decisions, protecting agents against the uncertain aggregate behavior of the population. Remarkably, we also observe that strategic robustness can lead to lower equilibrium costs for all agents, uncovering a “coordination-via-robustification” effect.

Keywords: Machine learning, Large-scale systems, Agents and autonomous systems
Abstract: We propose a Reinforcement Learning framework for sparse indirect control of large-scale multi-agent systems, where few controlled agents shape the collective behavior of many uncontrolled agents. The approach addresses this multi-scale challenge by coupling ODEs (modeling controlled agents) with a PDE (describing the uncontrolled population density), capturing how microscopic control achieves macroscopic objectives. Our method combines model-free Reinforcement Learning with adaptive interaction strength compensation to overcome sparse actuation limitations. Numerical validation demonstrates effective density control, with the system achieving target distributions while maintaining robustness to disturbances and measurement noise, confirming that learning-based sparse control can replace computationally expensive online optimization.

Keywords: Optimal control, Reduced order modeling, Predictive control for nonlinear systems
Abstract: In this paper, we propose a suboptimal and reduced-order Model Predictive Control (MPC) architecture for discrete-time feedback-interconnected systems. The numerical MPC solver: (i) acts suboptimally, performing only a finite number of optimization iterations at each sampling instant, and (ii) relies only on a reduced-order model that neglects part of the system dynamics, either due to unmodeled effects or the presence of a low-level compensator. We prove that the closed-loop system resulting from the interconnection of the suboptimal and reduced-order MPC optimizer with the full-order plant has a globally exponentially stable equilibrium point. Specifically, we employ timescale separation arguments to characterize the interaction between the components of the feedback-interconnected system. The analysis relies on an appropriately tuned timescale parameter accounting for how fast the system dynamics are sampled. The theoretical results are validated through numerical simulations on a mechatronic system consisting of a pendulum actuated by a DC motor.

Keywords: Game theoretical methods, Machine learning, Adaptive systems
Abstract: We reinterpret Generative Adversarial Networks (GANs) as Cournot-type duopoly games, where the generator and real-data source act as competing producers and the discriminator serves as a market-price oracle. Within this analogy, we (i) map discriminator outputs to per-unit prices recovering divergence-based objectives, (ii) show that an optimal discriminator yields a potential-game formulation, and (iii) derive local stability conditions for gradient-play dynamics in an aggregate model. Numerical simulations demonstrate the resulting learned-market discriminator framework.

Keywords: Autonomous robots, Differential algebraic systems, Safety critical systems
Abstract: This paper investigates the problem of safe and stable control of a point robot to navigate through a two-dimensional environment containing unknown convex obstacles. The robot can obtain only local sensor information about nearby obstacles that it encounters. To guarantee collision avoidance and mitigate chattering effects, we introduce the concept of virtual rigid obstacles that serve as safety buffers during the robot's navigation. The resulting dynamics are modeled as a differential complementarity system. We illustrate the controller's performance with numerical simulations of a robot that must navigate around a single elliptical obstacle.

Keywords: Autonomous robots, Agents and autonomous systems, Machine learning
Abstract: Gaussian Processes (GPs) are powerful non-parametric Bayesian models for scalar-field regression, typically assuming perfectly known measurement locations and Gaussian measurement noise. In many real-world mapping applications, however, sensor-equipped mobile robots collect measurements under imperfect localization, causing discrepancies between estimated and true measurement locations that degrade GP mean and covariance estimates. To address this issue, we propose a method for updating GP models as improved location estimates become available. By exploiting the differentiability of the GP kernel, we develop a second-order correction algorithm based on precomputed Jacobians and Hessians of the GP mean and covariance, enabling real-time refinement from measurement-location discrepancy data. Simulation results show improved prediction accuracy and lower computational cost than full retraining.

Keywords: Computer aided control design, Optimization algorithms, Robotics
Abstract: Bayesian optimization (BO) is increasingly used for system commissioning, but it usually requires a scalar objective that may be hard to obtain or may not fully capture expert judgement. We investigate preferential Bayesian optimization (PBO), which instead relies only on pairwise expert feedback, for commissioning a robotic planar pushing task. State-of-the-art PBO methods are evaluated in simulation and on a real robot. The results show that PBO can identify controller settings that satisfy an expert using only binary preferences. We further show that an expert-designed cost function is not fully consistent with the expert’s own choices. In contrast, a cost function learned from PBO preferences better matches the expert’s decisions and can be reused in conventional BO.

Keywords: Sensor and signal fusion, Autonomous robots, Neural networks
Abstract: In the era of autonomous vehicles, state estimation is crucial for planning and control. In this paper, improvements for the filtering of wheel odometry & GNSS measurements are proposed to increase the accuracy in the challenging city driving localization task. Although the wheel encoder-based estimation is a robust and cost-effective localization method in passenger vehicles, the accuracy is limited by the parameter uncertainty, thus the proposed method contains the online calibration of model parameters. The avoidance of pose estimation divergence can be guaranteed with accurate GNSS observations, thus the estimation of the current measurement noise is also considered. This information is fed to a neural network to predict optimal covariances for a Kalman-filter. The effectiveness of the proposed method through experimental measurement is illustrated, which is saved in urban driving, where the standard deviation of the GNSS is in the meter range. Furthermore, the proposed method can operate with any kind of noisy pose measurements, e.g. from vision even indoor, thus it has a wide range of practical applications in the field of autonomous robotics.

Keywords: Autonomous robots, Robotics, UAV's
Abstract: This paper presents Coverage-RRT⋆, a coverage planner that combines coverage-aware sampling with primitive-based steering to generate dynamically feasible trajectories, avoiding a costly global re-optimization. The planner biases tree growth towards uncovered high-utility regions, and efficiently generates trajectories, significantly accelerating convergence. Coverage-RRT⋆ is then further extend, introducing an anytime execution framework tailored for dynamic environments. A cost-to-go heuristic guides the tree expansion by estimating the remaining effort to complete coverage; while unpromising subtrees are discarded through a pruning mechanism. At each iteration, only a short portion of the planned trajectory is executed by the robot; meanwhile the planner continues to refine the remainder of the plan as new obstacles are sensed or the map changes. This framework enables online adaptation starting from partial prior knowledge of the environment, and supports plan improvements during its execution.

Keywords: UAV's, Autonomous robots, Agents and autonomous systems
Abstract: In this work we consider a multi-robot team operating in an unknown environment where one aerial agent is tasked to map the environment and transmit (a portion of) the mapped environment to a group of ground agents that are trying to reach their goals. The entire operation takes place over a bandwidth-limited communication channel, which motivates the problem of determining what and how much information the assisting agent should transmit and when while simultaneously performing exploration/mapping. The proposed framework enables the assisting aerial agent to decide what information to transmit based on the Value-of-Information (VoI), how much to transmit using a Mixed-Integer Linear Programming (MILP), and how to acquire additional information through an utility score-based environment exploration strategy. We perform a communication-motion trade-off analysis between the total amount of map data communicated by the aerial agent and the navigation cost incurred by the ground agents.

Keywords: Differential algebraic systems, Optimal control, Modeling
Abstract: Underactuated robots are characterized by a larger number of degrees of freedom than actuators and, if they are designed with a specific mass distribution, they can be controlled by means of differential flatness theory. This dynamic property enables the development of lightweight and cost-effective robotic systems with enhanced dexterity. However, a key challenge lies in managing the elastic-passive joints, whose control demands precise and comprehensive dynamic modeling of the system. To simplify dynamic models, particularly for low-speed trajectories, friction is often neglected. While this assumption simplifies analysis and control design, it introduces residual oscillations of the end-effector about the target position. In this paper, the possibility of using optimal control along with differential flatness control is investigated to improve the tracking of the planned trajectories. First, the study was carried out through formal analysis, and then, it was validated by means of numerical simulations.

Results highlight that optimal control can be used to plan the flat variables considering different (quadratic) performance indices: control effort, i.e. motor torque, and potential energy of the considered underactuated joint. Moreover, the minimization of potential energy can be used to design motion laws that are robust against variation of the stiffness and damping of the underactuated joint, thus reducing oscillations in the case of stiffness/damping mismatch.

Keywords: Feedback linearization, Algebraic/geometric methods, Nonlinear system theory
Abstract: We investigate the minimal order of input prolongations required for control systems to achieve differential flatness by pure prolongation.

Using the necessary and sufficient conditions recently established for pure-prolongation flatness, we revisit the classical two-input case and provide a simplified proof of the known tight bound 2n − 3.

We then extend the analysis to multi-input systems with two or more controls, showing that each input requires at most 2n − 3 prolongations, leading to the minimal total bound (m − 1)(2n − 3).

These results generalize previous findings and clarify the geometric structure underlying flatness by pure prolongation. The bound is being proven to be sharp by an example.

Keywords: Nonlinear system theory, Algebraic/geometric methods, Feedback linearization
Abstract: In this contribution we discuss flat discrete-time nonlinear systems in a general setting including two special subclasses, namely, forward- and backward-flat systems. We relate rank conditions for certain submatrices of the Jacobian of the flat parameterization to the mentioned subclasses. Motivated by these rank conditions, for the case of two-input systems that possess an (x,u)-flat output, we derive a simple type of dynamic extension for the purpose of an exact linearization.

Keywords: Control over networks, Network analysis and control, Nonlinear system theory
Abstract: Adaptive bitrate streaming (ABR) over the HyperText Transfer Protocol (HTTP), which raises numerous delicate questions, is nowadays almost the only approach to video streaming. This paper presents elementary solutions to three key issues: 1) A straightforward feedforward control strategy for the bitrate and the buffer level via flatness-based control. 2) Closing the loop permits mitigating unavoidable mismatches and disturbances, such as Internet fluctuations. This is adapted from the new HEOL setting, which mixes model-free and flatness-based controls. 3) An easily implementable closed-form estimate of the bandwidth via algebraic identification techniques is derived, perhaps for the first time. It permits handling severe variations in channel capacity. Several computer experiments and metrics for evaluating the Quality of Experience (QoE) are displayed and discussed.

