摘要
基于视觉的无人机地面目标跟踪状态估计为非线性滤波问题,针对使用一般粒子滤波算法存在粒子退化和计算量大的缺陷问题,提出了一种基于排序的粒子滤波算法,对粒子依误差大小进行排序并计算粒子权重。仿真试验表明,该方法减轻了粒子贫化的影响,提高了状态估计精度。
Vision-based ground target tracking state estimation is a non-linear filter problem. For particle degeneracy and large amount of computation phenomenon exist in the standard particle filter algorithm, a sort-based particle filter algorithm is proposed. In this algorithm, all the particles are sorted by measurement update error during weighting update process and the weight of the particles is calculated based on sequence. Compared to standard particle filter, the sort-based particle filter has lightened particle degeneracy, reduced the amount of computation and improved the precision.
出处
《计算机工程与应用》
CSCD
2012年第15期21-23,28,共4页
Computer Engineering and Applications
基金
航空科学基金(No.20080896009)
关键词
目标跟踪状态估计
粒子滤波算法
排序
粒子权重
target tracking state estimation
particle filter algorithm
sort
weight of particles
