In this study, the observer-based asynchronous stabilization is addressed for networked system under multichannel attacks, in which the asynchronous phenomenon refers to the mismatch between the controller mode and the actual attack mode. To accurately depict complex attack behaviors, a piecewise homogeneous semi-Markov chain (SMC) model modulated by a superstratum Markov chain is introduced, which can simultaneously describe the randomness of attack mode transitions and the time-varying nature of transition probabilities. Considering that the actual attack modes are inaccessible, an observer-based mode switching delay technique is designed to solve this challenge. Under the framework of a piecewise homogeneous SMC, a sufficient criterion is established to ensure the $\varsigma $ -error mean-square stability under random multichannel denial-of-service attacks by means of a Lyapunov function depending on observed attack modes, piecewise homogeneous variables, and elapsed time. Moreover, matrix decoupling and convexification techniques are employed to reduce the computational complexity. Finally, the effectiveness of the proposed method is demonstrated through two practical simulation cases.

In this paper, we address the problem of parameter identification for a Wiener nonlinear system with an autoregressive (AR) noise model. We propose two novel algorithms, namely the Wiener system generalized gradient iterative (WS-GGI) algorithm and the Wiener system zebra optimization (WS-ZO) algorithm. The WS-GGI algorithm is rooted in the gradient search principle, while the WS-ZO algorithm is a metaheuristic approach characterized by its robust parallel search capability. The core concept of the WS-ZO method is to identify the optimal solution by simulating the random movements of zebras within the search space and evaluating the objective function. The effectiveness of the proposed algorithms is demonstrated through experimental data and further compared with the recursive generalized least squares (RGLS) algorithm and the particle swarm optimization (PSO) algorithm. In conclusion, our findings indicate that these two new algorithms offer significant advantages in terms of accuracy and computational efficiency.

This article addresses the output feedback control problem for a specific class of discrete-time fuzzy singularly perturbed systems subjected to nonuniform sampling and a round-robin protocol. An innovative method for modeling nonuniform sampling periods through nonhomogeneous sojourn probabilities is proposed, offering a more intuitive and adaptable framework for system design and analysis. The round-robin protocol is applied to nonuniformly sampled outputs, optimizing information transmission efficiency and enhancing overall system performance. To tackle potential limitations in state data acquisition, a token-dependent static output feedback controller is developed that addresses the complexities introduced by nonperiodic sampling and asynchronous premise variables. Sufficient conditions are derived to ensure stochastic stability of the closed-loop system. Finally, two simulation examples are presented to validate and demonstrate effectiveness of the theoretical approach.

Autonomous aerial robots require accurate and efficient local environment representations to enable safe and agile navigation in cluttered and unknown environments. In this letter, we propose PolyMap, an online local polyhedral mapping-planning framework designed for aerial robots. PolyMap represents obstacles as a set of convex polyhedra constructed directly from raw point cloud, providing a compact and planner-friendly geometric abstraction. A novel concavity-aware decomposition algorithm is introduced to partition non-convex point cloud clusters into tightly fitting convex subcomponents, significantly reducing conservativeness while maintaining computational efficiency. Furthermore, we employ standard dual representation of convex polyhedra to achieve fast collision checking and enable seamless integration with optimization-based motion planners.High-fidelity simulations and real-world experiments are conducted to demonstrate the effectiveness and practicality of the proposed method.

As autonomous driving systems evolve towards higher levels of autonomy, large language models (LLMs) are increasingly being introduced for understanding complex traffic scenarios and motion prediction, demonstrating their potential in processing multimodal and unstructured information. The core argument of this paper is that the inherent characteristics of LLMs, such as their susceptibility to hallucinations and extreme sensitivity to input perturbations, fundamentally conflict with the determinism and robustness sought by classical control theory. This conflict is further amplified in the context of cyberattacks, potentially jeopardizing the closed-loop stability and safety of autonomous driving systems. This paper will discuss these issues from the core dimensions of control theory, including attack-detection-defend and control system, aiming to provide a critical perspective and potential research directions for building safe, reliable, and trustworthy next-generation intelligent transportation systems (ITS).

