This paper investigates a pursuit-evasion differential game problem of two groups of underactuated autonomous surface vehicles (ASVs) in a complex obstacle environment. It is assumed that only a portion of the ASVs from the opposing sides can access the location information of the adversary. A safety-critical pursuit-evasion game guidance strategy for multi-ASVs is proposed to achieve the collision-free pursuit-evasion mission of pursuing and evading ASVs. First, distributed estimators are designed for two groups of ASVs to achieve the distributed estimation of the velocity and position of adversarial counterpart. Second, leveraging estimated information, an optimal pursuit-evasion strategy for ASVs is proposed based on differential game theory and optimal control theory, featuring intra-team cooperation and inter-team competition. Finally, to ensure the safety of the pursuit-evasion ASVs, a quadratic programming problem is formulated by incorporating input constraints and rate-tunable control barrier functions. Through the solution of the quadratic programming problem, optimal safety-critical guidance laws for both the pursuing and evading sides are obtained. The closed-loop pursuit-evasion system is proven to be input-to-state stable through Lyapunov stability analysis. The effectiveness of the proposed method is verified by simulation and experiment results.

Ensuring collision-free navigation is essential for the safe operation of multiple autonomous surface vehicles. This paper addresses the problem of safe containment maneuvering of multiple underactuated ASVs in unknown environments. A distributed containment maneuvering controller based on local LiDAR perception is proposed, which allows the ASVs to move into the convex hull formed by multiple virtual leaders following parameterized paths, achieving coordinated formations while maintaining collision-free behavior. First, the containment maneuvering control law is developed based on the information of the neighbor information. Then, by introducing path parameter deviations among multiple virtual leaders, a path parameter update law is designed. Finally, obstacle information is obtained through LiDAR perception, and a Gaussian process is employed to construct a control barrier function-based safety constraint. The designed containment maneuvering control law is optimized via quadratic programming to ensure real-time safety. Lyapunov stability analysis demonstrates that the overall control system is asymptotically stable, ensuring the safety of the system. The simulation results confirm the effectiveness of the proposed approach.

This paper addresses the cooperative target tracking control problem for multiple tracking autonomous surface vehicles (ASVs) under false data injection (FDI) attacks from the target ASV. A model predictive control (MPC) - based resilient cooperative target tracking control scheme is proposed for mitigating the effect of the attacks and restoring the tracking performance. Specifically, a nominal cooperative target tracking control law is designed to achieve cooperative target tracking without FDI attacks where a finite-time extended state observer (ESO) is proposed to estimate the model uncertainties. Next, the FDI attacks from the target ASV are modeled by using a MPC method to hinder the cooperative tracking of multiple ASVs. Then, optimal resilient signals are developed via the MPC approach to counteract the attack effects and thereby recover cooperative tracking performance. The input-to-state stability (ISS) of the proposed finite-time ESO error dynamics is analyzed by employing a homogeneous Lyapunov function, besides in the entire closed-loop control system, all tracking error signals are uniformly ultimately bounded. The efficacy of the proposed resilient control method is confirmed through simulation results.

This paper addresses the distributed containment control problem for autonomous surface vehicles (ASVs) with model uncertainties and external environmental disturbances under a structured cost index. A distributed fuzzy optimal control method based on a predictor-actor-critic architecture is proposed to achieve containment tracking control for ASVs. First, a finite-time data-driven fuzzy predictor is developed using integral concurrent learning to estimate the lumped uncertainties. Next, using the predictor and reference trajectories information, a performance index including tracking errors and control inputs is constructed. The corresponding Hamilton-Jacobi-Bellman (HJB) equation is derived to obtain the optimal control solution. Then, critic fuzzy logic systems (FLSs) are developed to approximate the unknown partial derivatives coupled with the control input in the HJB equation. Furthermore, actor FLSs are trained using a gradient descent method to approximate the critic FLSs. Finally, a distributed fuzzy optimal controller is designed based on the actor FLSs. Through Lyapunov stability analysis, it is proven that the closed-loop system is input-to-state practically stable. Simulation results demonstrate that the proposed distributed containment tracking control method is effective for ASVs based on the predictor-actor-critic architecture.

