This article focuses on the secure predefined-time sliding mode control problem for a third-order heterogeneous vehicle platoon. For the purpose of reducing the effect due to the sensor measurement deviation and relaxing the system conservatism, a state estimation algorithm is designed based on the historical data and threshold judgment. Furthermore, a rate-coded reversible privacy-preserving mechanism with dual encryption is proposed and applied to vehicle-to-vehicle communications, which can guarantee the protection of the critical system data and the realizability of the desired predefined-time convergence performance by utilizing the origin outputs. In order to avoid the singularity problem and enhance the resistance of the vehicle platoon to the external disturbance, a nonsingular sliding mode surface and a corresponding predefined-time controller are designed. Based on the predefined-time stability and the string stability theorems, the vehicle platoon can be proved to be a practical predefined-time stable (PPTS) and string stable. Finally, adequate validations of the third-order heterogeneous vehicle platoon demonstrate the fast convergence speed and good robustness of the proposed control scheme.

Multi-agent trajectory prediction at signalized intersections is pivotal for the safety of autonomous driving and the efficiency of intelligent transportation systems. However, conventional vehicle-centric approaches are limited by restricted perception ranges and occlusion. Vehicle-to-Everything (V2X) cooperation is widely regarded as an effective approach to alleviating these limitations. Furthermore, existing cooperative systems often suffer from complex coupling and selection bias, hindering universal and real-time service. To address these challenges, this paper introduces a novel Infrastructure-to-Everything (I2X) collaborative prediction scheme. This scheme decouples infrastructure capabilities from vehicle requests by independently forecasting and broadcasting trajectories for all detected vehicles. Building on this scheme, we propose I2XTraj, a dedicated infrastructure-based model that leverages three core mechanisms. First, a continuous signal-informed mechanism to adaptively encode real-time traffic light information. Second, a maneuver strategy awareness mechanism that integrates intersection geometric constraints to estimate maneuver distributions. Third, a spatial-temporal-mode attention network to refine multi-agent interactions. Extensive evaluations on two real-world datasets, V2X-Seq and SinD, demonstrate the superiority of our approach. In both single-infrastructure and collaborative scenarios, I2XTraj outperforms state-of-the-art methods by over 30% and 15%, respectively, confirming its strong generalizability and robustness in complex intersection environments.

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.

With the widespread application of networked control systems in critical infrastructure, their security issues have become increasingly prominent. This paper investigates the synchronization control problem for T-S fuzzy multi-layer networks (MLNs) under the combined threats of deception attacks (DAs) and actuator faults (AFs). Due to the coupling characteristics of MLNs, the impacts of faults and attacks may propagate across layers, making it particularly difficult to guarantee both intra-layer synchronization (ALS) and inter-layer synchronization (RLS) simultaneously. To address this issue, an adaptive fault-tolerant control (AFTC) scheme integrated with a resilient dynamic event-triggered mechanism (RDETM) is proposed. In this scheme, the Bernoulli stochastic process is employed to describe the random occurrence of DAs, and an actuator fault model with unknown time-varying bias is established. The adaptive law is designed to compensate for fault parameters online, and the RDETM is introduced to reduce the frequency of data transmission, thereby decreasing the chance of data tampering by attackers. Theoretical analysis demonstrates that the proposed scheme can drive the synchronization error to converge to a bounded region related to the attack intensity and rigorously exclude Zeno behavior. Finally, numerical simulations verify the effectiveness of the proposed method.

Large-scale group decision making needs consensus rules that are robust to uncertainty yet easy to interpret. We propose TWD-SCD, a consensus framework built on three-way decision. At each step every decision maker is placed in a positive, boundary, or negative region by its distance from the group, and each region triggers a different update: keep the opinion, mix it mildly with the group mean, or average it over the social network. A lightweight Q-learning policy tunes the region thresholds online. We analyze the fixed-threshold kernel and the learning policy separately. For the fixed-threshold kernel, we give concrete, checkable conditions on the network and the thresholds under which a contraction holds, yielding almost-sure consensus when noise decays and a bounded error floor when noise is constant. A mean-field argument predicts a critical threshold, and a residual test bounds how well cooperative and non-cooperative decision makers can be told apart before consensus erases the evidence. We evaluate TWD-SCD against eight baselines on random graphs (up to N=3000), with further tests on the real email-Eu-core topology and on MovieLens ratings. It matches DeGroot’s consensus level at lower adjustment cost and lower variance under non-cooperative behavior.

