关键词: 目标跟踪, 卡尔曼滤波, 转换量测, 条件均值和协方差, 性能分析

Abstract:

The converted algorithms are approximations of the optimal solution based on different assumptions,leading distinct performance in different applications,thus it is necessary to uniformly compare and analyze the performance of different converted algorithms.In this paper,the typical converted measurement algorithms are demonstrated.Their performances in different scenarios are evaluated by simulations in terms of tracking accuracy,credibility and computation complexity.The averaged normalized estimation error squared of the best linear unbiased estimate (BLUE) is approximately equals to 1,which is theoretically optimal in the sense of minimum mean square error with superior tracking performance to other filters.

Key words: target tracking;Kaman filter;converted measurement;conditional mean and covariance;performance analysis