An outlier rejecting and adaptive filter algorithm applied in MEMS-SINS/GPS

An outlier rejecting and adaptive filter is proposed against the influence of outliers in measurements and inaccurate noise statistical characteristics on Kalman filter. This algorithm computes the fading factors or performs outlier rejecting calculations by real-time monitoring the predicted residuals, such that the filter can still be optimal when there are outliers in measurements and when statistical characteristics are inaccurate. This approach is applied to the MEMS-SINS/GPS integrated navigation system, and simulation results show that the new algorithm is effective at the aspect of reducing the impact of outliers and inaccurate noise statistical characteristics.