多源冲突证据的智能融合算法研究

与Dempster-Shafter理论（DST）相比，Dezert-Smarandache理论（DSmT）通过保留证据冲突项作为数据融合的焦元，从而可以很好地解决在证据发生高冲突情况下的信息融合问题。但是因为DSmT算法增加了矛盾焦元，致使推理过程中的计算量加大，更容易产生焦元爆炸的问题。针对上述问题，提出一种结合两者优点的DST-DSmT智能算法。该算法以证据之间的冲突质量作为判断依据，当冲突质量较小时采用DST算法，反之则采用DSmT算法，以期在保证融合效果的情况下，减小计算量。以P2-DX机器人为实验平台，以具体算例验证了方法的正确性和有效性。

Intelligent fusion algorithm of multi-source conflicting evidences

Compared with DST,DSmT can keep the conflict of evidences as a focus element in data fusion in order to resolve the difficulty in the high conflict.But the DSmT's computation will oversize more easily because more focus elements are additional in DSmT rule.And the fusion result is worse than DST's when the low conflict situation is occurred.Aiming at it,this paper proposes one kind of intelligent algorithms on basis of DST and DSmT,combining their advantages.The algorithm uses conflict mass as a judgment in...