Algorithm Research On Satelellite Autonomous Navigation Based On Synthetic Aperture Radar

With the development of the space technology, more and more spacecrafts were launched in the past years and the applications become more common. Therefore, the self-surviving performance of the spacecrafts is regarded as one of the key technology in the 21st century. The autonomous navigation technology is not only important aspect of spacecraft autonomy but also the trend of development of space technology. It has important meaning for the modern spacecraft to reduce the burden of the ground-assisted system, decrease the cost of maintenance, and enhance the survival ability in the war. This paper discussed a new method for the spacecraft's autonomous navigation, using the Synthetic Aperture Radar (SAR). The main contents of this paper as follows:SAR was usually used for earth observations; it can get the distance and the Doppler shift between the satellite and the landmarks. The new autonomous navigation method is based on the distance equations. This paper analyzed the principle of the method and established the observation equation which is based on the distance equations. And then it analyzed the main factors affect the observation equation, including the accuracy of the measurement, the error of the landmarks and the distance between the landmarks.In the condition that SAR can get serial observations, it design extended kalman filter, and analyzed the influence of the measurement accuracy, the error of the landmarks and the distance between the landmarks, and support the analysis mentioned before through simulation. EKF is based on the linearization, and ignore the high order parts. To overcome the defects of EKF, it studied another algorithm using Unscented Transformation (UKF), and made a comparison between EKF and UKF in the same simulation condition in this paper.In the interval condition that SAR can only get observation in the fixed area, such as in the inland of China. The interval between the first and the second observation may be several hours, which lead to the error of orbit determination reach to a peak. To solve the problem, this paper made a research on the kalman smoother, and design fixed interval smooth and the fixed point smooth based on EKF. Then simulation result and system performance is analyzed and discussed after numerical simulation.