MIMU/GPS Integrated Navigation Filtering Algorithm under the Condition of Satellite Missing

At present, Kalman filter is the most commonly used algorithm in the MIMU/GPS integrated navigation system. However, when the satellite signal is missing, inertial navigation system works independently instead of integrated navigation system, but with a serious problem of poor navigation accuracy due to the large drift error of micro inertial measurement unit. This paper puts forward a solution to this problem, which makes direct prediction of observation sequence in the absence of satellite signals. This method establishes the ARMA model of measurement values under the condition of historical observation, solves the parameters of the corresponding model, and obtains the predicted measurement values. The simulation experiments show that this measurement values obtained by the method can effectively modify the state error and further improve the precision of integrated navigation system.