存在目标交叉情形的扩展目标跟踪算法

发布时间：2018-05-18 04:39

  本文选题：扩展目标跟踪 + 概率假设密度滤波　； 参考：《系统仿真学报》2017年01期

【摘要】：扩展目标跟踪过程中,若出现目标交叉,直接采用扩展目标高斯混合概率假设密度滤波算法会出现目标漏估计。针对该问题,提出了一种改进算法。计算每一时刻跟踪到的目标间的欧式距离,以此判定目标是否处于临近区域。在下一时刻,若临近区域内跟踪到的目标数目突然变少,则对临近区域内目标对应的高斯分量权值进行补偿;否则看作是正常的目标消亡现象,不作处理。使用处理后的高斯分量进行目标估计和跟踪。改进算法解决了因量测集分布紧密而被划分到同一个子集带来的目标数目漏估计的问题。仿真实验结果表明了改进算法的精确性与有效性。
[Abstract]:In the process of extended target tracking, if the target cross occurs, the extended target Gao Si hybrid probability assumption density filtering algorithm will be used to estimate the target missing. To solve this problem, an improved algorithm is proposed. The Euclidean distance between the targets tracked at each time is calculated to determine whether the target is in an adjacent area. At the next moment, if the number of targets tracked in the adjacent region suddenly becomes smaller, the weight of the Gao Si component corresponding to the target in the adjacent region is compensated; otherwise, it is regarded as a normal phenomenon of target extinction and is not dealt with. The processed Gao Si component is used for target estimation and tracking. The improved algorithm solves the problem of missing estimation of the number of targets which is divided into the same subset due to the tight distribution of the measurement set. Simulation results show the accuracy and effectiveness of the improved algorithm.
【作者单位】： 西安工程大学计算机科学学院;
【基金】：国家自然科学基金(61201118) 中国博士后科学基金(2103M532020) 陕西省自然科学基础研究计划项目(2016JM6030)
【分类号】：TN713
，

本文编号：1904430