Haitong Xu

A uniform semiglobal exponential stability (USGES) proof for a time-varying vector field guidance law used for path-following control of vehicles is presented. A sliding mode control is introduced for heading autopilot design and Lyapunov methods are used to derive the control law. The equilibrium point of the autopilot error dynamics is proven to be globally exponentially stable (GES). The main result is a time-varying vector field guidance law in cascade with the autopilot. A theorem ensures that the equilibrium point of the cascaded system is uniformly semiglobally exponentially stable. Both straight-line and curved-path path following scenarios are considered in the presence of ocean currents. Simulation studies are carried out to verify the theoretical results. The time-varying guidance laws can also be applied to vehicles in general such as aircraft, underwater vehicles, drones and autonomous vehicles. © 2019 European Control Association

This paper proposes a hybrid dynamic path planning algorithm that integrates global path planning (GPP) and local path planning (LPP). The objective is to effectively address the challenges associated with path planning and collision avoidance for unmanned surface vehicles (USV) in complex environments. The bidirectional A* algorithm is used to search the optimized strategy based on the actual geographical information map for the navigation routes of USV. The Dynamic Window Algorithm (DWA) is employed, and a novel cost function is formulated considering the USV’s manoeuvring characteristics, dynamic constraints, and environmental information. The primary purpose is to ensure the compliance of the USV with COLREGs during dynamic obstacle avoidance. It accomplishes this by providing positive reinforcement for achieving the correct turning speed when the USV determines the orientation of an approaching ship. Another evaluation function is introduced to mitigate the problem of local optima in complex environments that the LPP may encounter. The actual work map is used to model the application scenario with the obstacles in the presence of incoming ships, the simulation results show that the proposed algorithm is capable of generating adaptive, collision-free routes while adhering to COLREGs rules. © 2024 The Authors

This comprehensive review examines the works of reinforcement learning (RL) in ship collision avoidance (SCA) from 2014 to the present, analyzing the methods designed for both single-agent and multi-agent collaborative paradigms. While prior research has demonstrated RL’s advantages in environmental adaptability, autonomous decision-making, and online optimization over traditional control methods, this study systematically addresses the algorithmic improvements, implementation challenges, and functional roles of RL techniques in SCA, such as Deep Q-Network (DQN), Proximal Policy Optimization (PPO), and Multi-Agent Reinforcement Learning (MARL). It also highlights how these technologies address critical challenges in SCA, including dynamic obstacle avoidance, compliance with Convention on the International Regulations for Preventing Collisions at Sea (COLREGs), and coordination in dense traffic scenarios, while underscoring persistent limitations such as idealized assumptions, scalability issues, and robustness in uncertain environments. Contributions include a structured analysis of recent technological evolution, and a Large Language Model (LLM) based hierarchical architecture integrating perception, communication, decision-making, and execution layers for future SCA systems, which prioritizes the development of scalable, adaptive frameworks that ensure robust and compliant autonomous navigation in complex, real-world maritime environments.