基于施蒂费尔流形和线性优化重采样的粒子滤波器

为了解决粒子退化问题,提出一种基于施蒂费尔流形和线性优化重采样的粒子滤波算法.将系统模型置于施蒂费尔流形之上,采用朗之万分布描述过程转移概率分布,用矩阵正态分布表示似然函数分布,从而得到一种较为通用的重要性概率密度函数选择方法;同时,将重采样中抛弃的粒子与复制的粒子按照一定的线性组合方式产生新粒子.仿真结果表明.该算法具有较高的滤波效率、滤波精度和较强的滤波鲁棒性.

Particle filtering based on Stiefel manifold and linear optimizing resampling

In order to resolve the problem of particle degeneration, a particle filter based on Stiefel manifold and linear
optimizing resampling is proposed. System model is placed on Stiefel manifold, and Langevin distribution is used as
prior density and matrix normal distribution as likelihood function, such that a relative universal importance probability
density function is acquired. Further more, new particles are generated through linearly combining the selected particles
and abandoned particles in resampling process. The experiment results show the proposed particle filter has better filtering
efficiency, precision and rubustness.