将秩滤波与粒子滤波相结合,提出秩粒子滤波。它首先由秩滤波方法得到重要性密度函数,然后再采用粒子滤波方法进行滤波。由于秩滤波产生的重要性密度函数包含了最新的观测信息,与真实状态概率密度的支集重叠部分更大,所以,秩粒子滤波比无迹粒子滤波和粒子滤波具有更高的精度,且计算简单,便于工程应用。大量仿真结果验证了这一结论。
Based on rank filter (RF) and particle filter (PF) , a rank particle filter (RPF) was presented. This method included the importance density function obtained from RF and the application of PF. Because of importance density function from RF containing the latest measurement information, its importance density function was more in line with the probability density of the truth state. So RPF has the higher accuracy than those of unscented particle filter (UPF) and PF, and it has been proved by lots of simulations. Furthermore, RPF is simple to calculate and easy to apply in engineering.