Keywords: Nonlinear system theory, Algebraic/geometric methods, Feedback linearization
Abstract: Determining whether a nonlinear multi-input system is differentially flat remains challenging. One way to obtain computationally tractable sufficient conditions is to give complete characterizations of flat normal forms. We introduce a structurally flat triangular form for control-affine systems with at least three inputs that is based on a multi-chained form. For two specific instances of this structure, we provide complete geometric characterizations, i.e., necessary and sufficient conditions under which a control-affine system is static-feedback equivalent to the respective triangular form. These characterizations yield sufficient conditions for differential flatness and, in turn, constructive procedures for computing flat outputs.

Keywords: Feedback linearization, Nonlinear system theory, Computational methods
Abstract: Differential flatness serves as a powerful tool for controlling continuous time nonlinear systems in problems such as motion planning and trajectory tracking. A similar notion, called difference flatness, exists for discrete-time systems. Although many control systems evolve in continuous time, control implementation is performed digitally, requiring discretization. It is well known in the literature that discretization does not necessarily preserve structural properties, and it has been established that, in general, flatness is not preserved under discretization (whether exact or approximate). In this paper, inspired by our previous work [Jindal et al., LCSS'23] and based on the notion of discretization maps, we construct numerical schemes that preserve flatness.

Keywords: Control education, Control courses and labs
Abstract: This paper describes a web-based engine that exposes Scilab/Xcos simulations through a web interface as part of a virtual laboratory. The engine was designed to make simulations, exchange data with a web interface, and make Scilab usable in online educational tools. It provides a REST API for starting and managing simulations and works inside a Docker container to keep the setup stable and easy to maintain.

The system manages the entire simulation workflow, from uploading and configuring models to executing computations and visualizing results. The engine integrates Scilab into a web service layer, where simulations can be parameterized, repeated, and compared directly in the browser. This approach brings the flexibility of desktop-based modeling to an online environment and allows students to explore system dynamics interactively, without installing any additional software.

The Reaction Wheel Pendulum experiment was used to demonstrate how the engine connects Scilab simulations with the virtual laboratory, illustrating how traditional desktop-based simulation environments can be integrated into modern web-based educational infrastructures.

Keywords: Control education, Computer aided learning, Control courses and labs
Abstract: In this paper we present a new version of a smartphone app that serves as a learning tool for the analysis of linear systems modelled as transfer functions. In particular, given a system, the app provides its step response, its pole-zero map, and its Bode and Nyquist plots. The tool is interactive, so that the student can perform a ``what-if'' analysis when a parameter is changed and understand the relationships between the parameters and the time and frequency responses. Illustrative examples are given to show the main features and how it can be used in control courses.

Keywords: Control education, Energy systems
Abstract: This paper examines the integration of the gender dimension in STEM and control education within the European 2030 policy framework, considering post-COVID-19 disruptions. Using Belgium and Italy as complementary case studies, we analyse how policy objectives translate across the educational pipeline, frontier research participation, and institutional governance. Evidence from Belgium shows that while STEM policies have improved participation and gender balance, post-pandemic trends indicate a stabilisation of growth, challenging expected trajectories. Analysis of ERC grant outcomes at Ghent University reveals persistent gender disparities at the level of research excellence, highlighting a gap between participation and competitive success. The case of the Italian National Research Council (CNR) demonstrates how data-driven governance and Gender Equality Plans can operationalise structural change. Overall, the study frames gender inclusion as a system-level design objective, requiring coordinated, data-informed interventions to enhance resilience, equity, and performance in STEM and control education systems.

Keywords: Agents and autonomous systems, Machine learning, Control courses and labs
Abstract: Mini-Auto-Drive (MAD76) is a new miniature lab system in the 1:76 scale for autonomous driving in education and research. In one of its possible configurations, MAD76 is a robot car racing game where autonomous robot cars compete against each other or against human players. In this contribution, we present the design, training, deployment, and testing of reinforcement learning (RL) agents for motion planning and control in competitive robot car racing. Soft-Actor-Critic (SAC) agents are trained to achieve minimum laptimes in races with up to four robot cars. We take a partial end-to-end approach in RL, where the observation space of the SAC agents is computed from vehicle states generated by computer vision and multi-object tracking, while their actions are steering and motor commands. The observation space and hyperparameters of the SAC agents are specifically designed to compensate significant time delays in cars with front-wheel steering and radio control. By applying model-driven software engineering (MDSE) with Simulink, Reinforcement Learning Toolbox, and Embedded Coder from The Mathworks, the trained SAC agents are deployed and tested as embedded software on the Raspberry Pi 5 computer of the real MAD76 system. In competitive racing, the robot cars perform collision-free passing, distance control, braking and evasive maneuvers.

Keywords: Control education, Control courses and labs, Computer aided learning
Abstract: This paper presents a modular, web-based framework designed for the automated generation and visualization of dynamic systems composed of mass–spring–damper elements. The proposed approach integrates automated 3D model generation using the Blender Python API, numerical simulation of system dynamics in Scilab, and visualization in the browser with Babylon.js. The modular design enables flexible creation of different mechanical configurations and intuitive exploration of their behavior. Three representative models are presented: a damped harmonic oscillator, a two-body system connected by a spring and damper, and two bodies between rigid walls. The proposed framework is scalable and extensible, providing a base for future work in multi-degree-of-freedom systems, and interactive control experiments.

Keywords: Stability of nonlinear systems, Control education, Nonlinear system theory
Abstract: The describing function method is a useful tool for the qualitative analysis of limit cycles in the stability analysis of nonlinear systems. This method is inherently approximate; therefore, it should be used for a fast qualitative analysis of the considered systems. However, plotting the exact describing function requires heavy mathematical calculations, reducing interest in this method especially from the point of view of control education. The objective of this paper is to enhance the describing function method by providing a new approach for the qualitative plotting of the describing function for piecewise nonlinearities involving discontinuities. Unlike the standard method, the proposed approach allows for a straightforward, hand-drawn plotting of the describing function using the rules introduced in this paper, simply by analyzing the shape of the nonlinearity. The proposed case studies show that the limit cycles estimation performed using the standard exact plotting of the describing function yields the same qualitative results as those obtained using the proposed qualitative method for plotting the describing function.

Keywords: Electrical power systems, Adaptive control, Predictive control for linear systems
Abstract: This paper presents a data-driven framework for frequency and voltage regulation in inverter-based microgrids. Unlike conventional model-based approaches, the proposed framework does not require prior knowledge of the system model. Instead, it utilizes online recursive estimation to estimate some relevant input-to-output model directly from real-time measurements, enabling adaptive control under time-varying operating conditions. A model predictive control problem is formulated that updates continuously as new data become available. The framework assumes that all inverter-based resources are interfaced through grid-forming inverters, allowing the controller to regulate frequency and selected bus voltages by adjusting the active and reactive power setpoints of controllable units. We showcase the controller performance via numerical simulations and controller hardware-in-the-loop testing.

Keywords: Predictive control for linear systems, Electrical power systems, Optimization algorithms
Abstract: This paper presents a hierarchical model predictive control (MPC) strategy for optimal charging and discharging of battery modules in a modular multilevel converter (MMC) system. The proposed two-layer control architecture consists of an offline optimization layer and an online predictive control layer. The offline layer determines upper bounds on the number of time slots per load factor that each battery module can be assigned. The online MPC layer utilizes these bounds to regulate module currents online, ensuring correct converter functionality and compliance with the planned energy distribution. This coordinated framework enables state of charge (SoC) balancing, allowing all battery modules to reach their terminal SoC targets while minimizing degradation and improving overall charging efficiency. Simulation results illustrate the effectiveness of the proposed approach compared to conventional charging strategies.

Keywords: Predictive control for linear systems, Stochastic control, Energy systems
Abstract: This paper presents a detailed formulation of Probabilistic-Robust (Probust) MPC and its application to a real microgrid, focusing on its capability to detect faults in real-time and apply adaptive mitigation strategies. Unlike classical residual-based methods for Fault Detection and Isolation (FDI), Probust MPC integrates probabilistic constraints and robust optimization techniques to enhance resilience against uncertainties in power generation and demand. The proposed method dynamically adjusts control actions in response to stochastic disturbances, ensuring effective fault detection and mitigation. By incorporating robust probabilistic constraints, this approach systematically identifies faults before they develop into critical failures, allowing for timely corrective actions. Simulation results demonstrate that the proposed approach enhances system reliability, optimizes energy management, and improves operational robustness under various conditions. These findings highlight the potential of Probust MPC as a scalable and effective solution for operating resilient microgrids, as demonstrated through its application to data from a real energy community on the Culatra Island, Portugal.

Keywords: Power electronics, Electrical power systems, Output regulation
Abstract: This paper extends the concept of the Power Flow Controller (PFC), initially developed for DC meshed grids, to hybrid AC/DC meshed grids. The coexistence of AC and DC power flows and the grid’s complex topology present significant modeling and control challenges. To address these, a bi-linear state-space model of a modular PFC integrated into an AC/DC meshed grid is derived. By introducing an appropriate change of variables, a formal analysis of its steady-state trajectories is given, revealing degrees of freedom and formal conditions for their existence. A particular choice of steady-state is formulated, leading to a constant equilibrium, thus a bilinear time-invariant model in error coordinates. A preliminary control law is proposed that renders this equilibrium globally asymptotically stable. Simulation results illustrate the effectiveness of the proposed approach.

Keywords: Power electronics, Stability of nonlinear systems, Energy systems
Abstract: This work presents a bounded nonlinear control strategy for a bidirectional DC–DC boost converter for ensuring output voltage stability. The proposed approach ensures closed-loop stability while maintaining the control signal strictly within admissible limits, overcoming a common limitation of existing unbounded designs. By reformulating the converter dynamics and equilibrium conditions, the proposed scheme guarantees voltage regulation under both charging and discharging modes. The control law explicitly accounts for system nonlinearities and ensures asymptotic stability while allowing for a proportional and integral control action. Simulation results carried out in MATLAB/Simulink using Simscape Electrical validate the effectiveness of the method. The bidirectional converter, connected to a supercapacitor energy storage device, achieves accurate voltage and current regulation under dynamic load variations. The results confirm that the controller achieves fast transient response, robustness, and seamless mode transition between current- and voltage-regulated operation, highlighting its potential for plug-and-play integration in DC microgrid applications.

