On Adapting Kalman Filtering for Adjusting Observation Noise Covariance in GPS/INS Integration

In GPS/INS integration algorithm,the effect of observation error on the integrated system should not be ignored,since the solution of GPS is not always precise enough.A new method is proposed that the measurement noise covariance matrix is adapted to improve the GPS/INS integration algorithm,based on the innovation sequence and the selected weight filtering.The experimental result shows that the abnormal errors are conspicuously controlled as well as the filtering precision of the parameters is improved.