A Method To Represent And Fuse The Heterogeneous Information Based On Random Sets Theory

Multi-source information fusion technology has been applied in many fields, but it is currently facing a prominent problem that the types of the information are different.The traditional methods of multi-source information fusion are lack of the mathematical basis when they are used to solve the representation and fusion problems of heterogeneous information, so it can not to do the analysis and evaluation effectively. In order to solve the problems of representing and fusing the heterogeneous information, a method based on random sets is proposed in this paper.On the basis of introducing the basic theories of random sets and the relationship of mutual transformation between random sets and traditional methods including D-S evidence theory and fuzzy sets, the method that solves the heterogeneous information problem in the field such as conditional monitoring and fault diagnosis is proposed in this paper. First, the concept of the global sensor is introduced into the case of using multi-sensor to monitor one of the factors that affect the condition; the random sets theory and the method of plausibility measure are used to obtain the value of the global sensor and the basic probability assignment of sensors data. Then, introduce the concept of experts'weight, and use the random sets to represent experts set and experts'opinions set; by some calculations of random sets, the basic probability assignment of experts' opinions can be obtained. Finally, Fuse the sensors data and the opinions of experts in the framework of random set, get the final result.In this paper, the simulation experiment and Matlab are used to process the data, and describe the representation and fusion process of the information provided by the sensors and the experts in detail. By the form of figures to analyze the experimental data, and the results shows that the method proposed in this paper can reduce the uncertainty range of sensor data and improve the accuracy of match results, so a more accurate basic probability assignment can be obtained. Besides, the information of experts is also represented by random sets, which makes it can be fused with the data of seasons and give a more accurate final result. The simulation experiment shows the validity of the method proposed in this paper.