Keywords: Power electronics, Electrical power systems, H2/H-infinity methods
Abstract: This paper presents an H-infinity model matching control approach to the problem of power hardware-in-the-loop (PHIL) interfacing. The objective is to interconnect a grid simulation and a physical device via an interface in a way that is stable and accurate. Conventional approaches include the ideal transformer method (ITM) and its impedance-based variants, which trade accuracy for stability, as well as some H-infinity control-based approaches, which do not make use of all the available information in their optimization for accuracy. Designing for transparency, as opposed to accuracy as existing approaches do, would achieve both stability and accuracy, while making use of all the dynamical information present in the idealized interconnection of the grid and device. The approach proposed in this paper employs model matching to formulate the PHIL problem as an H-infinity control problem using transparency as the explicit frequency-domain control objective. The approach is experimentally validated in a real-time resistive-load PHIL setup, and is found to achieve accuracy levels that are comparable or superior to those of an ITM-based interface.

Keywords: Distributed control, Fluid flow systems, Reduced order modeling
Abstract: This paper proposes a reduced-order Chorin- projection method based on proper orthogonal decomposition (POD) for optimal control problem governed by incompress- ible magnetohydrodynamic (MHD) equations. It is a velocity- pressure reduced-order model that guarantees divergence free approximations of velocity and magnetic field. Furthermore, it does not require satisfaction of inf-sup condition for mixedPOD subspaces with the help of in-built pressure stabilization provided by the projection method without adding extra stabilization terms. We derive error estimates for the reduced-order state, adjoint and control variables. Numerical experiments are performed to discuss the accuracy and performance of the new reduced order model by solving a velocity and magnetic field tracking problem in a closed cavity.

Keywords: Distributed parameter systems, Computational methods
Abstract: This paper provides a detailed implementation of a computational scheme that uses a moving sensor to reconstruct the state of a 3D advection-diffusion partial differential equation. To estimate a PDE state over extremely large spatial domains in real-time, a CFD approach combining both physical and computational domain decomposition is used. However, this method can only rely on estimators based on Luenberger observer design, which do not account for sensor noise. A specialized version of the Kalman filter, called the restricted Kalman filter for PDEs, is proposed to include sensor noise. It is shown that under certain conditions, this filter matches a Luenberger observer, with the scalar gain in the output injection term bounded based on the sensor threshold and noise statistics. Extensive simulation studies demonstrating the effects of the scalar gain are included.

Keywords: Stochastic systems, Computational methods, Markov processes
Abstract: We propose an end-to-end algorithm for simulating the noisy dynamics of nonlinear systems satisfying divergence-free conditions. Our main technical tool is the Kolmogorov partial differential equation describing time evolution of scalar functions of solutions, averaged over noise. To enable efficient simulation, we project the Kolmogorov equation onto the space of low degree polynomials and derive a rigorous upper bound on the resulting approximation error, which may be of independent interest. Finally, we demonstrate the efficacy of our algorithm by simulating it numerically for a paradigmatic nonlinear system, dynamics of which is described by the 2D Navier-Stokes equation.

Keywords: Distributed parameter systems, Delay systems, Output feedback
Abstract: This paper addresses the design of a control strategy for an input-delayed underactuated heat-heat cascade system. The scalar delayed control is applied to the boundary of one single heat equation, which is coupled to a second heat equation through a boundary condition. The reported control strategy takes the form of an output feedback using a measurement done on the latter heat equation. This work is the first to address an underactuated heat-heat cascade with an arbitrarily large input delay. For this aim, a spectral reduction performed at the cascade level and a predictor-based output feedback (leveraging the Artstein transformation), providing exponential stabilization with a user-prescribed decay rate.

Keywords: Distributed parameter systems, Agents and autonomous systems, Sliding mode control
Abstract: This paper addresses the leader–follower consensus problem for a network of dynamical agents modeled by a class of boundary-controlled wave PDEs affected by boundary perturbations. An agent acts as the emph{leader}, while the remaining ones, referred to as emph{followers}, are required to asymptotically track the leader state. When linear local interaction protocols are employed, the possibly time-varying leader input signal must be available to all followers, which represents a restrictive assumption. By introducing a decentralized discontinuous local interaction protocol inspired by the emph{Twisting} second-order sliding mode control algorithm, suitably adapted to the distributed-parameter multi-agent systems setting, the leader input signal no longer needs to be communicated to the followers. In addition, uniformly bounded boundary disturbances are fully rejected by the proposed discontinuous local interaction rule. Consensus is proven to be achieved in the L2(0,1) space and with an exponential convergence rate through a Lyapunov-based analysis by means of which constructive tuning rules for the control parameters are derived. Simulation results are presented and discussed to corroborate the theoretical findings.

Keywords: Distributed parameter systems, Energy systems, Observers for linear systems
Abstract: Precise estimation of the electrochemical states of a Li-ion battery benefits the design of a more efficient and reliable battery monitoring system. This paper explores state estimation approaches for a Single Particle Model with electrolyte (SPMe) dynamics with modified boundary conditions integrated into a Thermal model (SPMe-T). The integration of thermal dynamics into the SPMe enhances its capability to capture temperature-dependent behavior, while maintaining a balanced trade-off between computational efficiency. An internal average temperature model is incorporated into the battery model to account for temperature variations that affect the diffusion parameter. Battery parameters, such as the diffusion coefficient, vary with temperature. Incorporating a thermal model can improve subsequent state estimation. This study estimates the lithium concentration in both the electrolyte and solid phases by proposing closed-loop backstepping Partial Differential Equation (PDE) observers. The stability of all proposed observers and the convergence of the state estimation errors are proven.

Keywords: Nonlinear system identification, Modeling, Computational methods
Abstract: We discuss the use of symmetries for analyzing the structural identifiability and observability of control systems. Special emphasis is put on the role of discrete symmetries, in contrast to the more commonly studied continuous or Lie symmetries. We argue that discrete symmetries are the origin of parameters which are structurally locally identifiable, but not globally. We exploit this fact to present a methodology for structural identifiability analysis that detects such parameters and characterizes the symmetries in which they are involved. We demonstrate the use of our methodology by applying it to four case studies.

Keywords: Identification, Model validation, Biological systems
Abstract: Dynamic models of microalgal growth often oversimplify photoacclimation and pigment dynamics under varying light, complicating parameter identification due to the strong influence of preacclimation history. To address this, we present a dynamic model that explicitly incorporates photoinhibition and photoacclimation, paired with a novel two-step calibration method. After analyzing the model’s key dynamic properties, our approach employs a differential evolution algorithm to fit model parameters to both pigment content and growth rate data from the literature. To quantify uncertainty, the Jackknife method is used to estimate confidence regions for both the parameters and model outputs. Statistical results demonstrate high calibration performance, with a Root Mean Squared Error (RMSE) of 10⁻² µg_Chl.10⁻⁶cells and 1 g_O2.g_Chl^-1.h⁻¹ (for chlorophyll and growth rate, respectively) of the same order of magnitude as the experimental Standard Deviation (SD). Experimental validation along with a high coefficient of determination (R² = 0.961) confirm the framework’s robustness and accuracy, providing a reliable tool for dynamic parameter identification in photoacclimation studies.

Keywords: Identification, Computational methods, Optimization
Abstract: This work focuses on a mathematical model that describes the coupled dynamics of microbial contamination and organic matter, in both water and fresh produce, during industrial washing processes, with and without disinfection.

To enable reliable application in industrial environments, we address the challenge of estimating process unknown parameters. A structural identifiability analysis is conducted to determine the types of measurements required to uniquely estimate these parameters. Moreover, we apply Optimal Experimental Design (OED) techniques to compute the most informative experimental configurations, ensuring efficient and targeted data collection under typical industrial constraints.

Industrial relevance: The proposed virtual representation of the process offers the food industry a rigorous and adaptable tool for microbial risk assessment, process characterization, and optimization of fresh produce washing operations.

Keywords: Identification, Modeling, Biological systems
Abstract: Minimising the risks associated with antimicrobial resistance (AMR) requires a comprehensive understanding of its emergence and selection mechanisms across experimental, industrial, environmental, or clinical settings. Mathematical models describing microbial pharmacodynamics (PD) under drug exposure are widely used for this purpose. However, PD models may be calibrated against experimental data without any guarantee of structural or practical identifiability, leading to unreliable PD parameter estimates or even to erroneous conclusions drawn from the model.

To advance towards model-based strategies for understanding and controlling AMR, this work analyses the identifiability of a general PD model to determine optimal experimental designs for model calibration using standard laboratory protocols, specifically dose-response curves. The PD model considered here combines different models from the literature to describe key interactions between susceptible and resistant subpopulations under drug effect. The proposed framework is flexible enough to account for different AMR mechanisms and drug modes of action, thereby contributing to the standardisation of PD modelling in AMR. This has significant implications for optimising antimicrobial protocols in industrial applications or clinical settings, where informed therapeutic decisions are critical.

Keywords: Biological systems, Nonlinear system identification, Iterative learning control
Abstract: Accurate temperature prediction is crucial for optimizing microalgae growth in outdoor cultivation systems. This paper presents the application of the Simplified Auto Tuning Heat Exchange (SATHE) model to predict temperature dynamics in biofilm reactors under greenhouse conditions. The SATHE model, derived from first principles heat transfer equations, offers a simplified structure with only six parameters requiring identification. We demonstrate the model's identifiability and present an efficient parameter identification strategy. By coupling temperature predictions with a growth model, we estimate productivity of a rotating algal biofilm reactor. Validation using over nine months of data shows the model's capability to accurately predict temperature trends and support operational decisions for enhanced productivity.