This paper investigates fault-tolerant distributed Nash equilibrium (NE) seeking for noncooperative games in high-order nonlinear MASs over switching communication networks. Motivated by networked signal processing applications, the considered setting addresses practical challenges including limited information exchange, unmeasurable states, and actuator faults. The proposed distributed NE seeking scheme realizes gradient-play iterations within a backstepping-embedded distributed control architecture. An adaptive distributed observer is constructed to enable each agent to perform local information processing and communication using only neighbor information. In addition, fuzzy logic systems are incorporated as adaptive nonlinear approximators to compensate for uncertainties and actuator faults. Finally, rigorous convergence properties are established via Lyapunov analysis, and numerical simulations validate the effectiveness of the proposed approach.

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.

This article investigates a two-person zero-sum game for sequential decision making processes with dynamically evolving performance preferences. First, we incorporate this problem within a game theoretic framework, and establish the existence and uniqueness conditions for the temporal equilibrium strategy. Moreover, we present a compact formulation for one player's best response to the other's strategy. Furthermore, the specific convergence rate is achieved through the design of utility parameters. Finally, we validate our approach on the optimal liquidation and competitive supply chain models.

In this paper, the stabilization problem is concerned for a class of nonholonomic wheel mobile robots (WMRs). The main aim is to achieve control performance in sense of fixed-time stability in presence of uncertainties. Firstly, by virtue of differential flatness theory, the underactuated system is transformed to a dual-input-dual-output one. Taking advantages of tracking differentiator (TD), the moving trajectory is reasonably planned from initial to stabilization points. Secondly, considering lumped unmodeled dynamics and external disturbances, a reduced-order fixed-time extended state observer (FxTESO) is introduced for uncertainty estimation. Then, a FxTESO-based active disturbance rejection control law is developed for dynamic model, in which disturbance compensation is presented to attenuate uncertainty influence in real time. Finally, the effectiveness and advantages of the proposed methods are illustrated by numerical example. Consequently, not only the stabilization purpose, but also the transient performance can be desirably achieved for nonholonomic WMRs subject to uncertianties.

This article addresses the safe near-optimal formation control problem for multiagent systems in obstacle-cluttered environments. A fully distributed safe reinforcement learning (RL) framework is proposed to achieve near-optimal formation tracking and obstacle avoidance without requiring global topology information. First, control barrier functions (CBFs) are employed to characterize safety constraints, transforming the problem into a constrained optimization formulation. An explicit safe near-optimal control law is derived that decouples safety guarantees from the learning process, ensuring safety regardless of neural network (NN) convergence. Second, a critic-only NN with a finite-time learning law is developed, significantly reducing computational complexity and accelerating convergence compared with actor-critic methods. Third, a prescribed-time fully distributed estimator enables each agent to accurately estimate the leader state. The proposed approach is validated through both simulation and experiment, achieving formation convergence with guaranteed safety margins and demonstrating superior performance.

The problem of reinforcement learning (RL)-based fuzzy control for nonlinear systems with unknown dynamics via parallel composite policy iteration (PCPI) scheme is studied in this article. The main objective of this article is to solve the fuzzy algebraic Riccati equation (FARE), which is inherently complex and cannot be easily solved by traditional mathematical formulas. Policy iteration (PI) and value iteration (VI) algorithms proposed have been widely used to address this problem. However, these algorithms have the disadvantages of an initial stabilizing control policy, the persistent excitation (PE) condition, and huge amounts of data. To effectively alleviate these drawbacks, a novel PCPI algorithm is proposed in this article. Specifically, for each fuzzy subsystem, an adaptive parameter is designed to eliminate the requirement of an initial stabilizing control policy. In addition, an online model-free PCPI algorithm is proposed for the situation where the dynamic information of the fuzzy system is difficult to obtain. By substituting the stored historical data with online data, the PE condition is relaxed to the initial excitation (IE) condition. Concurrently, the corresponding algorithm can be executed independently and concurrently under each fuzzy rule, thereby fully exploiting the available computational resources. Finally, the effectiveness of the algorithms set forth in this article is verified through a single-link robot arm and quarter-car active suspension (QCAS) experiment.

This paper investigates the problem of optimal DoS attack strategies against distributed consensus fusion estimation. Using Q-learning algorithm in reinforcement learning and Markov Decision Process(MDP) in modelling attackers’ behavior, a novel optimal DoS attack strategy is developed, by iteratively updating the objective function, the method converges to the optimal attack strategy for varying network topologies. The proposed DoS attack strategy quantifies the impact of attacks on network connectivity (characterized by the smallest nonzero eigenvalue) and derives the optimal attack sequence under energy constraints through Q-learning iteration, significantly degrading the system’s consensus convergence speed. Compared with the existing literature, the proposed strategy resolves the inadequacy of current attack strategies in adapting to time-varying topologies. In the end, a numerical simulation example is presented to demonstrate the validity of the proposed DoS attack strategy against distributed consensus fusion estimation.

