基于密度模糊C均值量测集划分的多扩展目标跟踪算法

摘要

针对多扩展目标跟踪中的量测集划分问题,提出一种自适应门密度模糊 C 均值(fuzzy C-means,FCM)量测集划分算法.首先,该算法利用自适应门对量测集进行预处理,滤除部分杂波量测,在降低算法计算复杂度的同时提高了量测集划分的准确度.然后,采用密度函数法产生 FCM的初始聚类中心,并给出自适应门密度 FCM多扩展目标量测集划分算法的详细过程.最后,采用容积卡尔曼滤波求解非线性框架下具有封闭解的扩展目标高斯混合概率假设密度(ET-GMPHD)滤波算法.仿真实验验证了算法的有效性.%Aimed at the problem of measurement set partition of multiple extended target,a measurement set partition algorithm is proposed by fuzzy C-means (FCM)with adaptive gate density.Then,initial clus-ter center of FCM is generated by density function,and the detailed process of multiple extended target measurement set partition algorithm for FCM with adaptive gate density is given out.Finally,using the cubature Kalman filter to solve the extended target Gaussian mixture probability hypothesis density(ET-GMPHD)with closed solution in the nonlinear framework.Simulation results verify the effectiveness of the algorithm.