组合式快速JPDA算法

基于有序状态空间搜索方法,定义联合事件为问题节点,将多目标数据关联问题求解转化为状态空间问题求解。定义联合数据关联概率函数对数表达为节点的估价函数,减少扩展节点数,迅速产生和搜索Ｎ个最大联合概率事件。给出快速计算Ｎ个最大联合概率事件的ＪＰＤＡ公式,并提出一种确定新生节点是否是已生成节点的简便方法。算法的搜索次数不随回波数增加而增加,有效地解决了传统ＪＰＤＡ算法的实时性问题,兼有贝叶斯方法和非贝叶斯方法的各自优点。

COMBINATORIAL QUICK JPDA ALGORITHM

The JPDA algorithm has excellent performance on tracking multiple targets in dense clutter environment. But its complexity will increase rapidly with the nurnber of targets and re-turns,which makes it almost impossible to be realized on realtime,In this paper,a graph search-ing algorithm which can search out N events with the largest joint probabilities is presented and an approximate bayesian estimation formula with these N events is then developed.In addition,an approach to the calculation of valuation function and repetition is also put forward.In this way,we form a new quick algorithm.The algorithm is of the performance equivalent to JPDA and its comoutation can be reduced drastically.