论文部分内容阅读
本文提出了将集合论描述法应用到快速神经联合概率数据互联（FNJPDA）中，该方法综合了二者的优点，便于工程实现。同时本文还提出当目标数和杂波密度相当大时，将跟踪空间分成若干子空间，在每一个空间中应用该方法，以降低神经网络的维数，进一步提高运算速度，有效地完成多目标跟踪。
In this paper, we propose to apply Set Theory to FNJPDA, which combines the advantages of both and makes it easy for engineering implementation. At the same time, this paper also proposes that when the target number and the clutter density are quite large, the tracking space is divided into several subspaces, and the method is applied in each space to reduce the dimension of the neural network, to further improve the computing speed and to effectively accomplish the multi-target track.