This paper presents a parallel path following control scheme for cyber-physical maritime autonomous surface ships (MASSs) equipped with dual-podded propulsion systems. Specifically, an artificial MASS system is constructed using a deep neural predictor with a dual-time deep learning architecture, expanding the path following problem into the virtual space by integrating virtual-reality interaction. Then, based on information provided by the artificial MASS system, a parallel control law with integral action is formulated through computational experiments to ensure accurate tracking of the predefined path by the actual MASS. After that, the derived parallel control law is utilized to simultaneously drive the actual and artificial system during parallel execution. Stability analysis proves that the closed-loop control system is input-to-state stable, and all signals are uniform ultimate bounded. Notably, unlike the existing parallel path following methods that employ generalized forces as control inputs, the proposed approach directly regulates both of the rotational speed and pod angle, thereby enhancing the feasibility in practice. Simulation results are provided to illustrate the efficacy of the proposed approach for an MASS subject to environmental disturbances.

This paper addresses the problem of safety-critical path-guided herding control for multiple-input multiple-output multi-agent systems under incomplete state measurement, safety constraints, and limited communication resources. Specifically, an output-feedback finite-time neural predictor is proposed to identify both model uncertainties and unknown state information. Subsequently, a distributed path-guided herding control strategy is designed, including patrolling, gathering, enclosing, and expelling. Control barrier functions are formulated as safety constraints, and a quadratic optimization problem is established. By using the neurodynamic optimization to solve the optimization problem, the optimal control law satisfying both safety constraints and state constraints is generated. Furthermore, a dynamic event-triggered communication method is proposed to reduce unnecessary communication, especially during the transient phase. By using the proposed herding control approach, the collision-free herding is guaranteed for input-to-state stability, and simulation results are provided to demonstrate the effectiveness of the approach.

The International Maritime Organization (IMO) proposed the concept of a maritime autonomous surface ship (MASS), which could integrate perception, decision-making, and control functions to implement autonomous navigation in complicated environments. It is recognized as the fundamental component of the future new generation of shipping systems. The development of MASS requires the use of specific test tools and use cases for the evaluation of the safety, reliability, and functional performance of its navigation system. At present, the lack of standardized, modular, and serial testing paradigms and methods for MASS impedes the development and improvement of related products and applications. Therefore, it is essential to develop a practical and applicable testing framework. This paper presents a functional analysis of autonomous navigation systems and provides a comprehensive overview of current research status, potential solutions, and future challenges. Moreover, this paper also takes a step towards facilitating the functional testing of autonomous surface vehicles by proposing a hybrid virtual-real framework consisting of scenario generation, virtual simulation, model-scaled physical experiment, validation and evaluation. The research opportunities and future aspects are also addressed. This work may serve as a reference for academic and industrial researchers to investigate new methods and to develop prototype systems for future autonomous surface ships.

To address the challenges of traditional physiological signal monitoring—such as its invasive nature—and the limitations of existing computer vision solutions, including unreliable extraction of micro-expression features (e.g., eyelid closure, yawning frequency), high computational redundancy, and the resource constraints of edge devices, this paper proposes a real-time fatigue detection system based on an improved YOLOv11 algorithm, designed for operational scenarios such as cockpits. The core contribution involves an innovative redesign of the backbone network through the introduction of the EfficientNet compound scaling mechanism, which optimizes the network across depth, width, and resolution. This enhancement significantly improves feature detection capability for small-scale targets. Additionally, a new SDLoss function is designed to increase localization accuracy, effectively alleviating the high sensitivity of IoU to positional deviations in small objects. Experimental results demonstrate that the improved model achieves a 1.619

