Underactuated USV Path Following Control Approach for Inland River Search and Rescue

With the increasing importance and complexity of searching mission in inland river, it is a current research hot topic to require underactuated Unmanned Surface Vehicle (USV) to offset the large drift angle caused by the harsh environment, so as to accurately and efficiently arrive at the accident site to search and rescue people who fell into the water. In response to the problem, this article designs a Line-of-sight (LOS) guidance method based on the Extended State Observer (ESO) to accurately estimate the time-varying large sideslip angle, consequently, achieving the desired heading angle. Taking into account the presence of unknown time-varying disturbances and dynamic uncertainties, this paper introduces an estimator to improve the Radial Basis Function (RBF) neural network and designs an adaptive path-following controller. By employing a Radial basis function neural network that utilizes estimator-based error estimation, the approximation of unknown terms in the model is achieved, resulting in enhanced transient performance and robustness of the controller. Ultimately, the effectiveness and robustness of the algorithm are validated through stability analysis and comparative simulation experiments.