欠驱动UUV自适应RBF神经网络反步跟踪控制

摘要

An underwater unmanned vehicle ( UUV) model has an error that causes time-varying perturbation in the fluid.Therefore, the radial basis function ( RBF) neural network control technique was introduced in order to carry out an adaptive compensation estimate.Additionally, the backstepping method was utilized in order to design the position, attitude, and velocity controller of the UUV.Virtual speed was used in order to replace the attitude error for the purpose of converting the attitude tracking control to speed control.The simulation results show that this method was effective in improving the robustness and adaptability of the UUV.%针对水下无人航行器(UUV)模型存在一定的误差以及流体中存在时变扰动问题,引入径向基函数(RBF)神经网络控制技术进行自适应补偿估计,结合反步法设计UUV的位置、姿态及速度控制器,使用虚拟速度来代替姿态误差的控制手段,将姿态跟踪控制转化为对速度控制.仿真结果说明此种方法是有效的,提高了UUV的鲁棒性及自适应能力.