Improved algorithm based on interacting multiple model and H_∞ filtering for IMU/GPS navigation
Zheng Nansha · Journal of China University of Mining and Technology · 2014
For coping with nonlinear error model with respect to low-cost inertial measurement unit(IMU)/GPS,an alternative filtering algorithm was put forth in the paper.The algorithm integrates interacting multiple model(IMM)and robust H∞filtering with Cubature Kalman filtering,which intends to improve the performance of integrated navigation.After IMM methodology and integrated navigation dynamic model were introduced,two filtering schemes for dealing with IMU/GPS integrated navigation were performed.The results indicate that the improved algorithm can shrink the errors of position and velocity estimation,and the accuracies of estimated position and velocity were improved 41.4%and 36.0%respectively.In addition,the proposed algorithm can effectively mitigate the fluctuation of filtering while systematic noises have not been correctly identified in advance,in particular,are far from the optimal.In other words,the hybrid algorithm can mitigate the effects from observation outlier effectively.