Finite control-set model predictive control (FCS-MPC) has emerged as a promising control scheme for power converters. However, its dependence on precise physical parameters presents challenges to practical implementation. To address this inherent model dependence issue, this article proposes a data-driven predictive control strategy for LC-filtered voltage source inverters, utilizing full-form dynamic-linearization (FFDL) technique. To mitigate the inherent computational delay in digital implementations, conventional model-based and model-free MPC typically adopt a sequential prediction framework. In contrast to this widely deployed sequential prediction structure, this article constructs a holistic input-output data model that directly captures the mapping dynamics from the current instant to the delay-compensated future. Consequently, neither physical parameters nor delay compensation procedures are required in this design, rendering a simplified implementation. Furthermore, by leveraging the comprehensive data structure of FFDL, the proposed strategy achieves direct output capacitor voltage control in the absence of inductor current sensors and specific circuit parameters, thereby enhancing system reliability and reducing computational cost. Comparative experimental results on a three-level neutral-point-clamped (3L-NPC) inverter validate the effectiveness of the proposed method, demonstrating its computational efficiency and strong robustness against parameter mismatches compared to existing schemes.

Considering the actuator wear and energy consumption for the unmanned surface vehicle (USV) with dynamic environmental disturbances, a low-conservative intermittent activation area-keeping control scheme is proposed based on the adaptive event-triggered weather optimal control (WOC) in this article. Firstly, a WOC method is employed to keep the USV reach the desired position and optimal heading based on the direction of dynamic environmental disturbances. Then, an adaptive event-triggered mechanism is introduced into the WOC method to reduce the frequent updates of the desired position and optimal heading caused by dynamic environmental disturbances. During the process of USV area-keeping, a low-conservative intermittent controller is designed based on input-to-state safe control barrier function (ISSf-CBF) to achieve the reciprocating motion of USV with less energy consumption. The proposed ISSf-CBF-based method can transform the conservative and fixed activation boundary of controller into a low-conservative controller activation position, and achieve the safety-critical area-keeping control for the USV, which means that the USV can approach the safety area boundary as much as possible but does not exceed it. Finally, the obtained simulation results verify the effectiveness and advantages of the proposed control scheme.

Considering thruster faults and unknown external environmental disturbances, this paper proposes a prescribed-time fault-tolerant control scheme for a dynamic positioning (DP) vessel based on the fault reconstruction strategy. Firstly, a sliding mode iterative learning observer (SMILO) is designed to reconstruct thruster faults by combining iterative learning strategy with sliding mode technique. Then, utilising the reconstructed fault information, a prescribed-time iterative learning sliding mode fault-tolerant controller is further developed, which adaptively compensates for unknown disturbances and reconstruction errors to improve tracking accuracy. Furthermore, the prescribed-time sliding mode surface ensures that the DP vessel's motion states converge at a preset time, allowing for more flexible adjustment of the system's convergence time based on actual operational requirements. The stability of the SMILO and control scheme is analysed by Lyapunov theory, and simulation results verify the effectiveness of the proposed methods.

To reduce parameter sensitivity and computational complexity, this paper proposes an improved three-vector modelfree predictive current control. The proposed method utilizes an ultra-local model to replace the traditional predictive model and integrates an extended state observer (ESO) with an adaptive gain update mechanism for online estimation of total system disturbances and control gain, thereby enhancing prediction accuracy under parameter mismatch conditions. Furthermore, the durations of optimal voltage vectors are directly determined using the current error under the zero vector, which effectively reduces computational complexity. Simulation results demonstrate that the proposed method provides considerable improvement in parameter robustness compared with the conventional MPC scheme.

This paper studies the collision-free cluster formation control of autonomous surface vehicles subject to unknown internal uncertainties and unknown external disturbances. A collision-free cluster formation control method is presented based on fuzzy predictor to achieve safe formation tracking control in obstacle environments. Specifically, a fuzzy predictor is designed by using the fuzzy logic system. Then, a collision-free cluster formation controller based on artificial potential functions is proposed, which can autonomously avoid inter-agent collisions and static obstacle. By using the Lyapunov stability analysis, the closed system is input-to-state. Simulation results verify the effectiveness of the proposed method.

