Trajectory Tracking Control of Underwater Robots Based on Neural Network Backstepping Method

In terms of the trajectory tracking control of the spherical underwater vehicle, its movement is divided into the horizontal movement and the vertical movement. For the vertical movement, the motion model is built by transformation, and the trajectory tracking control law is designed by the neural network backstepping method. Simulation results and experiment results verify the effectiveness of the control law. The linear trajectory keeping method for the horizontal movement is designed according to the gyroscope approximation theory. Repeated experiments verify that the vehicle's high-speed rotation on the long axis can effectively weaken the course shift which is caused by flow interference.