A Potential Field-Based Model Predictive Target Following Controller for Underactuated Unmanned Surface Vehicles
Sen Han, Jiahao Sun, Shifeng Ding, Li Zhou
Abstract
This paper presents a target following controller for collision avoidance of underactuated unmanned surface vehicles (USVs). The primary aim is to control the USV from a starting point to a predetermined destination while avoiding collision with stationary obstacles and other moving vessels. The distinctive feature of this approach is the coherent combination among path planning, collision avoidance and motion control, all gathered in a nonlinear model predictive control (MPC) framework: i) an adaptive steering-constrained Theta* (ASC-Theta*) is proposed for generating a collision-free path considering the initial USV heading and turning angle; ii) potential functions are formulated to comply with various types of obstacles and COLREGs structures, hence ensuring safe navigation in narrow waters and complicated scenarios; iii) a nonlinear MPC controller is designed such that its objective includes potential functions along with the USV dynamics terms. The potential field-based model predictive target following controller is modeled and simulated for some intricate test scenarios of multi-USVs collision avoidance. The results show that USVs avoid the obstacles and observe COLREGs with appropriate USV dynamics.