水下无人航行器实时轨迹跟踪控制方法研究

Research on real time trajectory tracking control method for underwater unmanned vehicles

针对水下无人航行器受外界不确定因素影响产生航行准确度低的问题,研究水下无人航行器实时轨迹跟踪控制方法。分析水下无人航行器运动特性,设计基于贝叶斯神经网络的水下无人航行器自适应控制器,以水下无人航行器的参考速度信号为网络输入,通过网络的不断训练,输出控制力矩,促使水下无人航行器的航行点与预设航行位误差趋近于0,完成实时轨迹跟踪控制。实验结果表明,该方法可完成直线轨迹、圆形曲线轨迹的跟踪控制,且航行点与预设航行位误差小、收敛快、波动次数少,具备轨迹跟踪控制抗干扰能力强、适应能力强的优势。

To address the issue of low navigation accuracy caused by external uncertainties in underwater unmanned vehicles,a real-time trajectory tracking control method for underwater unmanned vehicles is studied.Analyze the motion characteristics of underwater unmanned aerial vehicles,design an adaptive controller for underwater unmanned aerial vehicles based on Bayesian neural networks,take the reference speed signal of the underwater unmanned aerial vehicle as the network input,continuously train the network,output control torque,promote the error between the navigation point of the underwater unmanned aerial vehicle and the preset navigation position to approach 0,and complete real-time trajectory tracking control.The experimental results show that this method can achieve tracking control of straight and circular curve trajectories,with small errors between the navigation point and the preset navigation position,fast convergence,and fewer fluctuations.It has the advantages of strong anti-interference ability and strong adaptability in trajectory tracking control.