关键词: 水下机器人;区域跟踪控制;滑模控制;神经网络控制

Abstract: The problem of region tracking control method for autonomous underwater vehicle is researched, and an adaptive region tracking control method based on PD-neural-sliding mode is proposed. Aiming at the problem of weak convergence capability of the radial basis function neural networks which only adjust the networks weights on line in the autonomous underwater vehicle adaptive controller, the simultaneous adaptive adjustment method for radial basis function neural networks weights, radial basis function centers and variance is given. The method makes radial basis function neural networks do not need to select radial basis function centers and variance offline, and online adaptive learning can be achieved. Considering that the system chattering problem which is easily caused by sliding mode item, a switch gain adjust method based on exponential function is proposed, which enables sliding mode switch gain real-time adjustment according to tracking error in order to reduce system chattering. The stability analysis of the proposed adaptive region tracking control method is carried out based on Lyapunov theory. The simulation experimental and pool experimental based on autonomous underwater vehicle are presented to demonstrate the effectiveness of the proposed method.

Key words: underwater vehicle;region tracking control;sliding mode control;neural networks control