随机摄动强跟踪粒子滤波算法

Study on stochastic perturbation strong tracking particle filter

如何解决粒子的退化问题和提高算法对突变状态的跟踪能力,是粒子滤波算法研究和应用中需要考虑的两个主要因素.传统的再采样算法虽然可以解决退化问题,但是容易导致粒子耗尽;扩展粒子滤波算法虽然可在一定程度上解决粒子耗尽问题,但其对突变状态的跟踪能力却不近人意;强跟踪粒子滤波算法可以提高对突变状态的跟踪能力,但却未能较好地改善粒子退化问题.针对上述问题,本文将随机摄动再采样方法引入强跟踪粒子滤波算法,提出了一种随机摄动强跟踪粒子滤波算法.当粒子退化问题严重时,对权值最大的粒子迭加随机摄动,用摄动粒子替换退化粒子以解决粒子退化问题,同时由于摄动粒子的加入增加了粒子集的多样性,可在一定程度上缓解粒子耗尽问题,提高算法对突变状态的跟踪能力.利用标准验证模型和分时恒定系统对所提出的算法进行了仿真验证,仿真结果证明了该算法的可行性和有效性.

To solve the degeneracy phenomenon and to improve the ability for tracking the breaking states are two difficult problems in the application of particle filter. Sequential important re-sampling can reduce orilliminate degeneracy, but the sample impoverishment is a secondary result. Extended particle filter can also reduce the degeneracy, but it cannot track the breaking states. The ability to track the breaking states can be improved by a strong tracking particle filter, but the degeneracy phenomenon will not be well solved still. A stochastic perturbation strong tracking particle filter is proposed for solving the above problems, in which a stochastically perturbative re-sampling is introduced into a strong tracking particle filter. Thus a stochastic perturbation is added to the particle with maximal weight to form some new particles, and the degenerative particles are displaced by the new particles to solve the degeneracy phenomenon and so the sample impoverishment improves the diversity of the samples. The ability of the proposed algorithm to track breaking states is also improved, and the feasibility and validity of the proposed algorithm are demonstrated by the simulation results of the standard validation model and the system with constants in different periods of time.