GPS/INS integration based on adaptive interacting multiple model

The extended Kalman filter (EKF) had widely been used in the inertial navigation system (INS) and global positioning system (GPS) integrated navigation system. Nevertheless, the EKF contains instability caused by linearisation and numerous calculations of the Jacobian matrix. To solve the problem, the adaptive interacting multiple model (AIMM) filter algorithm is proposed in order to achieve a better navigation solution. The soft-switching feature, which is offered by interacting multiple model, allows conversion of process noise between lower and upper limits, meanwhile the measurement covariance is adjusted online by Sage adaptive filter. Considering the need to update the pseudo-range and Doppler data, the updating strategy of classification measurement is proposed. The results of the GPS/INS integrated navigation are estimated in the form of data of real ship, and experimental results indicate that the higher position accuracy can be obtained in AIMM filter.