This paper addresses the fault-tolerant optimal tracking control design problem for marine dynamic positioning under actuator faults, disturbances, and model uncertainties. A novel control strategy integrating the adaptive dynamic programming (ADP) and the nonlinear observer is proposed to achieve precise trajectory tracking, energy optimization, and fault tolerance simultaneously. Firstly, an observer-based backstepping controller is designed to estimate composite term including external disturbances, uncertainties, and actuator faults. Then, the ADP technique with a single critic network is employed to solve the nonlinear optimization problem for the error system, realizing optimal control performance while minimizing the cost function. The stability of the closed-loop system is rigorously proven via Lyapunov theory, ensuring that all signals are uniformly ultimately bounded. Simulation results demonstrate the superiority of the designed strategy.

This paper studies the collision-free cluster formation control problem for high-order nonlinear multi-agent sys tems subject to completely unknown internal uncertainties, unknown external disturbances, and unknown control input gains. A collision-free cluster formation controller based on a finite-time data-driven fuzzy predictor is pro posed to achieve safe cluster formation tracking control in obstacle environments. Specifically, a finite-time data-driven fuzzy predictor is designed by using a fuzzy logic system with historical data information. The proposed predictor demonstrates the capability to simultaneously estimate internal system uncertainties, unmea surable external disturbances, and unknown control input gains, while maintaining estimation error boundedness in the absence of persistent excitation. Then, a collision-free cluster formation controller based on artificial po tential functions is proposed, which can autonomously avoid inter-agent collisions and multiple static obstacles. Through Lyapunov theory, the input-to-state stability of the closed-loop control system is demonstrated. The effectiveness of the collision-free cluster formation control method is verified via simulations of autonomous surface vehicles with entirely unknown system dynamics.

Berthing large vessels poses significant challenges due to reduced rudder effectiveness and substantial mass inertia under low-speed conditions. This paper presents a collision-free guidance and control method for collaborative berthing of autonomous tugboats subject to static obstacles, state constraints, as well as environmental disturbances. Specially, a two-layer control architecture is proposed for ATs to address these challenges. In the kinematic layer, a nominal line-of-sight guidance control law is designed for each AT. Based on the nominal velocity signals, a quadratic programming problem is formulated to generate safety-critical collaborative guidance velocity signals subject to contact force constraint, velocity constraints, payload constraints, and safety constraints which are encoded based on control barrier functions. In the kinetic layer, practical predefined-time extended state observers are employed to estimate model uncertainties and ocean disturbances. Based on the estimated information and the optimized velocity signals, anti-disturbance kinetic control laws are developed. It is proven that the collaborative berthing system is uniformly ultimately bounded. Simulation results demonstrate the efficacy of the proposed safety-critical collaborative berthing control methodology.

Autonomous Surface Vehicles (ASVs) in future mixed maritime traffic environments may need to avoid obstacles exhibiting rule-violating behaviors. This paper proposes an ASV collision avoidance framework targeting rule-violating behaviors, systematically integrating risk assessment, behavioral intention inference, and velocity planning that considers obstacle course angle uncertainty. A dual-ellipse compound domain model is designed to enable early warning, and a Hidden Markov Model (HMM) is employed to infer obstacle intentions based on observed behaviors. To enhance system robustness, course angle uncertainty is incorporated into the decision-making process. Based on statistical analysis of real-world maritime accidents, nine types of rule-violating behaviors are identified, and corresponding avoidance strategies are developed for each type, enabling the ASV to adaptively respond to different complex encounter scenarios. The proposed framework is validated in 16 typical scenarios, including head-on, crossing, overtaking, being overtaken, and multi-vehicle encounters. The proposed method is compared with the standard Velocity Obstacle (VO) method through Monte Carlo simulations in scenarios involving encounters among five vehicles, verifying its stability and effectiveness. Key hyperparameters influencing avoidance performance are then analyzed. This work provides a feasible solution to address the challenge of avoiding rule-violating obstacles in maritime navigation.

