基于改进自适应滤波算法的控制系统传感器故障诊断

The Improved Adaptive Filter Algorithm Applied to Sensor of Control System for Failure Detection and Identification

应用标准的多模自适应滤波算法进行故障诊断时,需要一组与假设模型等数量的卡尔曼滤波器进行滤波计算,计算量很大而且耗时,这对于实时性要求很强的工程应用是不合适的,因此提出了一种改进自适应滤波算法,它只需要单个卡尔曼滤波器就可得到与标准的多模自适应滤波算法等价的残差,因此可以有效地减少计算量和计算时间.将此算法应用于某无人机控制系统的传感器故障诊断,仿真结果验证了该方法的有效性.

Using standard multiple model adaptive filter to detect faults, a bank of Kalman filters are needed to finish filtering computation. When a large number of kalman filters are used, MMAE can be a very costly algorithm. An improved filter algorithm for MMAE is developed that uses a single Kalman filter to produce equivalent residuals of a complete Kalman filter bank. The computation cost of the method is less than the stand Kalman filter bank algorithm.