Keywords: Nonlinear system identification, Biological systems, Computational methods
Abstract: Optimal Experiment Design (OED) is an essential tool for selecting appropriate operating and sampling conditions for experimental runs that enable data collection for parameter estimation and model calibration. Based on an original Overall Mean Squared Error (OMSE) criterion recently proposed by the authors, which accounts for prior parameter uncertainty and measurement noise, this paper presents a MATLAB toolbox, OEDLab, that implements the method and its application to the identification of a HEK293 cell growth model. This case study highlights the selection of fed-batch input profiles and sampling schedules that minimize expected posterior uncertainty. Using Monte Carlo simulations with synthetic datasets generated with the OEDLab-optimal design, empirical parameter standard deviations align with root-mean-squared-error forecasts, and observed objective values mirror the predicted criteria.

Keywords: Stochastic control, Optimal control, Computational methods
Abstract: We present a novel infinite-dimensional linear programming (IDLP) based method for solving Partially Observable Markov Decision Processes (POMDPs) via Moment-SoS hierarchies. Standard belief-state IDLP formulations yield rational dynamics via the Bayesian filter, obstructing direct use of Moment-SoS relaxations. To address this, we reformulate the POMDP over unnormalized beliefs, yielding polynomial system dynamics. This formulation, however, breaks contractivity of the Bellman operator, rendering conventional IDLPs ill-posed. Our key contribution is to restore contractivity by introducing an ell_1-based Lyapunov weight and formulating the control problem in a weighted Banach space. The resulting Bellman operator becomes contractive, yielding a well-posed IDLP with strong duality. This leads to a generalized moment problem (GMP) formulation, solvable via Moment-SoS relaxations. We validate the method on the two-armed bandit and tiger problems, generating approximations for optimal policies and optimal values.

Keywords: Stochastic control, Optimal control, Predictive control for linear systems
Abstract: Optimal transport with quadratic cost provides a geometric framework for steering an ensemble, modeled by a probability law, with minimal effort. Yet ambient-space formulations become unwieldy in high dimensions, and sensing or actuation in practice often reveals only linear views of the state --- camera silhouettes, LiDAR beams, tomographic slices. We develop a sliced feedback controller for distribution steering: the evolving law is projected onto one-dimensional directions on the sphere, the optimal one-dimensional velocity is synthesized in each projection, and these velocities are averaged to produce a feedback control in the ambient space. The construction reduces to the Benamou--Brenier problem in one dimension. In addition, it is invariant under orthogonal transforms, nonexpansive under projections, and well posed on mathcal{P}_2(mathbb{R}^n). Computation proceeds by sampling directions on the sphere and solving independent one-dimensional subproblems, yielding a scalable method aligned with partial observations. In the Gaussian setting, we show that the developed sliced controller steers the law to the prescribed target. Furthermore, we derive an identity relating the energy consumption incurred by the controller to the sliced Wasserstein distance.

Keywords: Stochastic control, Stochastic systems, Game theoretical methods
Abstract: In this paper, we consider decentralized discrete-time stochastic dynamical optimal control problems with multiple control strategies operating under delayed-sharing information patterns, formulated within the framework of person-by-person (PbP) optimality. We invoke Girsanov’s theorem to characterize PbP optimality under a reference probability measure through value functions satisfying simplified dynamic programming (DP) equations, together with corresponding information states that serve as sufficient statistics for the strategies. The value functions and information states retain the fundamental properties of classical partially observable Markov decision problems (POMDPs), namely, both depend on the actions of the minimizing controls, rather than their strategies. The main distinguishing feature of our DP approach is that each control strategy estimates the unobservable state process and the private information components of all other strategies solely from its own private information and the delayed-sharing information components, using information states.

Keywords: Stochastic control, Optimal control, Stochastic systems
Abstract: We present a geometric coordinate-free solution to the isotropic Schrödinger bridge problem (SBP) for the kinematic equation on the Lie group SO(2). We consider the angular velocity of the system as the control input and assume that the given initial and terminal state probability density functions defined on SO(2) in our SBP are continuous and strictly positive. We solve the SBP by proving the existence and uniqueness of a solution to the so-called Schrödinger system of equations on SO(2), by showing that a fixed-point recursion is contractive in a complete metric space with respect to the Hilbert's projective metric. The geometric controller thus designed only uses the intrinsic geometric structure of SO(2) and does not embed it in the Euclidean plane to achieve the optimal density control. The numerical simulation verifies the validity of the theoretical construction of the Schrödinger bridge. The codes and animations are publicly available at https://gitlab.com/a5akhtar/sbp.

Keywords: Stochastic filtering, Variational methods
Abstract: Approximating the evolution of probability measures for nonlinear stochastic differential equations (SDEs) and the associated nonlinear filtering problems is a challenging problem as it involves solving high-dimensional differential equations. In contrast to classical variational inference methods which address this challenge by minimizing the Kullback--Leibler (KL) divergence between the true and approximate distributions, we propose a Wasserstein-based variational framework for approximating the laws of stochastic systems. In particular, instead of minimizing the KL divergence, our approach minimizes the Wasserstein-2 (W_2) distance between joint probability distributions of the state and observation processes. This formulation respects the underlying transport geometry and results in evolution equations for Gaussian parameters that provide an approximation of the dynamics of the true measure. An illustration is provided for some of our results with the help of an academic example.

Keywords: Stochastic filtering, Computational methods, Emerging control applications
Abstract: We investigate a state estimation problem for a linear time-invariant Gaussian system, based on a measured scalar privacy-sensitive signal. This is formulated as a Kalman filtering problem under an “event-level” differential privacy constraint, for which existing approaches, directly perturbing the measured signal, tend to degrade the estimation accuracy significantly. Here, we propose a two-stage differentially private architecture combining a finite impulse response (FIR) pre-filter with the Kalman filter. We cast the joint design of the FIR and Kalman filters as a (bilinear) optimization problem to improve the trade-off between privacy and estimation accuracy. Simulations on an epidemiological example demonstrate that the proposed method preserves estimation accuracy while ensuring event-level privacy.

Keywords: Transportation systems, Optimization
Abstract: Last-mile logistics is the final stage of the supply chain, where goods are delivered to their final customers. This is also the stage of the delivery chain characterized by the greatest inefficiency. This work focuses on the improvement of logistics operations in smart cities, highlighting the benefits of a cooperation scheme for a coalition of couriers to improve delivery activities in terms of cost minimization as well as sustainability and customer satisfaction maximization. In order to compare the case without cooperation among couriers with the collaborative case, multi-objective optimization formula tions are proposed to find the optimal assignment of vehicles to logistics services to be provided in the two cases. The final goal of this work is to highlight the benefits of the collaboration among couriers to favor the use of green solutions, such as electric vehicles and cargo bikes, for the deliveries in the city center. Some results obtained by applying the proposed schemes to a real case are discussed in the paper.

Keywords: Transportation systems, Optimization, Machine learning
Abstract: The rapid expansion of e-commerce has led to a sharp rise in urban parcel deliveries, placing growing pressure on city logistics systems. As freight activity intensifies in dense urban areas, the traditional model of door-to-door delivery becomes increasingly untenable due to failed delivery attempts, limited curb availability and infrastructural constraints typical of dense urban environments. In response, a broad spectrum of alternative strategies has emerged, with particular focus on last-mile optimization. Among these, microhubs, small decentralized logistics facilities located near the final delivery destination, have gained significant attention as a promising solution. This work proposes a data-driven methodology for identifying optimal locations for urban microhubs, based solely on GNSS data collected from commercial delivery vehicles. Building on the spatial distribution and duration of observed delivery stops, the approach combines clustering techniques with constrained optimization to minimize the distance between customers and microhubs. The underlying demand distribution is inferred from stop duration and spatial density, and is used to prioritize areas with higher delivery intensity in the placement process.

The resulting microhub configurations are evaluated through simulation and show the potential to reduce pollutant emissions, travel distances and delivery times compared to the traditional door-to-door delivery model. The method offers a practical and effective solution to the challenges of last-mile delivery, and is both scalable and transferable to other urban contexts.

Keywords: Traffic control, Transportation systems, Machine learning
Abstract: The trend of digitalization and smart cities is enabling new Traffic Signal Control methods, with urban transportation Digital Twins acting as a key asset, enabling data-based approaches for reducing urban traffic congestion. We propose a Digital-Twin-enabled Deep-Q-Network Traffic Signal Control scheme, trained to improve the signal control efficiency of a single intersection in spite of possible sensor failures. The controller is trained using a Digital Twin of the western half of Södermalm, Stockholm, with rewards mirroring the Max-pressure signal control. We demonstrate in simulations that our proposed control significantly outperforms the Max-pressure benchmark in cases involving sensor failures, while achieving comparable performance without them.

Keywords: Traffic control, Transportation systems, Autonomous systems
Abstract: We present a framework for predicting human driving behavior in mixed traffic where connected and automated vehicles (CAVs) coexist with human-driven vehicles (HDVs), and validate it using an open-source virtual reality (VR) testbed. We estimate the time-shift parameter of Newell's car-following model for individual drivers using Bayesian linear regression and derive analytical expressions for the mean and variance of predicted trajectories. These predictions are integrated into an optimal control framework for CAV trajectory planning. To address the scarcity of mixed-traffic data, we develop a VR platform supporting realistic, multi-user driving scenarios and provide a reproducible experimental framework with a dedicated tutorial website requiring only MATLAB and Unreal Engine. Results show our approach enables efficient HDV predictions, while the VR platform offers an accessible environment for studying human behavior in mixed traffic.

Keywords: Optimal control, Transportation systems, Optimization
Abstract: This paper presents an energy-optimal scheme to compute the speed trajectory of a single train over multiple stops subject to a relaxed timetable. Motivation for such an approach is that a train operator cannot modify the timetable itself, but can compensate for its suboptimal design (from the energy consumption perspective) by allowing variation in the strict station timings. A convex optimization approach is devised to allocate time supplements among the inter-station segments to achieve energy optimality. The proposed approach is demonstrated on an example of real timetable and simulated train dynamics. Comparison with the timetable-adhering solution revealed energy savings of more than 26 % when symmetrical 15-second relaxation around the timetable is allowed, demonstrating the energy savings potential of the proposed proof-of-concept solution.