Navigation planning in ice-covered waters is one of the key technologies for advancing polar shipping. While existing research has mainly addressed large-scale global route planning, effective local methods for navigating broken ice regions remain limited. This study proposes a local path planning method that identifies energy-optimal paths while satisfying maneuverability constraints of vessels operating in such environments. Vessel maneuvering in small-scale broken ice fields is evaluated with the Nonsmooth Discrete Element Method combined with the Maneuvering Modeling Group model (NDEM-MMG), while ice resistance is estimated via empirical formulas. Building upon the conventional Position-based Dubins-RRT* (P-Dubins-RRT*), a Heuristic P-Dubins-RRT* (HP-Dubins-RRT*) algorithm is developed by constructing a heuristic map based on ice field information. This map enables non-uniform probabilistic sampling that accounts for both ice resistance and distance, significantly improving planning efficiency. Simulation results verify the effectiveness and superiority of the proposed method. Compared to Hybrid A* and P-Dubins-RRT*, the HP-Dubins-RRT* algorithm generates lower-energy paths in shorter computation time. In simulation tests, the planned paths reduce energy consumption by an average of 22.7% compared to the shortest paths. This method is applicable to autonomous navigation and decision-support systems for polar transport vessels with limited icebreaking capability.

In this paper, a digital twin-enabled parallel collaborative berthing control method is proposed for autonomous tugs (ATs) subject to multiple safety constraints in the presence of ocean disturbances. Specifically, an artificial AT system (AATS) is developed based on the digital twin technique for each AT in cyber space. A three layer control architecture is developed, including a cyber kinematic layer, a cyber optimization layer, and a kinetic layer. At the cyber kinematic level, a virtual kinematic control law is designed for each AT based on the AATS established by a predictor using neural networks. At the cyber optimization level, using the guidance signals, a quadratic optimization problem is solved to obtain virtual collaborative velocity signals for each AT subject to intervehicle collision avoidance constraint, velocity constraints, payload constraints, and port shoreline constraints, which are encoded by reciprocal control barrier functions. At the kinetic level, based on the optimized virtual collaborative velocity signals, parallel collaborative berthing control laws are designed for each AT to simultaneously drive both AATSs and physical AT systems within the cyber-physical space. It is proven that the closed-loop control system is input-to-state stable and all signals are uniformly ultimately bounded. Numerical examples are shown to illustrate the effectiveness of the proposed parallel collaborative berthing control method for ATs.

This article investigates a secure pursuit-evasion problem involving autonomous surface vehicles (ASVs) under dual-channel false data injection (FDI) attacks. In this scenario, the pursuit process is interfered with by the evading ASV through independent tampering of either or both the guidance and control channels of the pursuing ASV. A nonlinear model predictive control (NMPC)-based dual-channel FDI resilience method is proposed for the pursuing ASV such that the capture is to be guaranteed regardless of attacks. In the guidance channel, a min-max pursuit-evasion guidance law is first developed based on NMPC. The pursuing ASV anticipates the worst-case strategy of the evading ASV, and its own optimal strategy is subsequently determined to ensure successful capture. Furthermore, within the NMPC framework, a min-max attack mitigation strategy is proposed to predict the most adversarial attack scenarios from the perspective of the evader and achieve proactive mitigation of the detrimental impacts induced by the attacks. In the control channel, a novel min-max resilience neural network-based control law is introduced. Sharing with the same proactive mitigation mechanism, the control law ensures accurate tracking of the guidance commands regardless of uncertainties, disturbances, and adversarial cyber attacks. Finally, simulation and experimental results validate the effectiveness of the proposed mitigation mechanism in countering dual-channel FDI attacks while ensuring pursuit-evasion performance.