基于证据理论的交互式目标识别算法

在充分利用信息处理中心的人工情报和其它信息源数据的基础上,提出了基于证据理论的交互式二级融合算法。对数据融合系统获得的不同信息源数据进行了详细分类,并阐明了非实时信息和半实时信息在数据融合前期及中期的不同作用。首先在多传感器数据时域融合的基础上,获得对目标的初步判定值;其次在信任加权的基础上接收其它识别单元对目标的判别概率,通过信息处理中心传来的目标数据,对不同识别单元的目标数据融合的结果进行修正,解决了问题规模扩大时计算量快速增长的问题。仿真结果证明,该算法能有效实现对目标的快速、准确性判别

Interacting target identification algorithm of data fusion based on evidence theory

An interactive algorithm of two-level data fusion based on evidence theory is presented. Different information-source data received by the data fusion system are classified in detail. Different functions of non-real-time information and semi-real-time information in the early and middle stages of the operation are studied. Firstly, on the basis of time-domain fusion of each sensor data, the preliminary judgement targets are fixed. Then using target data come-from information processing centre, the fused target data located at different identification units are revised. This method avoids the rapid increase of calculation scale with the growth of identification task. Simulation results illustrate that proposed method can identify targets efficaciously.