论文部分内容阅读
在实际测量中,由于野值的影响,Kalman滤波新息的特性遭到破坏,滤波不再准确甚至发散。针对此一现象,提出了基于“新息”序列和时间序列预测联合修正的方法对测量数据进行处理的新算法。该算法运用“新息”序列进行野值点判别,利用时间序列观测的方法对野值点处的“新息”进行修复。仿真证明,该算法可使状态估计与野值点判别同时进行,并能很好的抑制滤波的发散。
In actual measurement, due to the influence of outliers, the characteristics of Kalman filter are destroyed, and the filtering is no longer accurate or even divergent. In response to this phenomenon, a new algorithm based on “new interest ” sequence and time series prediction combined correction method to process the measurement data is proposed. The algorithm uses “new interest ” sequence to discriminate outliers, and uses the method of time series observation to repair the “new interest ” in the outliers. The simulation results show that the proposed algorithm can make the state estimation coincide with the outlier detection and suppress the filtering divergence well.