Kalman滤波新息正交性抗野值法研究

Study for outliers based on Kalman filtering

在Kalman滤波应用过程中,观测值中的野值是影响滤波效果的重要因素。当观测中含有野值时,破坏了Kalman滤波新息的正交性,从而造成估计不准,滤波精度下降。本文提出了修正Kalman滤波新息正交性的方法,使修正后的Kalman滤波新息能够保持修正前的新息正交性。仿真结果表明,本文提出的方法有效地抑制观测中的野值对系统滤波的不利影响。

In Kalman filtering applications the observation values including outliers are important effect on optimal filtering. The outliers affect the orthogonal property of Kalman filter innovation sequence, filtering accuracy and make estimation inaccurate. In this paper, the modifying orthogonal property of innovation sequence based on Robust Kalman filtering is presented, which can keep orthogonal properties of innovation sequence. Simulation results show that the modified algorithms are effectively resistant to outliers in sampling data.