A Robust Adaptive Filtering Algorithm for the IMU/GPS/BDS Tightly Coupled Integrated Navigation System based on Online Noise Feature Modeling

The Inertial Measurement Unit/Global Positioning System/BeiDou Navigation Satellite System (IMU/GPS/BDS) tightly coupled integrated navigation system and navigational positioning system uses observation information based on the code pseudorange. The system can effectively improve the accuracy of the pseudorange-observed quantity based on the carrier-phase smoothed pseudorange, thereby improving the accuracy of navigation and positioning. However, the measurement noise after pseudorange smoothing does not conform with the characteristics of white noise, which causes the Kalman filter to easily diverge. At the same time, the stability of the filter is more seriously affected due to the presence of cycle slips. In connection with the aforementioned problems, the characteristics of the smoothed pseudorange noise is analyzed and a noise model is established in this paper. On this basis, a robust adaptive filtering algorithm is designed to carry out online, real-time estimation and compensation on the measurement noise, which, in combination with the robust estimation theory, carries out filtering to reduce the effects on the filter brought by the level of measurement noise and the uncertainties of the model. Theoretical analysis and simulation results show that, in a complex environment, the robust adaptive tightly coupled integrated navigation system based on a carrier-phase smoothed pseudorange has higher positioning accuracy.