一种融合高斯聚类的粒子滤波方法

A Particle Filter Method with Gaussian Clustering

为了解决粒子滤波存在粒子退化、噪声干扰和计算量大等问题,提出了一种融合高斯聚类的粒子滤波方法。将给定粒子划分为许多类,采用使目标函数最小方法,把搜寻到的最佳隶属度和最佳粒子更新对应的位置作为聚类中心,有效地完成数据关联融合。实验证明,与传统粒子滤波算法相比,该方法提高了粒子滤波的运算速度,降低了计算量,具有更高的滤波效率和滤波精度。

This paper proposes a particle filter method with gaussian clustering in order to solve the problems of particle？ degeneracy and noise interference existing in particle filter, and large amount of calculation in the particle filtering algorithm before. The given particles are divided into many categories, and the minimum membership function of the objective function is used as the clustering center to locate the optimum particle size. Comprehensive experimental results demonstrate that the new algorithm improves the particle filter calculation speed, reduces the computational complexity of the traditional particle filter algorithm, and highly increases the filtering efficiency and accuracy.