机动目标跟踪典型算法评述

Review of Typical Maneuvering Target Tracking Algorithms

机动目标跟踪算法可以划分为两种:极大似然法和贝叶斯估计法,前者通过对目标量测序列的分析,采用与当前目标运动模式匹配度最高(极大似然性)的模型(包括系统模型和观测模型)对目标状态进行跟踪和估计;后者则首先假定一组模型,再根据量测序列进一步估计出其中每个模型的匹配概率,最后采用贝叶斯公式对全部或部分模型估计进行综合。在此框架下,论文对其中的一些典型算法进行总结和分析。

The algorithms used in maneuvering target tracking can be classified as methods based on maximum likelihood estimation and methods based on Bayesian estimation. The former try to find out one model （including system model and measurement model） that matches the target motion mode best （maximum likelihood） by analyzing the measurements sequence, then run a single filter based on the model to estimate the target state and track. The latter first assume a set of models as possible candidates of the true mode in effect at the time, then calculate the model probabilities based on the measurcments sequence, at last generate the overall estimations using results of all filters or part of them by Bayes＇ hale. In this framework, this paper reviews. and analyses some representative algorithms for maneuvering target tracking.