Abstract: For the problem that the navigation measurement noise of near-space vehicles becomes non-Gaussian in a high dynamic environment, a flicker noise model is used to simulate the non-Gaussian characteristics of measurement noise. Then, a high dynamic loop tracking method based on Huber Cubature Kalman Filter (Huber CKF) and an adaptive deeply integrated navigation model based on the altitude angle adaptive fading factor are presented on the basis of the Beidou/INS deeply integrated model. First, a flicker noise model is developed to simulate the effect of ionospheric flicker on the carrier phase in high-speed environments. Second, a robust filtering method is used to replace the phase-locked loop, which overcomes the problem that the accuracy of the phase discriminator is reduced under non-Gaussian noise conditions. Finally, an adaptive fading factor is designed based on the satellite altitude angle to adjust the weights of each channel, which can effectively improve the position accuracy. The simulation results show that the estimation accuracy in the position Root Mean Squared Error (RMSE) can be improved by 0.288 4 m, and the estimation accuracy in velocity RMSE can be improved by 0.018 7 m/s, when the measurement accuracy indicator of the gyroscope is 0.01 (°)/h and the measurement accuracy indicator of the accelerometer is 10^-5 g. The methods presented in this paper can effectively improve the navigation accuracy of the near-space vehicle, and provide theoretical references for the navigation system of the near-space vehicle in the future.

Key words: satellite navigation, integrated navigation, robust filter algorithm, adaptive filter algorithm, flicker noise