Wearable Network Data Fusion Algorithm Based On Evidence Theory And Its Application

With the development of science and technology,more and more wearable network devices are becoming the part of people's life.The data of various wearable sensors is mainly divided into different categories such as motion data,environmental data,and biological data.Moreover,the sensors' data in the same wearable network device is also different with each other,so it is necessary to select an appropriate data fusion algorithm to process the data properly.Evidence theory is an effective method of dealing with uncertain information,which can be employed the fuse of multi-source heterogeneous data from different sensors,Evidence theory has been widely applied in various fields.This thesis aims at the problems in the evidence theory and that in the wearable network data fusion algorithms which are designed based on evidence theory.The main contributions of this paper are summarized as follows:(1)The thesis summarizes the development history of the wearable network,introduces the theoretical basis and research status of the evidence theory in detail,and sets forth from the formula meaning and the main problems in the present stage.(2)The thesis analyzes the results of anti-intuitional error caused by conflict evidence in the evidence theory,this paper proposes a data fusion algorithm based on belief interva lnumber distance of basic belief function.In this paper,the method is different from the modification of the evidence combination formula,The algorithm is proposed to modify the basic belief function generated by the evidence source.It is suggested that the distance between the bodies of two evidences is expressed by the confidence interval distance obtained by the distribution of the basic belief,and the weights of all the evidence bodies are given according to the distance between the evidence bodies.The basic belief function is modified by using the weight to solve the problem of anti-intuitionistic error.In order to verify the feasibility and effectiveness of this algorithm,the proposed method is compared with the traditional method of evidence combination and the correction evidence combination.(3)The wearable network data fusion algorithm is studied,a data processing algorithm based on evidence theory is proposed and applied to the human body balance ability analysis model.First,this paper ranks the fall risks of the human body to different risk grades,and constructs basic belief function of the sensor data,by using membership function and simple supporting evidence algorithm.Then,based on belief interval number distance of basic belief function in this paper,perform data fusion on conflicting basic belief functions.As a result,according to the decision making threshold,this paper presents the fall risks of object and gives the corresponding improvement opinions for the existing problem.The data fusion algorithm based on evidence theory is compared with other data fusion algorithms to verify the advantage of this algorithm in dealing with uncertain data.