粒子滤波自适应部分系统重采样算法研究*

样本退化是基于序列重要性采样的粒子滤波中的一个主要问题，为了解决这个问题重采样被引入。常规的重采样算法可以解决样本退化问题，但容易导致样本衰竭，增加计算的复杂度。本文在部分重采样的基础上，提出了自适应部分系统分重采样算法，该算法自适应调整重采样的时间，重采样前按照粒子的权值对其分类，只对少数粒子进行重采样，不仅减少了重采样的时间而且增加了粒子的多样性，仿真结果表明该算法与部分重采样相比有效的提高了粒子滤波的性能，减少了运行的时间。

Study on adaptive partial systematic resampling algorithms of particle filter

Sample degeneracy is a major problem of particle filter which is based on the sequential importance sampling. In order to solve this problem, the resampling algorithm is introduced in particle filter. Regular resampling algorithm can solve the sample degradation, but it easily lead to sample depletion and increase the computing complexity In this paper, adaptive partial systematic resampling algorithm is proposed based on the partial resampling.The adaptive partial systematic resampling algorithm adjusts the resampling time adaptively, before the resampling, the particles are classified according to the weight, resampling is carries on the minority particles, so it not only reduces the resampling time and increases the diversity of particle. The simulation result indicates that compared with partial resampling it increase the particle filter performance and reduces the computation time.