BDS/VIO integrated navigation method based on improved unscented Kalman filter

In order to solve the problem that BeiDou satellite navigation system (BDS) satellite signals are susceptible to interference leading to degradation of navigation performance, the complementarity between BDS and VIO systems is applied to the combined navigation system by combining the advantages of monocular visual inertial odometry (VIO), which is not easily affected by external factors and has stable output. The combined BDS/VIO navigation is usually fused and optimized using untraceable Kalman filtering (UKF). The stronger the nonlinearity of the combined navigation system, the stronger the UKF filtering performance. However, the UKF cannot be adaptively adjusted when the system model is uncertain, which reduces the accuracy of the filter; to address the above problems, this paper optimizes the mean-square error array, adjusts the gain array of the filter filter in real time by introducing an asymptotic cancellation factor, and corrects the original error array by the new state-estimated mean-square error array when the system model is interfered with, so as to improve the robustness and Utility. Finally, the improved algorithm is compared with the standard UKF algorithm, and it is verified that the improved algorithm enhances the navigation accuracy of the BDS/VIO combination.