Publications

Autonomous marine vehicles (AMVs), including autonomous surface and underwater vehicles, are versatile means to explore, exploit, monitor, and protect marine resources and environments. Motion control is a fundamental enabling technique for state-of-the-art AMV development. Especially, guidance is a critical component in AMV motion control. In recent years, line-of-sight (LOS) guidance, as an efficient guidance method, has attracted tremendous interest from both theoretical and practical perspectives. In this paper, an overview of recent advances in LOS guidance for AMV path following is provided. First, a control objective for the path following of an AMV with a kinematic model is specified. Next, major LOS guidance laws for path following are reviewed in detail. Then, LOS guidance laws applicable to coordinated path following of multiple AMVs are elaborated. Finally, six challenging issues for future research are addressed.

This paper addresses the problem of simultaneous cyber attack estimation and sensor attack detection for connected automated vehicles (CAVs), where the vehicle-to-vehicle communication network suffers from false data injection attacks and vehicular radar sensors experience spoofing attacks. First, a delicate proportional-integral observer (PIO) is developed to estimate the unavailable longitudinal vehicle tracking errors and the injected attack signal in real-time. Second, leveraging a residual signal generated from the PIO, an effective detection mechanism, which involves residual evaluation and thresholding, is designed to identify the radar spoofing attacks. Furthermore, due to the concurrent effects of false data injection attacks and spoofing attacks on the estimation errors, formal stability and performance analysis in terms of robustness against false data injection and sensitivity against spoofing is then performed. Finally, simulation examples are provided to demonstrate the efficacy of the proposed attack estimation and attack detection method.

This article is concerned with the synchronization control problem for a class of discrete-time dynamical networks with mixed delays and switching topology. The saturation phenomenon of physical actuators is specifically considered in designing feedback controllers. By exploring the mixed-delay-dependent sector conditions in combination with the piecewise Lyapunov-like functional and the average-dwell-time switching, a sufficient condition is first established under which all trajectories of the error dynamics are bounded for admissible initial conditions and nonzero external disturbances, while the l <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> - l <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> performance constraint is satisfied. Furthermore, the exponential stability of the error dynamics is ensured for admissible initial conditions in the absence of disturbances. Second, by using some congruence transformations, the explicit condition guaranteeing the existence of desired controller gains is obtained in terms of the feasibility of a set of linear matrix inequalities. Then, three convex optimization problems are formulated regarding the disturbance tolerance, the l <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sub> - l <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">∞</sub> performance, and the initial condition set, respectively. Finally, two simulation examples are given to show the effectiveness and merits of the proposed results.

With the rapid advancement of artificial intelligence, multi-agent systems (MASs) are evolving from classical paradigms toward architectures built upon large foundation models (LFMs). This survey provides a systematic review and comparative analysis of classical MASs (CMASs) and LFM-based MASs (LMASs). First, within a closed-loop coordination framework, CMASs are reviewed across four fundamental dimensions: perception, communication, decision-making, and control. Beyond this framework, LMASs integrate LFMs to lift collaboration from low-level state exchanges to semantic-level reasoning, enabling more flexible coordination and improved adaptability across diverse scenarios. Then, a comparative analysis is conducted to contrast CMASs and LMASs across architecture, operating mechanism, adaptability, and application. Finally, future perspectives on MASs are presented, summarizing open challenges and potential research opportunities.

Stealing attack against controlled information, along with the increasing number of information leakage incidents, has become an emerging cyber security threat in recent years. Due to the booming development and deployment of advanced analytics solutions, novel stealing attacks utilize machine learning (ML) algorithms to achieve high success rate and cause a lot of damage. Detecting and defending against such attacks is challenging and urgent so that governments, organizations, and individuals should attach great importance to the ML-based stealing attacks. This survey presents the recent advances in this new type of attack and corresponding countermeasures. The ML-based stealing attack is reviewed in perspectives of three categories of targeted controlled information, including controlled user activities, controlled ML model-related information, and controlled authentication information. Recent publications are summarized to generalize an overarching attack methodology and to derive the limitations and future directions of ML-based stealing attacks. Furthermore, countermeasures are proposed towards developing effective protections from three aspects -- detection, disruption, and isolation.

This article is concerned with the leader–follower formation control problem of multiple networked mobile robots (MRs), where the information exchanges over communication networks among the MRs suffer from replay attacks. The central aim is to develop an effective secure control scheme for the multiple-MR system such that the desired leader–follower formation control objectives are accomplished regardless of the simultaneous presence of replay attacks, network-induced delays, system uncertainties, and external disturbances. Toward this aim, an extended state observer is first constructed to estimate the unknown nonlinear terms consisting of system uncertainties and external disturbances. Then, leveraging the time-stamp technique, a dedicated data packet analyzer is developed for each MR to detect and handle replay attacks. Furthermore, a secure leader–follower formation control scheme, consisting of a network-based cooperative kinematic control law and two local kinetic control laws, is designed. Finally, both simulation and experiment results are provided to validate the efficacy of the proposed secure leader–follower formation control scheme.

This article is concerned with the exponential mean-square stabilization problem for a class of discrete-time strict-feedback nonlinear systems subject to multiplicative noises. The state-dependent multiplicative noise is assumed to occur randomly based on a stochastic variable obeying the Gaussian white distribution. To tackle the difficulties caused by the multiplicative noise, a novel backstepping-based control framework is developed to design both the virtual control laws and the actual control law for the original nonlinear system, and such a framework is fundamentally different from the traditional n-step predictor strategy. The proposed design scheme provides an effective way in establishing the relationship between the system states and the controlled errors, by which a noise-intensity-dependant stability condition is derived to ensure that the closed-loop system is exponentially mean-square stable for exactly known systems. To further cope with nonlinear modeling uncertainties, the radial basis function neural network (NN) is employed as a function approximator. In virtue of the proposed backstepping-based control framework, the ideal controller is characterized as a function of all system states, which is independent of the virtual control laws. Therefore, only one NN is employed in the final step of the backstepping procedure and, subsequently, a novel adaptive neural controller (with modified weight updating laws) is presented to ensure that both the neural weight estimates and the system states are uniformly bounded in the mean-square sense under certain stability conditions. The control performance of the proposed scheme is illustrated through simulation results.