D-S证据理论在小型空地无人集群战场目标识别中的运用

Application of D-S Evidence Theory to Battlefield Target Recognition in Small Air-ground Unmanned Swarm

近年来,多传感器信息融合技术在无人集群战场目标识别等领域得到了广泛的关注,针对D-S证据理论在无人集群协同目标识别过程中潜在的冲突证据组合问题进行了研究。首先,提出了一种基于综合考虑目标识别主体可信性和证据间支持度的复合折扣系数计算方法,并引入了结果精炼机制;其次,对该方法在无人集群战场目标识别中的运用进行了探讨;最后,通过仿真算例验证了新方法能够克服低置信度、高冲突证据组合“一票否决”的问题,与其他几种典型的改进算法相比,尤其适用于冲突证据数目较多的情况。

In recent years,multi-sensor information fusion technology has received extensive attention in fields such as unmanned swarm battlefield target identification,and the problem of potential conflicting evidence combinations in the process of unmanned swarm collaborative target identification by D-S evidence theory is studied.Firstly,a compound discount coefficient calculation method based on comprehensive consideration of target identification subject credibility and inter-evidence support is proposed,and a result condensation mechanism is introduced;secondly,the application of the method in unmanned swarm battlefield target identification is discussed;finally,it is verified through simulation arithmetic that the new method can overcome the low confidence and high conflicting evidence combination“one vote”problem,and is particularly suitable for situations with a high number of conflicting evidence compared to several other typical improved algorithms.