一种密集多回波环境下的机动多目标跟踪起始算法

A TRACKING INITIATION ALGORITHM OF MULTIPLE MANEUVERING TARGETS IN A DENSE MULTI-RETURN ENVIRONMENT

本文研究密集多回波环境下的机动多目标跟踪起始问题。文中首先提出主跟踪子空间和边缘跟踪子空间的定义与性质,接着修正Bayes轨迹确定方法BTC,并将其与具有残差滤波的修正概率数据关联滤波算法MPDAF-RF有机地结合起来,提出一种适合高密集多回波环境的机动多目环跟踪起始方法——“全邻”Bayes跟踪起始算法ABTI。Monte Carlo仿真表明,本文所给出的算法不仅克服了一类概率数据关联滤波方法没有跟踪起始机理的缺陷,而且辨别目标与虚警的能力很强,不失为解决高密集多回波环境下机动多目标跟踪起始的有效方法。

In tnis paper, a tracking initiation problem of multiple maneuvering targets in a dense multi-return environment is investigated.First, the definitions and properties of main tracking subspace and margin tracking subspace are studied.Then, an algorithm of all-neighbors. Bayes tracking initiation suitd to higher dense multi-return environment is presented.Monte Carlo simulation shows that the algorithm overcomes the shortcoming of modified probabilistic data association filter without mechanism of tracking initiation, improves the ability of tracker differentiating between targets and false alarms greatly, and is an eflective tracking initiation method of multiple maneuvering targets in higher dense multi-return environment.