摘要
目标识别系统中所获得的信息常常是高度冲突和不确定的。基于DSmT理论的多传感器目标识别,可以解决证据高度冲突情况下的信息融合问题。然而由于DSmT理论融合结果分类精细而不利于判决,需要将某些分类结果进行重新分配。层次分析法(AHP)包含不确定知识矩阵,生成基本信度分配函数。基于此提出AHP-PCR5方法进行证据高度冲突情况下多传感器综合目标识别,不仅提高识别精度,降低识别过程的不确定因素,同时引入折扣系数灵活处理识别过程中具有不同可信度的多传感器融合问题。
Information obtained from target identification system is highly conflicting and uncertain. Multi-sensor target identification based on DSmT can solve this problem. But it is difficult to make decision,because there are too many categories arranged by DSm T theory,and it is necessary to redistribute the coefficient. Analytic Hierarchy Process( AHP) method uses the uncertain knowledge matrices to generate basic belief assignments( bba's). Multi-sensor target identification by AHPPCR5 method is produced to improve identification reliability and reduce the uncertainty in highly conflicting evidence situation. And the discounting coefficient is introduced to solve the fusion problem when multi-sensor has different reliability.
出处
《指挥控制与仿真》
2015年第1期64-67,共4页
Command Control & Simulation
基金
国家自然科学基金(11371052)