Integrated Navigation Based on Robust Estimation Outputs of Multi-sensor Measurements and Adaptive Weights of Dynamic Model Information

In order to control the influences of outlying measurements and the kinematic model errors on the integrated navigation results, a robust estimation method and an adaptive data fusion method are applied. The new integrated navigation procedure is different from the federated Kalman filtering in four aspects. A three-segment robust weight function is introduced to construct the equivalent weight matrices for all local sensor measurement vectors, and an existing adaptive factor is directly applied to balance the contribution of kinematic model information and the preliminary integrated navigation result from the robust local sensor outputs. The calculation structure is given. An integrated navigation example using simulated data is illustrated in which three calculation schemes are performed.