摘要

Multi-target tracking system is used to distinguish the class from different targets and obtain the trajectory of all targets from multiple information systems. As a significant and novel information fusion method, evidence theory has been used in multi-target tracking system. The evidence combination rule which is the core of evidence theory is turning into a new research emphasis. Yager combination rule which defined in DST is an effective solution. PCR5 and PCR6 are proposed in DSmT which is an extension of DST. But because of the huge computation problem, the use of PCR5 and PCR6 in information fusion is restricted especially in real-time environment. In order to overcome the huge computation drawback of PCR5/6, a novel evidence combination rule, mixture combination rule (MCR) is proposed in this paper. MCR combines the advantages of PCR5/6 and YGR and switches its combination rule between PCR5/6 and YGR after making a judgment of decision stability. The simulation test on MCR is also done and the test results show that the computation load of MCR is largely reduced compared with PCR5/6.

全文