Robust Integrated Navigation Filtering Method Based on Unknown Input Estimator
Zihe Mao,Guochang Zhang,Ya Zhang,Qiang Hao,Fei Yu
DOI: https://doi.org/10.1007/978-981-19-6613-2_651
2023-01-01
Abstract:Aiming at the problem that the uncertainty of the system model and the interference of environment lead to the decline of the filtering performance of the integrated navigation system, firstly, the model of SINS/GPS integrated navigation system was established, the characteristics of this system were pointed out, and the reasons why the existing relevant algorithms can’t be directly applied to integrated navigation system were summarized. Then, the unknown input estimator was established according to the innovation and the value of states in the past, and the exponential weighted sliding window was used to judge the position of the unknown input and decouple it. Finally, the effectiveness of the algorithm proposed in this paper was verified by simulation experiments and on-vehicle experiments. Through the comparison with Kalman filter and adaptive Kalman filter, the results showed that the filtering effects of these three algorithms are almost the same under normal conditions, and when there was unknown input, the performance of proposed algorithm was significantly better than the other two algorithms, which means proposed algorithm significantly improved the robustness of integrated navigation system.