水下机器人的神经网络自适应控制

Neural network adaptive control for underwater vehicles

研究了水下机器人神经网络直接自适应控制方法,采用Lyapunov稳定性理论,证明了存在有界外界干扰和有界神经网络逼近误差条件下,水下机器人控制系统的跟踪误差一致稳定有界.为了进一步验证该水控制方法的正确性和稳定性,利用水下机器人实验平台进行了动力定位实验、单自由度跟踪实验和水平面跟踪实验等验证实验.

A neural network direct adaptive control method is studied in this paper. By using Lyapunov theory, we proved that the closed-loop tracking error of the underwater vehicle is uniformly ultimately bounded （UUB） in the presence of external bounded disturbance forces and the neural network approximation error. In order to further verify the correctness, validity and stability of the proposed underwater vehicle control system, several pool experiments were also performed using an underwater vehicle experimental platform. These experiments included dynamic positioning experiment, single-degree-of-freedom trajectory tracking experiment and trajectory experiment in horizontal plane.