基于数据融合的多目标无源跟踪的弹性网络方法

弹性网络被应用于多目标／多传感器无源识别和跟踪问题。在传感器仅能获得目标方位角的条件下，首先基于多传感器数据融合技术，分析了目标的聚类特征；然后借助期望模板提出了一种非全连接环形结构的弹性网络模型，分析了该模型识别和跟踪目标的动力学机制，讨论了网络应场内的ＷＴＡ竞争学习机制降低了网络的计算量，感应场的自适应性可以使神经元跳出局部极小点，提高目标识别率。

Elastic Net Method for Multi-Target Multi-Sensor Passive Trac king Based on Sensors

The elastic net is applied to solve the problem of mult i-target multi-sensor passive recognizing and tracking. At first the assemblag e features of the target are described, based on sensors fusion technique on con dition that the sensors detect only bearing angles. And then an elastic net mode l with unfully connected ring architecture is presented by using an expectation template and its dynamics for locking targets is analyzed. The elastic learn rat e and the role of the neurons expectation value and the noise coefficient in red ucing the net noise sensitivity are discussed. At last the simulational results show that WTA mechanism in the receptive field may reduce the computational cap acity a nd the neurons may escape poor local minimums through its adaptive receptive fie ld.