The lunar south polar region’s extreme illumination conditions impose strict energy constraints for solar-powered rover operations. Traditional Sun-synchronous path planning relies on dynamic time-dependent illumination evaluation, leading to high computational costs. We present CIRsE-Net, a spatiotemporal deep learning model that generates 72 h continuous illumination maps from hourly sequential illumination data. The model integrates a lightweight SST-VGG encoder (a customized 14-layer CNN for spatial feature extraction), BiGRU temporal modelling (a bidirectional recurrent network for capturing forward and backward temporal dependencies), and a consistency-aware spatiotemporal attention mechanism. On three different lunar illumination datasets (20 m/pixel, 5 m/pixel, and 20 m/pixel with a 2 m panel height), the model achieves Dice scores up to 0.983 and accuracy up to 0.985. When integrated with an enhanced 3ST-A* planner, the framework converts dynamic path planning into a static search task, reducing computational overhead while preserving path optimality and satisfying slope and illumination constraints. This work provides a validated methodological framework for Chang’E-7 mission planning and future lunar polar exploration missions by transforming dynamic path planning into a static search task.

Power cable insulation failures pose a significant risk of power outages, rendering insulation monitoring essential for ensuring the reliability of power systems. Existing schemes generally suffer from offline assessments, safety concerns, low sensitivity, and complex implementation. To overcome these challenges, this paper introduces a novel distributed online monitoring scheme that employs noncontact high-frequency (HF) injection to track insulation aging by quantifying changes in relative HF leakage current (Delta I %) associated with variations in insulation capacitance (Delta C %). In this scheme, an HF signal, with a frequency in the range of several kilohertz, is injected into the power cable and monitored through noncontact HF injection and sensing units positioned at existing cable connectors, all without disrupting system operation. This method also negates the need for precise data synchronization by relying solely on current amplitude measurements. The effectiveness, high sensitivity, and robustness of the proposed scheme are validated through both experimental and simulation results. Notably, at an optimal injection frequency, measurable Delta I % can approximately linearly reflect unmeasurable Delta C %. The minimum detectable capacitance variation is 1.11 %, while the maximum deviation of Delta I % remains below 5 % despite variations in system.

This study investigates the asynchronous controllability of non-homogeneous Markov switch generalized asynchronous Boolean control networks (NMHGABCNs) and random switching signals in these networks, aiming to follow a non-homogeneous Markov process. The controllability of the proposed networks is achieved using the discrepancy between the Markov chain mode and the control mode. Assisted by the semi-tensor product (STP), the algebraic forms of the NMHGABCNs are obtained, and the sufficient and necessary criteria for their asynchronous controllability are derived. The effectiveness of controllability is demonstrated through two examples, which validate the theoretical results.

This article investigates the stabilization problem of Interval Type-2 (IT-2) fuzzy semi-Markov asynchronous switching Boolean control networks (S-MASBCNs). The asynchronous phenomenon is a key consideration here. It arises from the lag between the jumping speed of the controller and that of the system.We aim to describe uncertain biological systems more accurately. To this end, we construct a global fuzzy semi-Markov jump asynchronous model. The construction is based on the semi-tensor product method. Subsequently, the largest invariant set method provides sufficient and necessary verification conditions for system stabilization. We also design a dwell-time-dependent control Lyapunov function. The design follows stochastic theory. This function is combined with a semi-Markov kernel and the IT-2 fuzzy method. Based on this strategy, we propose a relevant theorem. The theorem provides rigorous verification of system stabilization. It confirms whether the system can achieve the desired stability. Two examples demonstrate the effectiveness and superiority of the research results.

In this paper, the modeling of multi-agent systems(MASs) over finite fields is completed by using finite field networks(FFNs). Based on the combination of set stability and Lyapunov function (LF), the synchronization problem of finite-field networks (FFNs) and switched finite-field control networks (SFFCNs) is studied. First, the synchronization problem of FFNs is transformed into set stability, and the LF function of FFNs synchronization is given. Meanwhile, the necessary and sufficient conditions of SFFCNs synchronization are put forward, the algorithm to obtain the common control LF (CCLF) and mode-independent state feedback controller (MIDSFC) matrix X between each switched subsystem is provided. Moreover, SFFCNs with asynchronous state feedback control (ASFC) is taken into our study, the sufficient condition for its synchronization is presented. Two examples are given to verify the correctness of results.