一种基于重要性权重的证据合成方法及其应用

摘要

针对Dempster方法在合成高度冲突证据时结果与直观结论相悖的问题,提出了一种改进的基于重要性权重的证据合成方法,并将其应用到海上目标的识别.首先,根据证据之间的相似性程度判断证据间是否存在冲突,对于有冲突的情形,由基于P距离的权重分配模型得到各证据的重要性权值;然后,对BPA函数进行修正,采用Dempster规则修正后的BPA函数进行合成;最后,利用雷达和舰船的对应关系,将侦察雷达探测到的目标雷达信号载频、脉宽、重频转化为雷达参数的概率分配,并运用改进的证据合成方法进行了目标识别.仿真计算结果证明:该方法能有效处理冲突证据组合,克服了一票否决问题,减小了识别结果的不确定性,并保留了Dempster方法处理非冲突证据组合的良好特性.%To deal with the unrealistic results arising in using Dempster s method for the combination of highly conflicting evidences,an improved evidence combination method is presented which is based on evidence weight values.Firstly,the degree of similarity between evidences is depended on to judge whether conflicts exist.In the case of conflicts,a distance P can be used to calculate the distance between individual evidences so as to obtain the evidences' weight values.Then,BPA functions are modified before using Dempster rule for combination.This improved method can be applied to the recognition of sea targets.According to the correlation between radar and ship,it is possible to convert the radio frequency,pulse width,repeated frequency of the target radar signals detected by the reconnaissance radar to a probability assignment function of radar parameters,by which to recognize targets at sea.The simulation results show that this improved method can deal with conflicting evidences effectively,reduce more uncertainty in recognizion,and maintain the advantages of Dempster s method in tackling non-conflicting evidences.