Nonlinear model predictive position control for a tail-actuated robotic fish

Abstract

Position control is a significant technique for the underwater application of robotic fish; however, it is also very challenging due to the underactuated property and input coupling of system dynamics. In this article, a two-stage orientation–velocity nonlinear model predictive controller is proposed to solve this problem. A scaled averaging model of tail-actuated robotic fish is constructed at first. Then, the novel strategy based on orientation and velocity control is developed as well as proved to be equivalent with position control in the sense of Lyapunov. Furthermore, a nonlinear model predictive controller with a two-stage switching strategy is designed to regulate the orientation and velocity error. Finally, the simulation results demonstrate the superiority of the proposed control algorithm compared with other methods. Particularly, there exists an interesting twist-braking behavior in simulation, which indicates that the proposed method makes better use of the system dynamics. The proposed method is efficient for not only bionic robotic fish but also other aquatic underactuated robots, which offers new insight into the position control of underwater robots.