基于一步异相关法则的kalman滤波抗野值方法

Study on Restraining Outliers of Kalman Filter Based on One-Step Singular Correlation Law

分析了野值对kalman滤波的影响机理;从随机序列的自相关函数出发,利用序列的一步差分方差判别野值的一步异相关法则,提出基于一步异相关法则的kalman滤波新方法,并与传统的改进方法进行了仿真比较.仿真实验表明,该方法简单易行,可以有效抑制观测值中野值对kalman滤波的不利影响,提高滤波精度和稳定性.

This paper analyses the mechanism that the outliers affect Kalman filters. According to auto-correlation function of random series and one-step difference variance for outliers identification,a new method based on one-step singular correlation law for Kalman filters is proposed, and compared by simulations with the traditional improved methods. The results show that this method is simple and easy, and it can effectively restrain the bad effects from the outliers in observation values during filtering and improve the filtering accuracy and stability.