Keywords: Optimal control, Transportation systems, Optimization algorithms
Abstract: This paper investigates whether the Green Light Optimized Speed Advisory (GLOSA) framework—extensively studied for passenger cars—can also bring benefits to urban public transport, specifically trams. While the underlying principles remain the same—a traffic light controller broadcasts its signal plan, allowing approaching vehicles to adjust their speed to avoid stopping at red lights and subsequent acceleration—both quantitative and qualitative differences exist. The former include the tram’s substantial inertia, and the latter the requirement to adhere to a fixed timetable. To assess potential benefits, historical data logged by onboard units were analyzed and compared with simulated optimal trajectories that incorporated traffic light signal plans communicated via vehicle-to-everything (V2X). The results suggest that, when signal plans of upcoming intersections are known, noticeable energy savings can indeed be achieved. However, further analysis is required, as the logged operational data do not capture all information available to human drivers during actual operation—for instance, the imminent crossing of pedestrians outside designated crosswalks, which may have prompted the driver to slow down and thus resulted in trajectories appearing suboptimal.

Keywords: Stochastic control, Adaptive control, Machine learning
Abstract: A key challenge in model-based reinforcement learning (MBRL) is the exploration--exploitation tradeoff: an agent must collect informative data to improve its dynamics model while simultaneously minimizing cumulative cost. A principled approach to provably efficient exploration is the optimistic strategy, which selects policies that are optimal under the most favorable plausible model. While theoretically sound, this strategy requires solving a joint optimization over policies and model classes that is generally intractable. In this paper, we outline how insights into the connections of various paradigms within probabilistic optimal control can be brought to bear on this challenge. The key observation is that the optimistic exploration problem can be relaxed into a fully regularized probabilistic control formulation that naturally admits tractable analytical solutions. We survey its main properties, connect them to the regret minimization for efficient MBRL, and identify open questions for future investigation.

Keywords: Medical signal processing, Biomedical systems, Fault detection and identification
Abstract: This paper presents an AI-assisted emergency monitoring framework designed for smartwatch platforms, integrating multi-level activity intensity estimation with apnea-like event detection. Rather than relying on conventional categorical human activity recognition, the proposed approach reformulates activities into five ordered intensity levels, allowing continuous characterization of motion dynamics and abnormal transitions. To support this objective, a novel dataset, textit{Dyn-Intensity}, was developed to capture vigorous and emergency-like wrist movements that are rarely represented in existing benchmarks. Computationally efficient features based on Signal Vector Magnitude (SVM) and its temporal variation are employed to enable lightweight on-device inference.

In addition, infrared PPG signals are utilized to detect apnea-like events through subject-wise baseline normalization and a logistic regression classifier, improving robustness against inter-individual physiological variability. Experimental results show that the MLP-based intensity model achieves strong ordinal consistency (QWK = 97.27%), while apnea-like detection reaches an F1-score of 91.8% under subject-independent validation. These findings demonstrate the practicality of unified motion–physiological monitoring for real-time, resource-aware emergency detection on wearable devices.

Keywords: Cooperative autonomous systems, Robust adaptive control, Autonomous systems
Abstract: begin{abstract} We propose Lie Group Distributed Gyroscopic Consensus Flow (LGDGCF), a decentralized algorithm for pose synchronization in multi-agent systems on non-Abelian Lie groups (SO(3) and SE(3)). Agents broadcast their estimates; the flow combines a preconditioned dissipative term (negative gradient of a geodesic disagreement potential, weighted by graph adjacency and local covariance estimates) with a gyroscopic term constructed from Lie bracket operations on local relative log-errors, with gains tuned conservatively.

The gyroscopic component exploits non-commutativity to induce controlled circulation of inconsistencies on cycles, reducing command jitter (RMS and peak) and enhancing robustness to packet loss and delays. We prove local geodesic energy decay, provide bounds on steady-state command jitter under explicit Baker--Campbell--Hausdorff remainder and communication assumptions, and demonstrate that the flow is not a gradient flow (even with deformed metrics) via non-closed 1-form analysis on a triad example (Appendix).

Numerical experiments on SO(3) and SE(3) with realistic noise, drift, and delayed/asynchronous communication demonstrate reduction in command jitter (RMS and peak) and significantly lower phase-slip probability compared to intrinsic/extrinsic gradient baselines and simple port-Hamiltonian formulations. The approach is lightweight and fully decentralized, suitable for embedded swarm applications. end{abstract}

Keywords: Filtering, Signal processing, Neural networks
Abstract: 重力异常的实时测量与处理 数据对于精确导航具有重要意义， 地球物理勘探。传统上，高阶有限 脉冲响应（FIR）滤波器被广泛应用于 从复杂噪声中提取有效重力信号 背景。然而，这些滤镜不可避免地引入了 固定时间延迟，严重影响实时 系统性能。而神经网络则 展示了拟合复杂滤波的能力 过程并消除相位延迟，它们的实际操作 部署受到以下要求的严重限制 大量训练数据和耗时的初始化过程 过程。为应对这些挑战，本文提出了一个 基于仿真的实时神经过滤框架 预训练和现实世界的微调策略。最初， 基于物理的大规模模拟数据集 重力异常和传感器噪声的特征为 构建

Keywords: Biological systems, Discrete event systems, Modeling
Abstract: Although excitable neuromorphic circuits are physical models which produce continuous-time trajectories, these trajectories consist of sequences of discrete events. This mixture of the continuous and the discrete allows us to do hierarchical control, that is, reference tracking and decision-making, with the same building blocks. In order to investigate the possibility of using automated synthesis for generating neuromorphic circuits whose ordering of events is fixed, we explore the concept of forcible events to make the model hierarchical.

Keywords: Applications in neuroscience, Nonlinear system theory, Biological systems
Abstract: We study the contraction of Hodgkin-Huxley model and its role in the reliability of spike timings. Without input, the model is contractive in the region of physiological interest. With impulsive synaptic inputs, contraction is retained provided that the input events are sparse enough. Contraction is lost when the input firing rate is too high. Spike timings are shown to be reliable in the contracting regime.

Keywords: Stability of hybrid systems, Lyapunov methods, Applications in neuroscience
Abstract: Event-based control, unlike analogue control, poses significant analytical challenges due to its hybrid dynamics. This work investigates the stability and inter-event time properties of a control-affine system under event-based impulsive control. The controller consists of multiple neuronal units with leaky integrate-and-fire dynamics acting on a time-invariant, multivariable plant in closed loop. Both the plant state and the neuronal units exhibit discontinuities that cancel if combined linearly, enabling a direct correspondence between the event-based impulsive controller and a corresponding analogue controller. Leveraging this observation, we prove global practical stability of the event-based impulsive control system. In the general nonlinear case, we show that the event-based impulsive controller ensures global practical asymptotic stability if the analogue system is input-to-state stable (ISS) with respect to specific disturbances. In the linear case, we further show global practical exponential stability if the analogue system is stable. We illustrate our results with numerical simulations. The findings reveal a fundamental link between analogue and event-based impulsive control, providing new insights for the design of neuromorphic controllers.

Keywords: Applications in neuroscience, Emerging control applications
Abstract: The describing function (DF) and phase response curve (PRC) are classical tools for the analysis of feedback oscillations and rhythmic behaviors, widely used across engineering, biology, and neuroscience. However, these methods are not directly applicable to excitable systems, as they violate the underlying assumptions of harmonic balance and periodicity. Therefore, the paper proposes a novel approach to extending these classical methods tailored to excitable systems. Our methods rely on an abstraction to a discrete-event map from input events to output events. The methodology is illustrated on the excitability model of Hodgkin--Huxley. The proposed framework provides a basis for designing and analyzing central pattern generators in networks of excitable neurons, with direct relevance to neuromorphic control.

Keywords: Applications in neuroscience, Cellular dynamics, Nonlinear system theory
Abstract: Many essential biological functions, such as breathing and locomotion, rely on the coordination of robust and adaptable rhythmic patterns, governed by specific network architectures known as connectomes. Rhythmic adaptation is often linked to slow structural modifications of the connectome through synaptic plasticity, but such mechanisms are too slow to support rapid, localized rhythmic transitions. Here, we propose a neuromodulation-based control architecture for dynamically reconfiguring rhythmic activity in networks with fixed connectivity. The key control challenge is to achieve reliable rhythm switching despite neuronal degeneracy, a form of structured variability where widely different parameter combinations produce similar functional output. Using equivariant bifurcation theory, we derive necessary symmetry conditions on the neuromodulatory projection topology for the existence of target gaits. We then show that an adaptive neuromodulation controller, operating in a low-dimensional feedback gain space, robustly enforces gait transitions in conductance-based neuron models despite large parametric variability. The framework is validated in simulation on a quadrupedal gait control problem, demonstrating reliable gallop-to-trot transitions across 200 degenerate networks with up to fivefold conductance variability.

Keywords: Hybrid systems, Lyapunov methods, Emerging control theory
Abstract: In this paper, we provide systematic design methods for the stabilization of equilibria of linear time-invariant (LTI) systems with the restriction that the input signal consists of fixed-amplitude spikes. The motivation for studying spiking control comes from the field of neuromorphic engineering. This field aims to recreate, in engineered systems, the benefits of energy efficiency, low latency, and adaptability witnessed in biological neurons by control and communication through short asynchronous pulses (spikes). The spiking nature of the input signals poses challenges to the controller design, which requires event-based and fixed-amplitude control design methods. In this paper, we pursue a Lyapunov-based greedy method in which a spike, modeled as a Dirac pulse, is fired when it strictly reduces a weighted stabilization error in the form of a quadratic Lyapunov function, compared to the non-spiking case. We propose schemes with and without time regularization, proving in both cases the absence of Zenoness such that an infinite number of spikes in a finite-length time window does not occur. Moreover, we establish a practical stability property for the closed-loop system, being global for the solution without time regularization and semi-global for the time-regularized approach. The proposed approaches are illustrated through numerical case studies.