This article investigates a connected vehicle platoon with uncertain input delays in the absence of communication between vehicles. The primary control objective is to stabilize the platoon while estimating these uncertain input delays, ensuring that all vehicles maintain the same speed and a safe following distance. An observer is introduced to estimate the velocity and acceleration of the leading vehicle based on position information obtained from onboard sensors. In addition, an adaptive switching logic algorithm is designed to estimate the uncertain input delays for all vehicles. An improved decentralized controller is also proposed to enhance the control of the connected vehicle platoon. The stability analysis demonstrates the potential for achieving string stability within the system. Simulation results further validate the effectiveness of the proposed observer and controller.

This article addresses the mean-square exponential synchronization problem of reaction-diffusion neural networks (RDNNs) subject to stochastic switching and communication constraints. Different from existing Markov jump formulations that require exact transition probabilities or rely on constant sojourn probabilities, a dwell-time-dependent sojourn-probability switching rule is constructed to characterize random mode evolution in a more tractable form. To reduce the communication burden under random sampling, a random adaptive event-triggered protocol (RAETP) is developed, in which the triggering threshold is adjusted online according to the state error and the active sampling interval. Furthermore, a random spatiotemporal sampled-data control (RSTSDC) scheme is established by jointly introducing random temporal sampling, random spatial sampling, and switching gains into the RDNN synchronization framework. Based on this model, sufficient conditions are derived to guarantee mean-square exponential synchronization. Comparative simulations show that the proposed design achieves faster convergence and lower communication cost than several representative benchmark strategies.

This paper addresses the problem of distributed bipartite time-varying formation-containment (BTVFC) control for multi-agent systems (MASs) operating under Markov switching topologies and communication noise, with an active leader. The communication topology is modeled as a time-varying Markov chain, incorporating the effects of communication noise in real-world scenarios. The proposed framework enables multiple leaders to achieve a desired bipartite time-varying formation (BTVF) while tracking a virtual leader, while the followers maneuver within the convex hull defined by the leaders. A class of stochastic approximation controllers is introduced to handle the agent dynamics. Two algorithms are proposed to determine the control parameters, ensuring that the agents can meet the desired formation and containment criteria. The theoretical feasibility of achieving BTVFC is demonstrated using Lyapunov stability analysis with infinitesimal generators and indicative functions. Numerical simulations validate the effectiveness of the proposed method in achieving stable BTVFC, even under the influence of communication noise and topological uncertainties.

This paper proposes a novel robust tracking model predictive control (MPC) method within the Koopman framework for unknown nonlinear constrained systems. The key idea is to continuously update the Koopman surrogate model using new input-state measurements, while ensuring recursive feasibility of the MPC scheme under modeling errors through constraint tightening. The proposed method assumes that the modeling error is Lipschitz-continuous and estimates the corresponding Lipschitz constant from the training error in the dataset. Based on this estimate, a bound on the prediction error over the horizon is derived, leading to a convex robust optimal control formulation. Compared to existing Koopman-based approaches, the proposed method demonstrates superior modeling performance and improved data efficiency. Despite the need for online identification of the surrogate model and the construction of tightened constraint sets, the method remains real-time implementable. Furthermore, the input-to-state stability (ISS) of the closed-loop system is rigorously established. The effectiveness of the proposed method is demonstrated through numerical examples. (c) 2026 Published by Elsevier Ltd.

This article investigates the problem of asynchronous attack-compensated control for discrete-time nonlinear semi-Markov jump systems (S-MJSs) subject to false data injection (FDI) attacks. To mitigate the adverse effects of malicious data, a resilient controller is constructed by combining a false-signal observer with a compensation-based control law. Concurrently, considering limited network resources, a memory-based adaptive event-triggered scheme is proposed to enhance control performance while significantly reducing communication overhead. Recognizing the practical constraints in transition information identification, the semi-Markov kernel (SMK) and the high-level homogeneous Markov chain are assumed to be partially available. By employing a mode-rule-dependent Lyapunov function together with the linear matrix inequality method, sufficient conditions are derived to guarantee the $H_{\infty }$ performance of the control systems. Finally, a single-link robot arm model is employed to validate the efficacy of the proposed compensation control strategy.

