星载GPS低轨卫星定轨算法仿真研究

Simulation Study of Low-Earth-Orbit Determination Algorithm from Space-based GPS

在对Kalman滤波在星载GPS低轨卫星定轨的研究中,由于系统和观测方程均为非线性,在线性化过程中引入线性化误差造成滤波发散现象,以及由于截断误差使得定轨精度差的问题,因此将平方根UKF方法应用于天基GPS低轨卫星定轨中,为提高定轨精度、速度及定轨稳定性。利用UT变换良好的期望及方差传递特性,无须对系统方程及观测方程进行线性化,避免了传统的EKF算法由于线性化引入的误差,同时平方根滤波方法有效地克服了协方差阵非正定问题。仿真结果表明,在相同条件下比EKF有效地抑制了滤波发散,且提高了定轨精度。

SR_UKF was used in low-earth-orbit determination by space-based GPS in order to solve Kalman filtering divergence because of the problems of non-linear system equation and measurement equation,filter divergence due to linearization error,and the bad precision of orbit determination due to truncation error.Using this method can improve the precision and speed of orbit determination and stability.This approach adopts the favorable transfer characteristic in expectation and variance for unscented transformation,and needs not to be linear for the system and observation equation,thereby,can avoid importing error owing to conventional EKF algorithms.At the same time,the method of square-root filtering effectively overcomes the covariance matrix non-positive.The simulation experiment results show that the proposed method has efficiently restrained filtering divergence,and better positioning precise than EKF under the same condition.