Target tracking of Autonomous underwater vehicle using a neural network disturbance observer

Abstract

This paper presents dynamic modeling and control studies of an autonomous underwater vehicle, tracking targets by considering water wave disturbances. A Sliding mode control approach is implemented to minimize the tracking inaccuracies. A neural network based disturbance observer is designed and added to the control loop to minimize the effect of disturbances. The under-actuated vehicle dynamics with three input torques completely guide the vehicle motion. The work can be extended for tracking of moving objects also. It is observed that the controller with NN observer is very effective in low speeds of operation. Even though the concept of proposed autonomous underwater vehicle control is not new, some highlights towards the practical perspective are given at the end.