Keywords: Optimal control, Applications in neuroscience, Autonomous systems
Abstract: We study minimum-energy state transfer near linearized dynamical saddle points in one and two dimensions using Pontryagin's Maximum Principle (PMP). We show that the optimal controls can be expressed not only as explicit functions of time, but also as nonlinear, task-conditioned state-feedback primitives. This suggests a neural interpretation in which individual primitives are realized by neurons, while opposing primitives are implemented by a gated agonist-antagonist neural circuit. In the free-horizon across-saddle setting, these primitives reduce to rectified linear units (ReLUs). Using a simple pendulum, we further show that a local saddle-based controller approximates the nonlinear optimum near the upright equilibrium. These results identify a class of local control problems with linearized dynamics for which optimal feedback naturally acquires neuron-like nonlinear structure.

Keywords: Predictive control for nonlinear systems, Machine learning, Mechatronics
Abstract: Learning-based model predictive control has emerged as a powerful approach for handling complex dynamics in mechatronic systems, enabling data-driven performance improvements while respecting safety constraints. However, when computational resources are severely limited, as in battery-powered tools with embedded processors, existing approaches struggle to meet real-time requirements. In this paper, we address the problem of real-time torque control for impact wrenches, where high-frequency control updates are necessary to accurately track the fast transients occurring during periodic impact events, while maintaining high-performance safety-critical control that mitigates harmful vibrations and component wear. The key novelty of the approach is that we combine data-driven model augmentation through Gaussian process regression with neural network approximation of the resulting control policy. This insight allows us to deploy predictive control on resource-constrained embedded platforms while maintaining microsecond-level inference times. The proposed framework is evaluated through numerical simulations and hardware experiments on a custom impact wrench testbed. The results show that our approach successfully achieves real-time control suitable for high-frequency operation while maintaining probabilistic constraint satisfaction and improving tracking accuracy compared to baseline speed control.

Keywords: Predictive control for nonlinear systems, Machine learning, Adaptive control
Abstract: Closed-loop performance of sequential decision making algorithms, such as model predictive control, depends strongly on the choice of controller parameters. Bayesian optimization allows learning of parameters from closed-loop experiments, but standard Bayesian optimization treats this as a black-box problem and ignores the temporal structure of closed-loop trajectories, leading to slow convergence and inefficient use of experimental resources. We propose a time-series-informed multi-fidelity Bayesian optimization framework that aligns the fidelity dimension with closed-loop time, enabling intermediate performance evaluations within a closed-loop experiment to be incorporated as lower-fidelity observations. Additionally, we derive probabilistic early stopping criteria to terminate unpromising closed-loop experiments based on the surrogate model’s posterior belief, avoiding full episodes for poor parameterizations and thereby reducing resource usage. Simulation results on a nonlinear control benchmark demonstrate that, compared to standard black-box Bayesian optimization approaches, the proposed method achieves comparable closed-loop performance with roughly half the experimental resources, and yields better final performance when using the same resource budget, highlighting the value of exploiting temporal structure for sample-efficient closed-loop controller tuning.

Keywords: Biological systems, Predictive control for nonlinear systems, Identification for control
Abstract: Agricultural production is under increasing pressure from population growth and climate change, making efficient and high-yield farming essential. Greenhouse cultivation offers a controlled environment to meet these demands. This work investigates Data-Enabled Predictive Control (DeePC), a model-free method, for optimizing lettuce yield in a simulated greenhouse. We benchmark its performance against a nonlinear Model Predictive Control (NMPC) approach, which relies on a detailed physics-based model. Both controllers were tasked with maximizing final yield over a 40-day growth cycle under winter and summer weather conditions. In the winter scenario, which matched its training data, DeePC achieved 94% of NMPC's final biomass (125.81 g/m^2 vs. 133.30 g/m^2) and a slightly lower economic return (2.164 vs. 2.516 Hfl/m^2). In addition, DeePC was five times more computationally efficient. However, when the winter-trained DeePC was applied to the summer scenario, it frequently violated critical temperature constraints, despite achieving a high yield. These findings demonstrate that DeePC is a promising alternative for economic greenhouse control, but its reliability and safety depend heavily on the representativeness of its training data.

Keywords: Maritime, Optimal control, Machine learning
Abstract: Autonomous marine vessels require robust control systems capable of handling model uncertainties and environmental disturbances during precision manoeuvrers. This work presents three main contributions to address these challenges: First, a novel overactuated Reinforcement Learning-based Nonlinear Model Predictive Control (RL-NMPC) on the 4-azimuth thruster milliAmpere 1 research ferry with control allocation directly integrated in the optimal control problem. A combination of Q-learning and model-learning is employed to simultaneously improve the control goal of the NMPC and adapt the model online. The integrated allocation in the optimal control problem avoids allocator-controller conflict and improves robustness under wind disturbances and parameter drift. Second, a hybrid reference filter that solves coordinated surge, sway and yaw trajectory profiling without weight distortion between Degrees-of-Freedom (DoF). The filter combines trapezoidal velocity generation with body-frame velocity saturation and third-order filtering to provide smooth, predictable trajectory references for multi-DoF manoeuvres. Third, a comparative study of RL-NMPC, NMPC, and dynamic positioning control systems under 40% modelling errors and 4.0 m/s wind disturbances on the established 8-Corner Test trajectory. Simulation results demonstrate that RL-NMPC achieves superior tracking performance, reduced energy consumption, and lower actuator wear compared to both NMPC and DP controllers, while maintaining robust control under environmental disturbances.

Keywords: Robotics, Mechatronics, Predictive control for nonlinear systems
Abstract: We presents an auto-tuning framework for torque-based Nonlinear Model Predictive Control (nMPC), where the MPC serves as a real-time reactive controller for optimal joint torque commands. The MPC cost function weights and the low-level controller gains, are optimized using high-dimensional Bayesian Optimization (BO) techniques, specifically Sparse Axis-Aligned Subspace (SAASBO) with a digital twin (DT) to achieve end-effector trajectory real-time tracking on an UR10e robot arm. The simulation model allows efficient exploration of the high-dimensional parameter space, and it ensures safe transfer to hardware. Our simulation results demonstrate significant improvements in tracking performance (+41.9%) and reduction in solve times (-2.5%) compared to manually- tuned parameters. Moreover, experimental validation on the real robot follows this trend (with a +25.8% improvement), emphasizing the applicability of digital twins in optimization.

Keywords: Automotive, Safety critical systems, Robust control
Abstract: Safety-critical motion planning in mixed traffic remains challenging for autonomous vehicles, especially when interactions between the ego vehicle~(EV) and surrounding vehicles~(SVs) are significant. In dense traffic, the feasibility of a lane change depends strongly on how SVs respond to the EV motion. This paper presents an interaction-aware safety framework that incorporates EV--SV coupling into a control barrier function~(CBF)-based safety assessment. The proposed method learns a joint action-conditioned interaction model from data that captures reactive SV behavior and embeds it into a predictive environmental CBF~(ECBF) evaluated over a finite horizon. To address residual mismatch between the learned model and true SV behavior, a continuous observer is designed that estimates and compensates SV acceleration mismatch online. Compared with classical ECBF methods that neglect SV reactions, the proposed approach provides a less conservative and more informative safety representation for interactive traffic scenarios.

Keywords: Optimization, Optimization algorithms, Robotics
Abstract: — We present a GPU-accelerated trajectory optimization framework for scalable, real-time multi-agent motion planning in both 2D and 3D environments. The proposed method refines nominal trajectories using a model predictive control (MPC) approach, incorporating collision avoidance with obstacles and neighboring agents. By formulating the problem to exploit parallel computation, the optimizer efficiently handles tens of agents simultaneously. The algorithm supports dynamic environments by allowing local updates in response to unforeseen perturbations. We validate our method through a diverse set of simulated scenarios, demonstrating smooth and feasible trajectories with low computational overhead. Comparative analysis in 2D scenarios against the RVO2 baseline highlights improvements in smoothness and path quality

Keywords: Agents and autonomous systems, Decentralized control, Adaptive systems
Abstract: This paper addresses the problem of multi-agent coordination and obstacle avoidance for systems with limited sensing. Inspired by tensegrity structures, a nonlinear, distance-based control law composed of networked virtual tensile and compressive members is proposed. An adaptive gain relaxation on tensile forces is added, allowing formations to smoothly deform around convex obstacles before reforming back to the designed configuration. The obstacle avoidance protocol only relies on local spatial measurements, making it suitable for agents with limited sensing. Simulations demonstrate that a simple formation utilizing tensegrity-based control successfully navigates obstacles and is robust to noise and dynamic disturbances.

Keywords: Cooperative control, Optimization, Decentralized control
Abstract: Unmanned aerial vehicles (UAVs) are transforming urban mobility, logistics, and traffic monitoring. Among their diverse applications, parcel delivery and road monitoring have emerged as particularly influential. While delivery UAVs aim to minimize response time and optimize routing, monitoring UAVs seek to maximize spatiotemporal coverage of transportation networks. This paper presents a unified decision-making framework that enables UAVs to jointly perform delivery and monitoring tasks in a coordinated manner. The urban environment is represented as a graph-based road network where each segment’s importance value increases over time and is refreshed upon UAV visits. To address this problem, we design a knowledge-aware rolling horizon planning framework that operates in a fully distributed manner, allowing each UAV to re-plan its trajectory iteratively based on local state and shared global knowledge. This design ensures scalability and practical deployability. Simulations on real-world road networks show that our method achieves a superior balance between delivery efficiency and monitoring coverage compared to existing baselines.

Keywords: Optimal control, Decentralized control, Concensus control and estimation
Abstract: This paper presents a novel hybrid motion planning method for holonomic multi-agent systems. The proposed decentralised model predictive control (MPC) framework tackles the intractability of classical centralised MPC for a growing number of agents while providing safety guarantees. This is achieved by combining a fully decentralised version of the alternating direction method of multipliers (ADMM) with a centralised high-order control barrier function (HOCBF) architecture. Simulation results show significant improvement in scalability over classical centralised MPC. We validate the efficacy and real-time capability of the proposed method by developing a highly efficient C++ implementation and deploying the resulting trajectories on a real industrial magnetic levitation platform.

