科研之友
微信
新浪微博
Facebook
分享链接
摘要
This note focuses on the waypoints-based path-following problem of underactuated ships, where the reference path is surrounded by multiple static and moving obstacles. By virtue of the improved dynamical virtual ship (DVS) principle, a novel guidance law with multi-obstacles avoidance is proposed to generate the real-time attitude reference. In this approach, the manoeuvering tasks are specified as three priority levels: the static obstacle avoidance, the moving obstacle avoidance and the path-following mission. And the detailed design abides the International Regulations for Preventing Collisions at Sea (COLREGs) to ensure the sailing safety, especially for moving obstacles or other vessels. Furthermore, a robust neural algorithm is developed by using the neural networks (NNs) approximation and the robust neural damping technique. The system gain uncertainty of actuators is tackled and less information about the hydrodynamic structure, the actuator model and the external disturbances are required. Considerable efforts are made to obtain the semi-global practical finite time bounded (SGPFB) stability. The proposed scheme is with advantages of concise structure and improved autonomy. Finally, two experiments are employed to verify the effectiveness of the proposed strategy.
