抗差EKF算法及其在GPS/BDS动态相对定位中的应用

ROBUST EKF AND ITS APPLICATION TO COMBINED GPS /BDS KINEMATIC RELATIVE POSITIONING

扩展Kalman滤波算法(etended Kalman filter,EKF)对观测值质量要求较高,当观测值带有粗差时,EKF算法收敛失真甚至发散。提出采用抗差估计与EKF相结合的抗差EKF方法处理GNSS动态定位中的粗差问题。结果表明,抗差EKF方法可以有效抑制粗差观测值的影响,其定位精度相对标准EKF方法改善率优于70%。

Extended Kalman filter （EKF） algorithm is widely used in all kinds of nonlinear systems. However, EKF algorithm has high requirement to the quality of GNSS observation data. If observation data contains some gross errors, the convergence of EKF algorithm will be distorted. A combined method of robust estimation and EKF is proposed to deal with the gross errors in GNSS kinematic positioning. According to the test data, the new method can effectively control the influence of the gross errors and improve accuracy as 70% or more.