基于改进PID神经网络算法的AUV垂直面控制

针对一类小型低速自主水下航行器(AUV)的垂直面运动控制问题,设计了一种改进的PID神经网络控制器,实现对水下航行器在垂直面内深度和俯仰角的全局控制。利用REMUS水下航行器模型搭建了Simulink下AUV垂直面仿真控制系统,仿真结果表明,改进的控制方法克服了原方法中饱和区过大的问题,具有良好的动态性能同时能够适应不同的学习速率和网络初始权重,对水下航行器的工程实际应用具有一定参考价值。

AUV Vertical Plane Control Based on Improved PID Neural Network Algorithm

An improved PID neural network controller is designed for the movement control of a small low-speed autonomous underwater vehicle (AUV) in vertical plane, and the global control of the depth and pitch angle of the underwater vehicle in vertical plane is obtained. The AUV simulation control system is built by using REMUS underwater vehicle model in Simulink. The simulation results show that the improved control method with better dynamic performance has solved the original excessive saturation issue, and can adapt to different learning rates and network initial weight,and is of certain reference value to the practical application of the underwater vehicle.