改进的非线性鲁棒EKF算法及其应用

Modified nonlinear robust EKF algorithm and its applications

非线性系统由于噪声统计的不确定性,经常使得常规扩展Kalman滤波器滤波误差放大,甚至引起滤波发散,为此提出了一种采用虚拟噪声的补偿方法,对时变的噪声统计进行在线估计且加以补偿,提高了非线性滤波的鲁棒性。以星载GPS定轨为研究对象,应用该方法进行滤波定轨,结果表明,方法具有滤波稳定性和定轨结果的精确性。

Because of the uncertain noise statistics in nonlinear system,the filter error of the classical EKF（Extended Kalman Filter） will be enlarged even lead to filter divergence.To overcome this limitation,a novel method is proposed,it uses the method of fictitious noise compensating,and time-varying noise statistics can be estimated and compensated on line,so it improves the robust of nonlinear filter.At last,the proposed method is used in space-borne GPS orbit determination,the result shows that the novel method has better stability of filter and accuracy of orbit determination.