This article addresses the problem of encryption -decryption-based bipartite synchronization control for a class of discrete-time coupled neural networks (CNNs), in which the nodes exhibit both cooperative and antagonistic interactions. Initially, a Markov chain with concealed operating modes is used to describe Markov jump CNNs (MJCNNs) with switching topologies (STs). In this framework, a hidden Markov model (HMM) is incorporated, whose emission values express the system mode. Next, the decentralized adaptive event-triggered strategy is proposed to alleviate the communication burden caused by interactions between nodes. Moreover, an encryption-decryption algorithm (EDA) that takes into account identity authentication is programmed to encrypt the data at the triggering moment of each node, thereby securing the data interaction privacy. Then, the observation-mode-based bipartite synchronization control law is formulated to fulfill the control demands of the plant. Furthermore, some sufficient conditions for the networks to be mean square synchronized and satisfy the H-infinity performance are obtained based on the Lyapunov stability theory. Finally, two simulation examples involving chaotic neural networks (NNs) are presented to verify the effectiveness of the proposed method.

This article focuses on the self-triggered estimator for nonlinear Markov jump systems (MJSs) subject to stochastic hybrid attacks, including deception attacks (DAs) and denial-of-service (DoS) attacks. To alleviate communication pressure and eliminate the need for continuous detection, a dynamic self-triggered mechanism (DSTM) is proposed, wherein the dynamic variable is adaptively adjusted to conserve network resources more efficiently. Considering the inherent openness of communication networks and their susceptibility to attacks, two Markov chains are employed to characterize the random switching between different attack strategies. With the help of a mapping technique, the system modes and attack modes are integrated into a new Markov chain. In addition, a hidden Markov model (HMM) with incomplete transmission probabilities is considered to represent the mismatched modes and the covert characteristic of attacks. Then, sufficient conditions ensuring system stability are derived, and a dissipative estimator is developed. Finally, two examples are utilized to illustrate the efficacy of the derived theoretical results.

Focusing on resilience enhancement, this article develops a novel control framework for quadrotor formation under Denial-of-Service (DoS) attacks, during which each quadrotor cannot accept any data from neighboring quadrotors or virtual leader. First, a fixed-time observer is proposed to estimate the virtual leader’s information under DoS attacks, and the explicit relationship between its convergence time and the attack parameters is established. Second, based on the estimated information, a decentralized position tracking algorithm is designed and the corresponding thrust and attitude extraction methods are developed. Then, an attitude control strategy is devised to effectively avoid the unwinding phenomenon. Through a comprehensive simulation example, the effectiveness of the proposed control scheme is finally confirmed.

This paper addresses the event-triggered fixed-time cluster synchronization of coupled reaction-diffusion neural networks under mixed cyber-attacks. First, some improved Lyapunov-based criteria for fixed-time stability are derived, offering a less conservative upper-bound estimate of the settling time. A novel event-triggering mechanism with a time-varying threshold is then introduced, and a security-based cluster synchronization scheme is developed based on this mechanism. By leveraging the piecewise Lyapunov function approach and hybrid systems analysis methods, we establish sufficient conditions to ensure fixed-time synchronization between follower nodes with interacting clusters and target nodes, even in the presence of mixed cyber-attacks. Moreover, this paper quantifies the relationship between mixed cyber-attacks and the convergence rate of the synchronization strategy. The effectiveness of the proposed approach is validated through comprehensive simulations and experiments, demonstrating its superiority over existing fixed-time stability strategy and event-triggered synchronization protocol.

Dear Editor, This letter proposes a distributed pursuit framework for multiple evaders with identical motion capabilities in obstacle environments under state measurement noise. The framework integrates dynamic pursuers allocation, chance-constrained collision avoidance Voronoi cell (C3AVC) construction, and path controller optimization to ensure reasonable allocation of multiple pursuers and probabilistic collision avoidance during the pursuit process, thereby addressing the challenges of multi-evader pursuit under imperfect perception conditions. Comparative simulations and experimental results validate the effectiveness of the proposed framework.

