基于信息扩散的Kalman滤波抗野值研究

Study for Outliers Based Information Diffusion with Kalman Filter

标准Kalman滤波在理想条件下是系统状态最优滤波,当观测值中含有野值时,将严重影响滤波的精度。采用新息扩散原理对滤波方程中的新息序列进行修正,不仅可以抑制野值的影响,而且不受野值分布情况的限制,仿真结果验证了它的有效性。

Traditional Kalman filtering is the robust estimation under desirable conditions,but it may be deep influenced by outliers appeared in observation. A method based on information diffusion revising the observation difference is proposed. It can not only resist the outliers but also be applicable to different distribution. Simulation results has validated this method.