Keywords: Complex systems, Distributed cooperative control over networks, Agents and autonomous systems
Abstract: In this work, we consider multi-agent systems that are subject to a set of potentially coupled tasks expressed in Signal Temporal Logic (STL). The agents are equipped with a trust evaluation mechanism that allows them to determine the trustworthiness of the other agents in a quantitative manner. Given the local trust levels of the agents, we first introduce a novel trust-based STL robustness metric that propagates the agents' trust levels in the operator level. As a second contribution, a distributed control framework is proposed that is comprised by a slow and a fast time scale control algorithm. In the fast time-scale agents determine their local trust levels in a distributed manner. In the slow time scale the agents decide their actions aiming towards maximizing the newly proposed robustness metric using only information received from their immediate neighbors. The effectiveness of the proposed scheme is demonstrated in a time-varying formation control scenario.

Keywords: Cooperative control, Agents and autonomous systems, Decentralized control
Abstract: This work presents a cooperative trajectory planning framework for multiple agents, enabling safe traversal through a narrow gap, which permits a limited number of simultaneous traversals. The gap is partitioned into multiple feasible, traversable slots, and an auction-based allocation assigns a slot to each agent. A trajectory generation method guides each agent to its assigned traversable slot while ensuring alignment with a specified normal to the gap. Additionally, a cooperative speed-adjustment rule is introduced to ensure collision avoidance among agents. The key highlight of this work is a cooperative trajectory design, enabling safe and efficient multi-agent gap traversal. Numerical simulations demonstrate the efficacy of the proposed framework.

Keywords: Game theoretical methods, Optimization, Randomized algorithms
Abstract: Resource allocation problems across multiple contests are ubiquitous in adversarial settings, from military operations to market competition. While Colonel Blotto and General Lotto games have provided valuable theoretical foundations for such problems, their equilibrium characterizations typically permit resources to be arbitrarily allocated across all contests -- a flexibility that rarely aligns with practical constraints. This paper introduces a novel constrained variant of the General Lotto game where one player is restricted to allocating resources to only a single contest. In this model we provide lower and upper bounds on the security values for this constrained player, quantifying how the inability to distribute resources across multiple contests fundamentally changes optimal strategic behavior and performance guarantees. These findings contribute to a broader understanding of how operational constraints shape strategic outcomes in competitive resource allocation, with implications for decision-makers facing similar constraints in practice.

Keywords: Game theoretical methods, Electrical power systems, Agents networks
Abstract: As electricity consumption grows, reducing peak demand—the maximum load on the grid—has become critical for preventing infrastructure strain and blackouts. Pricing mechanisms that incentivize consumers with flexible loads to shift consumption away from high-demand periods have emerged as effective tools, yet different mechanisms are used in practice with unclear relative performance. This work compares two widely implemented approaches: emph{anytime peak pricing} (AP), where consumers pay for their individual maximum consumption, and emph{coincident peak pricing} (CP), where consumers pay for their consumption during the system-wide peak period. To compare these mechanisms, we model the electricity market as a strategic game and characterize the peak demand in equilibrium under both AP and CP. Our main result demonstrates that with perfect information, equilibrium peak demand under CP never exceeds that under AP; on the other hand, with imperfect information, the coordination introduced by CP can backfire and induce larger equilibrium peaks than AP. These findings demonstrate that potential gains from coupling users' costs (as done in CP) must be weighed against these miscoordination risks. We conclude with preliminary results indicating that progressive demand cost structures—rather than per-unit charges—may mitigate these risks while preserving coordination benefits, achieving desirable performance in both deterministic and stochastic settings.

Keywords: Game theoretical methods, Stability of nonlinear systems
Abstract: We consider a class of continuous-time dynamic games involving a large number of players. Each player selects actions from a finite set and evolves through a finite set of states. State transitions occur stochastically and depend on the player's chosen action. A player's single-stage reward depends on their state, action, and the population-wide distribution of states and actions, capturing aggregate effects such as congestion in traffic networks. Each player seeks to maximize a discounted infinite-horizon reward. Existing evolutionary game-theoretic approaches introduce a model for the way individual players update their decisions in static environments without individual state dynamics. In contrast, this work develops an evolutionary framework for dynamic games with explicit state evolution, which is necessary to model many applications. We introduce a mean field approximation of the finite-population game and establish approximation guarantees. Since state-of-the-art solution concepts for dynamic games lack an evolutionary interpretation, we propose a new concept - the Mixed Stationary Nash Equilibrium (MSNE) - which admits one. We characterize an equivalence between MSNE and the rest points of the proposed mean field evolutionary model and we give conditions for the evolutionary stability of MSNE.

Keywords: Game theoretical methods, Network analysis and control, Statistical learning
Abstract: This article describes a strong yet unexpected parallel between Shapley Values Additive Explanations (SHAP), a prevalent feature attribution method in eXplainable Artificial Intelligence (XAI), and causal inference in networks of dynamic systems: the computation of Shapley values can be interpreted as the average over multiple network identification procedures when no a priori assumptions are made about the causal structure underlying the input features and the model output. In contrast, causal inference algorithms for network reconstruction leverage a priori structural knowledge (such as faithfulness or partial causal orderings) and also apply targeted conditional independence tests to prune spurious connections. This analogy provides a novel theoretical framework for defining causally informed variations of Shapley values, where a priori knowledge and auxiliary independence tests are incorporated into the computation, enabling more causally consistent explanations. Beyond this conceptual contribution, the article introduces needed practical algorithmic refinements by modifying the kernel weighting in KernelSHAP to fit these specific variations to ensure computational consistency and scalability without resorting to exhaustive combinatorial enumeration. The resulting framework bridges causal reasoning and explainability, offering both theoretical insights and practical tools for interpreting black-box models acting on dynamic data.

Keywords: Machine learning, Game theoretical methods, Statistical learning
Abstract: We focus on stochastic generalized Nash equilibrium problems (SGNEPs) to establish probabilistic guarantees on the quality of the solution produced by a data-driven version of the extragradient (EG) algorithm for stochastic generalized Nash equilibrium (SGNE) seeking. Specifically, we consider a setting where the distribution of the random variable influencing the multi-agent process is unknown, and only a finite number of realizations is available for its characterization. While convergence to a SGNE should not be expected with finite data, we leverage a sample average-based approximation of the game mapping to derive distribution-free certificates bounding the distance between the true SGNE and the solution produced by the data-driven EG method. Consistently, the proposed bound shrinks as the number of samples grows to infinite.

Keywords: Transportation systems, Game theoretical methods, Control over networks
Abstract: Mobility systems are complex socio-technical environments influenced by multiple stakeholders with hierarchically interdependent decisions, rendering effective control and policy design inherently challenging. We bridge hierarchical game-theoretic modeling with online feedback optimization by casting urban mobility as a tri-level Stackelberg game (travelers, operators, municipality) closed in a feedback loop. The municipality iteratively updates taxes, subsidies, and operational constraints using a projected two-point (gradient-free) scheme, while lower levels respond through equilibrium computations (Frank-Wolfe for traveler equilibrium; operator best responses). This model-free pipeline enforces constraints, accommodates heterogeneous users and modes, and scales to higher-dimensional policy vectors without differentiating through equilibrium maps.

On a real multimodal network for Zurich, Switzerland, our method attains substantially better municipal objectives than Bayesian optimization and Genetic algorithms, and identifies integration incentives that increase multimodal usage while improving both operator objectives. The results show that feedback-based regulation can steer competition toward cooperative outcomes and deliver tangible welfare gains in complex, data-rich mobility ecosystems.

Keywords: Autonomous robots, Robotics, Intelligent systems
Abstract: This work proposes a COLREGs-compliant collision-avoidance (COLAV) framework for autonomous docking transits based on Control Barrier Functions (CBFs). Starting with the definition of a safety set using the target-ship domain, we introduce a higher-order CBF (HOCBF) to strictly enforce domain exclusion, together with a scenario-dependent weighting of the CBF constraint to handle thrust saturation while preserving feasibility. The CBF acts as a reactive safety filter on top of conventional guidance and control, producing minimally modified inputs. We validate the approach in the high-fidelity milliAmpere1 simulator across COLREGs Rules 13–16 and in field trials (overtaking, head-on, give-way). Our results demonstrate reliable, rule-compliant avoidance and robust handling of time-critical obstacles, while also revealing tuning interactions with the heading/docking controller.

Keywords: Autonomous robots, Observers for nonlinear systems, Quantized systems
Abstract: In this paper we study the observability properties of visual-place-recognition (VPR)-based localization systems for mobile robots. In our setup, the environment is split into discrete cells and the robot’s camera, through a VPR module, provides a discrete place label together with probabilistic distribution over cells, indicating the likelihood of being in each cell. An idealized, quantized version of this measurement is introduced to enable a rigorous observability analysis according to [1]. Using a smooth surrogate of the discontinuous measurement map, we apply the Hermann–Krener rank test to characterize weak local observability as the measurement resolution increases. The theoretical results identify conditions under which localization is possible or ambiguous given region-based sensing. A particle filter implementation using the probabilistic measurement model is employed to validate the analytical findings in simulation, confirming the predicted observability properties.

Keywords: Autonomous systems, Nonlinear system theory, Robotics
Abstract: This paper presents a framework for safe navigation of a unicycle point robot to a goal position in an environment populated with obstacles from almost any admissible state, considering input limits. We introduce a novel QP formulation to create a Cinf-smooth vector field with reduced total bending and total turning. Then we design an analytic, non-linear feedback controller that inherently satisfies the conditions of Nagumo's theorem, ensuring forward invariance of the safe set without requiring any online optimization. We have demonstrated that our controller, even under hard input limits, safely converges to the goal position. Simulations confirm the effectiveness of the proposed framework, resulting in a twice faster arrival time with over 50% lower angular control effort compared to the baseline.

