In order to improve the sensitivity of small abrupt fault and slow change fault detection of micro-electro-mechanical-inertial navigation system/global navigation satellite system(MINS/GNSS)measurement information,a robust adaptive filtering algorithm for the detection of innovation change rate in finite memory period is proposed.Based on the sequential probability ratio test,combining robust estimation and sequential filtering,precise detection and processing of individual measurement information are carried out using the innovation rate of change in a finite memory period.Simulation and experimental results show that the detection sensitivity and filtering accuracy of the proposed algorithm are improved compared with the residual Chi-square detection method and the robust adaptive filtering algorithm based on residual Chi-square detection.In the vehicle abrupt fault experiment,the missed alarm rates of latitude and eastward velocity are reduced by 83.3% and 90.5%,respectively,and the errors of latitude and eastward velocity are reduced by 17.1% and 25.3%,respectively.In the vehicle slow change fault experiment,the missed alarm rates of latitude and eastward velocity are reduced by 68.8% and 66.7%,and the errors of latitude and eastward velocity are reduced by 54.7% and 23.9%,respectively.
维普
万方数据
