Minimum error entropy iterative extended Kalman filter for robust GPS/IMU/visual-integrated navigation under GPS anomalies

The integrated navigation of global satellite navigation system (GNSS) and inertial measurement unit (IMU) faces the problems of satellite signal limitation and IMU error accumulation in complex environments such as forests and long tunnels. Adding visual sensors is a good solution. In the majority of existing researches, non-Gaussian ambient noise is mostly regarded as Gaussian noise to simplify the processing, how to deal with non-Gaussian ambient noise remains a burning issue. In this paper, a minimum error entropy (MEE) based iterative extended Kalman filtering (IEKF) method for GPS/IMU/Visual-integrated navigation is developed. First, IEKF is employed to process GPS abnormal data and gradually approach the real state through an iterative optimization. Second, the MEE criterion is combined to enhance the robustness of the system to non-Gaussian noise, and the fixed-point iteration method is used to calculate the state estimation to improve the positioning accuracy of the system. Finally, the Canadian and Katwijk navigation data sets are utilized to validate the effectiveness of the proposed method. The results show that compared with the traditional EKF-based method, the proposed method can reduce the longitude error by 5.9% and 29.2%, and the latitude error by 19.3% and 56.6%, respectively, in GPS abnormal and non-Gaussian noise environments, and the accuracy and robustness of the navigation system are significantly improved.