基于LMS算法的水下航行器传感器故障检测

Fault Detection of Sensor for AUV Based on the Least-mean-square Algorithm

为了对水下航行器控制系统的传感器故障进行在线检测,采用有限脉冲响应(FIR)滤波器对水下航行器进行在线自适应建模,利用LMS算法来调节自适应滤波器的权系数,通过对FIR滤波器权系数的分析,实时检测出系统的传感器故障。通过Matlab仿真表明,该方法简单、可靠,可对水下航行器控制系统的传感器故障进行实时检测。

The paper uses a finite-impulse-response （FIR） filter for online adaptive modeling, and a leastmean-square （LMS） algorithm to minimize the output errors between the monitored system and the FIR filter in the modeling process. By analyzing the resulting adaptive FIR filter coefficients, Information were extract about fault occurrence. This method was applied to detect sensor faults of the control system of Autonomous Underwater Vehicles （AUV）, and evaluate the performance with sensor failures. Simulation results show that the method is simple to implement and effective in fault detection.