Maximum Correntropy Kalman Filter Based Self Calibration of Airborne SINS Adopting Aided by Star Sensor

In order to strength the observability of velocity and attitude in the self-calibration process of airbrone strapdown inertial navigation system(SINS), and to improve the estimation accuracy of gyro draft and accelerometer zero bias, the attitude from star sensors and velocity from the GPS are introduced to calibrate in the strapdown inertial navigation system. Meanwhile the measurement noise of attitude and velocity becomes non-Gaussian distributed, and the statistical characteristics of noise is inaccurate, the attitude and velocity are affected by pulse noise and electromagnetic interference under the actual flight conditions, which leads to the degeneration of Kalman filter performance. In order to utilize the high-order moment information of measurement signals, the minimum mean square error criterion is replaced by maximum correntropy criterion in Kalman filter, which is used to calibrate the airbrone SINS aided by the star sensor and the GPS. The simulation results show that the accuracy of calibration is improved by the maximum correntropy Kalman filter. And the speed and accuracy in calibration are improved by the aide of the star sensor.

For the airborne self calibration of airborne SINS, in order to improve the calibration accuracy of gyroscope drift and accelerometer bias, combined with the observability analysis of the previous SINS error model, star sensors and GPS signals were used to assist in the airborne self calibration of airborne SINS. Due to the non Gaussian noise caused by factors such as aircraft maneuvers and body vibrations during aerial flight, the maximum correlation entropy Kalman filtering method was adopted to eliminate the impact of non Gaussian noise on self calibration accuracy. In the principle of maximum correlation entropy, the introduction of higher-order moment information can enhance the robustness of the filter, thereby significantly improving the self calibration accuracy and speed of airborne SINS errors. To verify the effectiveness of the proposed self calibration method, based on the SINS model and aircraft motion model, the aircraft maneuvering trajectory is designed to stimulate various SINS errors, and the maximum correlation entropy Kalman filtering algorithm is used to estimate the error terms. At the same time, a comparison was made with the error terms obtained from SINS aerial calibration without star sensor assistance. Through comparative analysis, it can be seen that with the assistance of star sensors, the maximum correlation entropy Kalman filtering algorithm has significantly improved the calibration accuracy of gyroscope drift and acceleration bias, especially for the drift of celestial gyroscopes. The calibration speed and accuracy have been improved. However, there is divergence in the height channel, so it is necessary to use other methods to improve the calibration accuracy of the vertical accelerometer bias. Meanwhile, when establishing the error model, this article only analyzes and discusses the constant value error, scale factor error, and installation error of inertial components, without in-depth analysis of temperature and high-order term errors of inertial components. In subsequent work, a more comprehensive error model can be established to further improve the calibration accuracy of error terms.