Mutli-sensor Target Identification based on Evidence Theory
Zhuang Miao · Fire Control and Command Control · 2010
When the information of multi-sensors highly conflict in target identification system,the information fusion result will be unreasonable by DS combination rule.In order to solve the problem,evidence expectation approach is proposed.When several pieces of evidences are combined,the support degree for one another can be calculated according to the conflict information.Normal eigenvector for the maximal eigenvalue of evidence support degree matrix is considered to be weight coefficient.Then,expectation evidence can be gained,and it is combined by DS rule multiple times.Finally,several ways are compared and analyzed through numeric simulation,and the result suggests evidence expectation approach can still get ideal fusion result when the information of multi sensors highly conflict.