Keywords: Optimization, Optimization algorithms, Autonomous systems
Abstract: We explore a GPU-accelerated framework for real-time pose graph optimization in stereo camera-based Simultaneous Localization and Mapping (SLAM). As pose graphs increase in size, due to larger sliding windows and the integration of diverse sensor modalities, real-time optimization becomes challenging for traditional CPU-based solvers. Our approach exploits GPU parallelism to efficiently handle large-scale graphs, achieving optimization of graphs with over 10,000 edges. Each iteration takes only 3 milliseconds on an NVIDIA RTX 2060. Additionally, our work includes a Lie Algebra toolkit with clear documentation to support fast prototyping of GPU-based pose graph solvers. The full implementation is open-sourced at https://github.com/HaktanM/S-PGO.

Keywords: Sensor and signal fusion, Robotics
Abstract: Visual-Inertial Odometry (VIO) is a staple for reliable state estimation on constrained and lightweight platforms due to its versatility and demonstrated performance. However, pertinent challenges regarding robust operation in dark, low-texture, obscured environments complicate the use of such methods. Alternatively, Frequency Modulated Continuous Wave (FMCW) radars, and by extension Radar-Inertial Odometry (RIO), offer robustness to these visual challenges, albeit at the cost of reduced information density and worse long-term accuracy. To address these limitations, this work combines the two in a tightly coupled manner, enabling the resulting method to operate robustly regardless of environmental conditions or trajectory dynamics. The proposed method fuses image features, radar Doppler measurements, and Inertial Measurement Unit (IMU) measurements within an Iterated Extended Kalman Filter (IEKF) in real-time, with radar range data augmenting the visual feature depth initialization. The method is evaluated through flight experiments conducted in both indoor and outdoor environments, as well as through challenges to both exteroceptive modalities (such as darkness, fog, or fast flight), thoroughly demonstrating its robustness. The implementation of the proposed method is available at: https://github.com/ntnu-arl/radvio.

Keywords: Delay systems, Stability of linear systems, Optimization
Abstract: In this article, we invite the reader to take a different viewpoint on the analysis and stabilization of time delay systems; while most existing literature point out the negative impact of delays on system stability, this article demonstrates that deliberate introduction of delays in the controller dynamics can be advantageous; we show that delays can aid in system stabilization in settings where static output feedback controllers are ineffective. Moreover, in numerous applications, including power systems and machining, an accurate system description leads to time-periodic models. Accordingly, we propose a systematic framework for the design of delay-based controllers to stabilize systems governed by time-periodic delay differential algebraic equations, where not only the feedback gains but also the intentional delays are control parameters. We formulate a non-smooth and constrained optimization problem rooted in minimizing the spectral radius of the monodromy operator of the closed-loop system. We conclude our study with an evaluation of the efficacy of our method and a comparative analysis against existing approaches, illustrated through two numerical case studies.

Keywords: Stability of linear systems, Delay systems
Abstract: In this paper, we investigate a pinned inverted pendulum system controlled by delayed full-state feedback. By using the delay as a control parameter, we propose a prescribed stabilization method based on the partial pole placement paradigm, which in turn relies on the multiplicity-induced dominancy (MID) property for linear time-invariant dynamical systems described by delay differential equations. The validity of the proposed stabilization approach is analyzed through a parameter-based study of a semi-algebraic system of equations derived from the so-called elimination-produced polynomial, along with additional polynomial inequalities. This analysis is carried out using cylindrical algebraic decomposition and computer algebra tools. Our results yield an accurate characterization of the regions in the parameter space—defined by the mechanical parameters and the feedback delay in the control loop, where feedback stabilization is achievable through the assignment of a multiple real root of the corresponding characteristic function that explicitly defines the decay rate of the solution of the corresponding closed-loop system.

Keywords: Linear time-varying systems, Stability of linear systems, Robust control
Abstract: This paper falls under the umbrella of the literature on slowly time-varying linear systems with the most salient assumption being that the state matrix A(t) is pointwise Hurwitz. In many papers, this property is referred to as stability at each frozen time. Under mild regularity conditions imposed on A(t), the central question considered in the related literature and in this paper is: How large can the rate of variation of A(t) be while providing a guarantee of global exponential stability? In this setting, our main claim is that the rate bounds obtained by previous authors can be quite conservative and often can be significantly improved upon. To this end, we assume a Lipschitz constant L to parameterize the largest rate at which A(t) can vary. Then, by treating L as a variable parameter which is carried through each step of the Lyapunov analysis, we obtain an estimate L* for the so-called tolerable rate of variation. Our main result, Theorem 1, indicates that if L < L*, global exponential stability is guaranteed. The paper also includes three examples for which this new estimate is compared with existing rate bounds and interpretation of our results is given in a robust control context.

Keywords: Optimization algorithms, Stability of linear systems, LMI's/BMI's/SOS's
Abstract: This paper proposes a new numerical solution for finite time stability (FTS) of a linear time invariant (LTI) system. For this purpose, we use piecewise polynomial functions to recast a Differential linear matrix inequality (DLMI) into a time dependant quadratic inequality. Second, we express the time dependant resultant vector of this quadratic constraint as a Linear Fractional transformation (LFT). Finally, we use a S-Procedure lemma to convert this quadratic time-dependent constraint over a finite time interval into an efficient, tractable Linear Matrix Inequality (LMI). Then, the proposed approach is used to get new computationally efficient sufficient LMIs for checking the FTS of an LTI system.

Keywords: Stability of linear systems, Switched systems, Optimization
Abstract: Employing emph{sparse inputs} is a way to use resources efficiently in networked control systems. This work considers the problem of stabilizing continuous linear time-invariant (LTI) systems using s-sparse inputs. Using switched systems theory, we show that any stabilizable continuous-time diagonal LTI system can be stabilized using a 1-sparse input. Next, we design s ge 1 sparse inputs that minimize the feedback gain, leading to the design of inputs with reduced control energy. We show that an upper bound on control energy required grows inversely proportional to s^2 while providing the same exponential rate of system stability for all values of s. We corroborate our theoretical results via numerical simulations.

Keywords: Stability of linear systems, Distributed parameter systems, Computational methods
Abstract: This work addresses the stabilization problem for a pair of linear Korteweg–de Vries (KdV) equations coupled through dispersive effects on a bounded domain. The model is motivated by a physical phenomenon in which two long internal waves, having nearly equal phase speeds but corresponding to different modes, interact. Control inputs are applied at the left (Dirichlet) boundaries, while the right boundaries remain uncontrolled. Our goal is to design a feedback control law based on the backstepping method to achieve stabilization for any domain length. Finally, numerical simulations are provided to illustrate and validate the theoretical results.

Keywords: Optimal control, Switched systems, Lyapunov methods
Abstract: Simple Clock Auctions (SCA) are a mechanism commonly used in spectrum auctions to sell lots of frequency bandwidths. We study such an auction with one player having access to perfect information against straightforward bidders. When the opponents’ valuations satisfy the ordinary substitutes condition, we show that it is optimal to bid on a fixed lot over time. In this setting, we consider a continuous-time version of the SCA auction in which the prices follow a differential inclusion with a piecewise-constant dynamics. We show that there exists a unique solution in the sense of Filippov. This guarantees that the continuous-time model coincides with the limit of the discrete-time auction when price increments tend to zero. Moreover, we show that the value function of this limit auction is piecewise linear (though possibly discontinuous). Finally, we illustrate these results by analyzing a simplified version of the multiband Australian spectrum auction of 2017.

Keywords: Switched systems, Fault detection and identification, Signal processing
Abstract: Motivated by a study on deception and counterdeception, this paper addresses the problem of identifying an agent’s target as it seeks to reach one of two targets in a given environment. In practice, an agent may initially follow a strategy to aim at one target but decide to switch to another midway. Such a strategy can be deceptive when the counterpart only has access to imperfect observations, which include heavily corrupted sensor noise and possible outliers, making it difficult to visually identify the agent’s true intent. To counter deception and identify the true target, we utilize prior knowledge of the agent’s dynamics and the imprecisely observed partial trajectory of the agent’s states to dynamically update the estimation of the posterior probability of whether a deceptive switch has taken place. However, existing methods in the literature have not achieved effective deception identification within a reasonable computation time. We propose a set of outlier-robust change detection methods to track relevant change-related statistics efficiently, enabling the detection of deceptive strategies in hidden nonlinear dynamics with reasonable computational effort. The performance of the proposed framework is examined for Weapon-Target Assignment (WTA) detection under deceptive strategies, using random simulations in the kinematics model with external forcing.

Keywords: Switched systems, Optimal control, Chaotic systems
Abstract: This paper studies the differentiability of the value function of switched linear systems under arbitrary switching and controlled switching, referred to as worst-case and optimal value functions respectively. First, we show that the value functions are Lipschitz continuous, when the cost function is Lipschitz continuous. Then, as the central contribution of this work, we show with examples that each of these functions can be non-differentiable on dense subsets of the state space, even if the cost function is smooth and Lipschitz continuous. This has implications for optimal control and reinforcement learning since it implies that the exact computation of these value functions requires templates involving functions that are non-differentiable on dense subsets.

Keywords: Hybrid systems, Stability of hybrid systems, Biomedical systems
Abstract: The problem of keeping the output of a continuous plant within a predefined corridor of values is considered for a class of positive single-input single-output Wiener systems. A pulse-modulated feedback controller is employed to stabilize the closed-loop system along a periodic solution satisfying the corridor condition. The return map describing the propagation of the continuous plant state through the impulse instants is used to represent the hybrid closed-loop dynamics. Fixed points of this map correspond to periodic solutions of the closed-loop system. The class of periodic solutions featuring a single impulse of the pulse-modulated feedback within the minimal period (a 1-cycle) is treated. The control design problem is then solved in three steps. First, for the given continuous plant dynamics, the fixed point of the return map corresponding to a 1-cycle satisfying the corridor condition is computed. Second, this fixed point is stabilized by selecting the slopes of the modulation functions at the fixed point. Third, the modulation functions are adjusted to obtain the desired period and amplitude of the periodic solution. The results are illustrated by simulation of muscle-relaxant agent administration in general anesthesia.