首页 | 本学科首页   官方微博 | 高级检索  
     

高斯混合概率假设密度滤波器在多目标跟踪中的应用
引用本文:吕学斌,周群彪,陈正茂,熊运余,蔡葵.高斯混合概率假设密度滤波器在多目标跟踪中的应用[J].计算机学报,2012,35(2):2397-2404.
作者姓名:吕学斌  周群彪  陈正茂  熊运余  蔡葵
作者单位:四川大学计算机学院视觉合成图形图像技术国防重点学科实验室 成都 610064
基金项目:本课题得到国家"八六三"高技术研究发展计划项目基金,国家自然科学基金
摘    要:实现了基于随机集和点过程理论在目标数未知或随时间变化的多目标跟踪滤波算法.研究成果包括:(1)分析了基于随机有限集的多目标跟踪模型;(2)分析推导了基于随机集和点过程理论的概率假设密度滤波递推表达式;(3)实现了在线性高斯条件下的概率假设密度滤波的一种解析滤波算法;(4)仿真实验验证了算法的性能,比较了在杂波强度和检测概率变化的情况下和联合概率数据互联算法相关性能;(5)指出了算法的一些不足以及改进的研究方向.

关 键 词:高斯混合概率假设密度(PHD)滤波器  概率假设密度滤波器  随机集  多目标跟踪  联合概率数据互联

The Gaussian Mixture Probability Hypothesis Density Filter and Its Application to Multi-Target Tracking
LV Xue-Bin , ZHOU Qun-Biao , CHEN Zheng-Mao , XIONG Yun-Yu , CAI Kui.The Gaussian Mixture Probability Hypothesis Density Filter and Its Application to Multi-Target Tracking[J].Chinese Journal of Computers,2012,35(2):2397-2404.
Authors:LV Xue-Bin  ZHOU Qun-Biao  CHEN Zheng-Mao  XIONG Yun-Yu  CAI Kui
Affiliation:LV Xue-Bin ZHOU Qun-Biao CHEN Zheng-Mao XIONG Yun-Yu CAI Kui(College of Computer,Key Laboratory of Foundamental Science for National Defense,Sichuan University,Chengdu 610064)
Abstract:A algorithm based on random sets and point process theory is proposed for jointly estimate the time-varying number of targets and their states.The main contributions include:(1) Analyze multi-target tracking model based on random finite sets;(2) The Probability Hypothesis Density recursive formulas are deduced based on random sets and point process theory;(3) A analytic implementation of the Probability Hypothesis Density Filter is proposed under the linear Gaussian assumptions;(4) Two simulation results validate GMPHD performance and then compare GMPHD and JPDA performance under clutter and detection probability change;(5) Point out some the algorithm’s lack and research direction.
Keywords:GMPHD  probability hypothesis density filter  random sets  multi-target tracking  JPDA
本文献已被 CNKI 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号