Researching Multi-sensor Information Fusion Using D-S Theory And Fuzzy Set Theoy

Information fusion technology stems from the early period of military applications, this technology can process and handle incomplete information obtained form different information sources, and make sure these information can be complementary between their superiority, thus weaken the uncertainty included in the information sources, so which make a consistent description process that relatively complete to system environment. This description process can support and help people to make a decision, and perceive overall information connotation in further, finally completes accurately predict to the situation or behavior.By analyzing the base theory and several typical methods of multi-sensor information fusion, we especially study the information fusion method of the Dempster-Shafe evidence theory and the Fuzzy set theory.For resolving the conflict of Dempster-Shafe evidence theory, this paper introduced the concept of information capacity provided evidence. In the first place, we obtained the size of evidence information capacity for each, and implemented the normalization processing. Then we combined with the principle of the probability that supported evidence confliction, which weighted distributed by the average level of support, and obtained the average support rating, ultimately we proposed an improving composite formula of D-S.On the basis of the full study of Fuzzy set theory, by dint of the Fuzzy set similarity matrix, we created support rating of evidence and the reliability function, and obtained the weight coefficient of each evidence and the value of the basic assignment function, then we made the appropriate changes in model structure of data conflict, and combined with the basic combination rules of D-S evidence theory,finally we proposed a multi-sensor information fusion weighted average algorithm base on fuzzy similarity matrix, meanwhile, through combining with the Fuzzy set theory , we proposed a multi-sensor information fusion method of fuzzy comprehensive evaluation.At last we analyzed and verified the validity of information fusion application process of D-S evidence theory in the context of sensor information fusion of multi-sensory intelligent robot.