Abstract: The centralized, distributed and combined information fusion methods are usually used by the traditional multi-sensor integrated navigation system, butthe multi-level information algorithm is the upgrade of the above methods, and can fulfill the fusion at the high grade space. Thus, the multi-levelinformation fusion method of the multi-sensor integrated navigation system without feedback information was studied, and then the theoretical analysisshows that this method is equivalent to the federated Kalman filter algorithm without feedback information. Finally, the SINS/GPS/SST/altimeter was usedto simulate and verify the above methods.

Key words: Federated Kalman filter, Multi-level fusion, Fusion algorithm, Integrated navigation system