基于高斯混合概率假设密度滤波器的扩展目标跟踪算法  

Extended targets tracking algorithm based on Gaussian-mixture probability hypothesis density filter

作　　者：曹倬[1] 冯新喜[1] 蒲磊[1] 张琳琳 CAO Zhuo FENG Xinxi PU Lei ZHANG Linlin(Information and Navigation Institute, Air Force Engineering University, Xi＇an 710077, China Air force Dalian Communications Noncommissioned Officers school, Dalian 116600, China)

机构地区：[1]空军工程大学信息与导航学院,陕西西安710077 [2]空军大连通信士官学校,辽宁大连116600

出　　处：《系统工程与电子技术》2017年第3期494-499,共6页Systems Engineering and Electronics

基　　金：国家自然科学基金(61571458)资助课题

摘　　要：针对杂波环境下多扩展目标跟踪中航迹起始和量测集划分问题,提出了一种基于高斯混合概率假设密度滤波器的扩展目标跟踪算法。在航迹起始阶段利用最近邻指数法对量测集进行聚类趋势分析,接着通过改进OPTICS(ordering points to identify the clustering structure)算法,建立一个增广数据集排序来表示量测集的密度结构,该算法对参数选择、初始点选择均不敏感,可以滤除量测集中的杂波。仿真结果表明,在航迹起始阶段本文所提算法在保证起始性能的同时计算代价明显减少,在量测集划分过程中,所提算法能够有效划分不同形状、密度的扩展目标,自适应地确定划分数目,减少算法运行时间。A multiple extended targets tracking algorithm based on Gaussian-mixture probability hypothesis density filter is proposed to solve the problems of track initiation and measurement partition of the extended target in the clutter environment. The proposed algorithm takes clustering tendency of measurement sets into account at the stage of track initiation. Then, the improved ordering points to identify the clustering structure （OPTICS） algorithm is used to extract the measurement cluster. Through the establishment of an augmented data set sequencing to represent the measurement set density structure, input parameters and initial point selection are not sensitive to the choice. Besides, the algorithm can filter the measurement from clutter. Simulation results show that the proposed algorithm has a better computational efficiency over the traditional algorithm at the stage of track initiation. In the clustering process, the proposed algorithm can be used to partition the measurement sets of different densities and shapes, adaptively determine the number of partitions and increase the computational efficiency.

关 键 词：目标跟踪 多扩展目标 聚类趋势 量测集划分 

分 类 号：TN953[电子电信—信号与信息处理]