Improved Adaptive Fading Memory Filter Based on State Switching Under Anomalous Measurement Data

With the swift evolution of integrated navigation technology in our country, it is essential to realize more accurate positioning of integrated navigation systems. In this paper, the INS/GNSS integrated navigation method is studied. Firstly, the system model is established according to the INS/GNSS integrated navigation principle. Then, based on the system model, the data fusion algorithms of Kalman filter memory filter (KF), FKF, and adaptive filtering are studied, to solve the problem of excessive use of fading factor in fading memory filtering, there is a state-switching-based fading memory filtering method (MFKF), however, this algorithm only considers the imprecise modeling of the Integrated Navigation System or the mismatch of the noise characteristics and does not consider the situation of the abnormal measurement data. To solve this problem, an improved adaptive fading memory filter based on state switching (IMFKF) is proposed by combining the innovation-based adaptive filtering method with the state switching fading memory filter and building the algorithm model that is more suitable for precise positioning. Simulation experiments verify the validity of the algorithm.