Kalman滤波器的抗干扰算法

The anti-jamming algorithm of Kalman filter

在Kalman滤波意义下,提出将观测值划分为正常和病态两种属性,病态值可以是通常的野值,也可以是部分的正常值,这种模糊划分降低了辨识的难度,也比较符合客观实际,避免了由于划分不当所造成的误判或漏判情况的出现;为了提高Kalman滤波器的抗干扰能力,通过对新息分量统计特性的逐一修正,提出了重新构造状态估计的改进算法.由于充分利用了新息各分量服从正态分布的统计特性,所以算法具有结构简单,计算量小的特点.仿真实验表明,不论出现何种形式的干扰,算法都具有稳定的滤波性能,不仅克服了经典Kalman滤波器易受干扰,滤波效果不稳定的问题,而且克服了众多Kalman滤波改进算法的不足之处,适用范围更为广泛.

In the sense of the Kalman filter, a method in which samples may be divided into two groups： those for normality and those for ill-conditioned value, is presented. The ill-conditioned value that can usually be the boundary value or can also be a part of the normal. This fuzzy method reduces the identification difficulty and fits actual case. This method avoids ill-judged or missing case coming from improper classification. In order to enhance anti-jamming ability, the im- proved algorithm with reconstructing state estimation by modifying the statistical characteristic of innovation component one by one is presented. Because the statistical property of obeying normal distribution in each innovation component is taken into account, the framework of algorithm is rather simple and the algorithm can run fast. The simulation experiment shows that no matter what form of interference will be, the filter with this algorithm will be good at stable performance. It is not only overcoming the defect of classical Kalman filter in anti-jamming and stability, but also overcoming the defect of improved algorithms. It will be applied more widely.