This paper presents a safe reinforcement learning (RL) framework for multi-agent systems based on data-driven distributed robust model predictive control (D3RMPC). Our approach leverages a parameterized D3RMPC as a function approximator, optimizing closed-loop performance through online updates to determine optimal policies for each agent. Unlike traditional MPC and existing MPC-based RL methods, D3RMPC relies solely on past system data and an implicit model based on behavioral system theory, eliminating the need for an explicit state-space model. By combining D3RMPC with the RL approach, we achieve explainable safety guarantees while overcoming the challenges of model mismatch. A formal theoretical framework ensures the preservation of safety, stability, and feasibility throughout the system's learning process and closed-loop operation. The efficacy of the proposed method is demonstrated through two numerical simulations.

This paper investigates the problem of resilient filter design for networked control systems (NCSs) subject to hybrid cyber attacksand measurement-channel input constraints. The considered hybrid attacks include denial-of-service (DoS) attacks, deception attacks, and replay attacks. A resilient filter incorporating a dynamic weighting parameter is designed to regulate the utilization of measurement information. Moreover, an adaptive event-triggered communication scheme is developed, where the triggering threshold is computed by the filter at triggering instants. An augmented filtering error system is constructed, and sufficient conditions are derived by employing a Lyapunov-Krasovskii functional together with the average dwell-time technique to guarantee the mean-square exponential stability of the resulting filtering error system. Finally, simulation results on a single-link robotic arm system demonstrate the effectiveness of the proposed filter design.

This article addresses the fuzzy formation tracking of uncertain nonlinear systems with inelastic performance and input constraints in a leader-follower configuration, where the leader dynamics are partially available to the followers. A distributed prescribed performance observer is developed for the followers to estimate the full states of the leader with a specified level of accuracy within a given time frame, improving both transient and steady-state performance of the estimation errors compared to existing approaches. Then, a deferred performance constraining function and a distance-dependent error transformation are introduced to ensure that all tracking errors converge to a compact set within a predefined time, under arbitrary initial conditions. Since the control magnitudes required to enforce strict performance metrics or handle large initial conditions are subject to input constraints, a fuzzy saturation-tolerant prescribed performance control scheme is proposed by employing reference modification systems. This approach mitigates the conflict between input saturation and inelastic performance specifications without explicit knowledge of the initial conditions. Finally, simulations and experiments with unmanned ground vehicles are conducted to validate the theoretical results.

Deploying multi-agent reinforcement learning (MARL) in safety-critical systems faces significant challenges due to insufficient agent exploration and inadequate safety constraint guarantees. Current approaches are constrained by two fundamental limitations: inefficient exploration leading to suboptimal policies, and expected-cost-based constraint frameworks failing to ensure full-process safety. To address these challenges, this paper proposes a novel safety-aware maximum entropy MARL framework using Conditional Value-at-Risk (CVaR) as a joint safety metric, which quantifies constraint satisfaction under worst-case scenarios for multi-agent systems. Moreover, we develop the Worst-Case Multi-Agent Soft Actor-Critic (WCMASAC) algorithm, incorporating sequential update mechanisms and maximum entropy optimization for heterogeneous agents, enhanced with distributed safety critics. Theoretically, we establish the monotonic improvement property, guaranteed constraint satisfaction, and convergence to a generalized Nash equilibrium for WCMASAC. Extensive experiments on Safety-Gymnasium based benchmarks demonstrate that WCMASAC outperforms state-of-the-art baselines in both task reward acquisition and safety constraint violation reduction, while exhibiting superior exploration efficiency and risk-aware control capabilities.

In this paper, an event-triggered prescribed-time resilient control scheme for uncertain Euler-Lagrange (EL) systems with deception attacks and deferred constraints is presented. First, to mitigate the effects of false data injection (FDI) attacks in the sensor channel, a novel coordinate transformation and the Nussbaum gain technique are applied under the framework of the backstepping method. The unknown actuator attacks are compensated by the application of a radial basis function neural network (RBFNN). Then, distinguishing from most existing resilient control algorithms that do not consider the transient characteristics of the output signal, the prescribed-time performance functions (PTPFs) and a barrier Lyapunov function (BLF) are merged in this paper, so that the output signal can converge to an adjustable set within a predefined time. To guarantee that the system satisfies the deferred constraint, a transformation function is introduced. What's more, we construct an improved event-triggered mechanism (ETM), which can utilize communication resources more efficiently. Finally, simulation results based on a two-link manipulator model are depicted to showcase the effectiveness of the proposed method.