摘要
针对多传感器多目标跟踪中数据关联的计算复杂性问题,提出了一种多传感器序贯势分布概率假设密度滤波算法.利用序贯滤波的方法将单传感器的势分布概率假设密度滤波扩展到多传感器情况,并给出了高斯混合实现的序贯势分布概率假设密度(Gaussian mixture sequential PHD,GMSPHD)滤波的递推算法.仿真实验结果表明,文中提出的GMSCPHD滤波算法具有较高的多目标状态估计和目标数目估计精度,是一种有效的多传感器多目标跟踪方法.
In view to the complicated computation problem of data association for multisensor muhitarget track ing, a new mutlisensor cardinalized sequential probability hypothesis density filtering algorithm is proposed. In this algorithm, using the method of sequential filtering, the single sensor cardinalized probability hypothesis den sity filtering is extended to the multisensor condition. And the recursive algorithm of Gaussian mixture sequential cardinalized probability hypothesis density(GMSCPHD) filtering is given. Simulation results demonstrate that the proposed algorithm has significant improvement in estimating the multitarget state and target number, and it is an effective muhisensor multitarget tracking method.
出处
《江苏科技大学学报(自然科学版)》
CAS
2012年第6期587-592,共6页
Journal of Jiangsu University of Science and Technology:Natural Science Edition
基金
海军装备预研基金资助项目(2010401010202)
江苏科技大学博士科研启动基金资助项目(35031103)
关键词
势分布概率假设密度
多传感器多目标跟踪
随机有限集
高斯混合
ardinalized probability hypothesis density
multi-sensor multi-target tracking
random finite set
Gaussian mixture