A Novel Real-Time Obstacle Avoidance Method in Guidance Layer for AUVs' Path Following
Chengren Yuan, Changgeng Shuai, Jianguo Ma
Abstract
This article addresses the real-time obstacle avoidance problem in 2D/3D path following missions for autonomous underwater vehicles (AUVs). To efficiently solve this problem, a novel universal obstacle avoidance framework based on the guidance layer is proposed for under-actuated AUVs to follow arbitrary referenced path. Based on this, a real-time avoidance line-of-sight guidance law is designed to avoid multi-static or slow time-varying obstacles that appear on the desired route and are pretreated by the oriented bounding box method. The strategy of single-obstacle avoidance, multi-obstacle avoidance, and vertical-obstacle avoidance is designed according to the geometrical features of unknown obstacles. Following the command of guidance law, kinematic and dynamic controllers are designed based on integral terminal sliding mode control to enhance robustness under environmental disturbances and parameter perturbation whether in path following or obstacle avoidance missions. Finally, Lyapunov theory proves stability of control system while numerical simulations demonstrate effectiveness and robustness of proposed control scheme.