基于小波网络的水下机器人故障诊断研究

Study on the fault diagnosis for underwater robots based on wavelet neural network

由于水下机器人系统的复杂性、不确定性及强非线性等使得对其建模异常困难,采用一种最小调整的小波神经网络对其进行运动建模以获得理想的建模效果.通过该网络的自学习,调节小波函数的伸缩因子与平移因子以及网络连接权值,既能以任意精度逼近函数的整体轮廓,也能捕捉函数的变化细节,使得函数的逼近效果较好.对比模型的输出与实际传感器测量值来生成残差,通过分析残差特性来提取故障诊断判据,进而完成推进器故障诊断.完成了推进器故障诊断的仿真试验,仿真结果验证了该方法的有效性和可行性.

It is very difficult to model the system because of the complexity,uncertainties and strong nonlinearity of underwater robots.In order to make the modeling perfect,a wavelet neural network based on the least adjustment is applied.The adjustment of the scale factors and shift factors of wavelet and weights of WNN is studied.The WNN has the ability not only to approach the whole figure of a function at high accuracy but also to catch detail changes of the function,which makes the approaching effect preferably.Residuals are acquired by comparing the output of neural network with the sensor output.Fault diagnosis criteria are distilled from the residuals to execute thruster fault diagnosis.The simulation trials for thruster fault diagnosis are finished.The validity and feasibility of the method presented are verified by simulation results.