Application of Unscented Kalman filter in celestial/geomagnetic autonomous navigation algorithm of aerocraft

The low accuracy of geomagnetic model would causes navigation accuracy insufficient and induces limitations when adopting celestial navigation method based on the orbital dynamics equation. To overcome these, an aircraft autonomous navigation method using celestial/geomagnetic information as measurements is put forward. Meanwhile the dynamics equations are established, and the observation equation is derived. This method takes the angle between starlight and geomagnetism vector as the observation value and uses the UKF algorithm to estimate the position and velocity. The simulation results demonstrate that this algorithm's navigation accuracy is far better than that of the geomagnetic navigation. In addition, the filter convergence and stability are satisfactory, and the error does not accumulate with the time. In this light, this method is valuable for engineering application.