MEMS冗余惯组小幅值故障检测的PCA算法设计和优化

MEMS Redundant Inertial Measurement Unit Small Value of the Fault Detection of PCA Algorithm Design and Optimization

当组成冗余惯组的陀螺为MEMS陀螺,而故障又为幅值比较小的阶跃故障时(仿真发现,阶跃故障幅值为陀螺噪声方差的10倍左右大小时),使用以往的PCA故障检验方法,陀螺噪声会造成故障检测的困难。针对这种情况,提出了MEMS冗余惯组小幅值故障检测的PCA(Principal Component Analysis)算法设计和优化的方法,在以往PCA算法的基础上结合假设检验算法,考虑了陀螺噪声的影响,降低了MEMS陀螺噪声对故障检测的影响。通过仿真验证,证实了本文算法比以往PCA算法具有更高的检测准确性和灵敏度,这对于提高整个系统的可靠性具有重要意义。

When the composition of redundant IMU is MEMS gyroscope and is small amplitude step fault(simulation finds that step fault magnitude about 10 times the variance of noise in gyro), the noise of the gyro will cause the difficulty of fault detection if using previous PCA fault detection method. To solve this problem, this paper proposes MEMS redundant inertial measurement for small value of the fault detection of PCA algorithm design and optimization. This measure adds hypothesis testing algorithm based on previous PCA algorithm, through considering the influence of the gyro noise, which reduces the effect of noise of MEMS gyroscope of fault detection. The simulation results proved that this algorithm has higher detection accuracy and sensitivity than the previous PCA algorithm, which has great significance for improving the reliability of the whole system.