基于随机有限集的多拓展目标跟踪算法研究

项目名称： 基于随机有限集的多拓展目标跟踪算法研究

项目编号： No.61304261

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 自动化技术、计算机技术

项目作者： 闫小喜

作者单位： 江苏大学

项目金额： 24万元

中文摘要： 随着现代雷达分辨率的不断提高，每个目标已经能够具有多个散射中心；体现在目标量测上，每个目标在每个时刻产生多个量测而非单个；此种情况下的目标跟踪称为拓展目标跟踪，它是当今目标跟踪领域新的研究热点。基于每个目标在一个时刻至多产生一个量测假设的传统多目标跟踪算法，已经不再适用于多拓展目标跟踪。本课题在多目标Bayes滤波理论框架内，利用随机有限集解决多拓展目标跟踪问题，主要解决多拓展目标运动特性与量测似然建模、多拓展目标概率假设密度滤波器推导及多拓展目标概率假设密度滤波器实现中的目标数目与状态联合估计问题。本课题的目标是，开发一种基于随机有限集的多拓展目标跟踪算法以充分利用高分辨率所产生的多目标量测信息。

中文关键词： 拓展目标跟踪；随机有限集；量测集分割；分量删减；强度估计

英文摘要： As the resolution capabilities of modern radars are increasing, there are several scattering centers existing in each target. It indicates on target measurement that several measurements rather than single are produced by one target at each time. The target tracking under this case is called extended object tracking, which is the novel research spot in the field of target tracking. The traditional multiple target tracking algorithms, which are based on the assumption that each target at most produces one measurement at each time, are not appropriate for multiple extended objects tracking yet. The problem of multiple extended objects tracking is solved by the tools of random finite set under the framework of multiple targets Bayes filtering in this topic. The topic is made up of three parts, the modelling of motion characters and measurement likelihood of multiple extended objects, the derivation of probability hypothesis density filter of multiple extended objects, and the development of combined estimation algorithm of target number and target states in the implementation of probability hypothesis density filter of multiple extended objects. The goal of this topic is to develop the multiple extended objects tracking algorithm based on random finite to make full use of the information of multiple target measurem

英文关键词： extended object tracking；random finite set；measurement set partition；component pruning；intensity estimation