Research On Multi-sensor Information Fusion Technology Based On D-S Evidence Theory

In recent years,with the advancement of electronic information technology and the widespread application of sensors,multi-sensor information fusion technology has developed rapidly.As a key step in the field of multi-sensor information fusion,the processing of uncertain information has made many new developments driven by the needs of automation and intelligence,but it is also facing great challenges due to the higher requirements of the system for its real-time and reliability.Dempster-Shafer evidence theory provides powerful tool support for uncertain information processing research with its solid theoretical foundation.However,in practical applications,it also has problems such as the generation of Basic probability assignment(BPA),the applicability of Dempster combination rules,and the synthesis of conflicting evidence.Facing the demand of multi-sensor information fusion system to improve the processing capacity of uncertain information,this paper conducts research on multi-sensor information fusion algorithm based on D-S evidence theory.The main research contents are as follows:In the process of generating basic probability assignments,the model has poor universality,and the attribute weights cannot fully reflect the reliability of the system.Aiming at this problem,a BPA generation method based on the cloud model is studied.Firstly,a method for constructing a propositional model function of a sample attribute single subset based on the normal cloud model is proposed,and then the numerical characteristics of the composite subset model function are obtained.Secondly,the attribute weight is calculated according to the model overlap degree of the test sample under some single sub-focal elements,so as to take into account the reliability of the information source.Finally,use the attribute weight to modify the output result of the model function to obtain the BPA.In the same simulation scenario,the proposed method has higher recognition accuracy than traditional methods,and is suitable for situations with fewer samples.The method of correcting conflict evidence is not comprehensive and objective when combining the internal and external conflict information provided by the source of evidence.To solve this problem,a method of evidence revision based on Singular Value Decomposition and information entropy is studied.Firstly,a reconstruction technique based on singular value decomposition to extract the main direction corresponding subspace of the evidence matrix is proposed to filter out matrix interference information.Secondly,according to the number of main directions of the evidence matrix and the information entropy,the highly conflicting evidence is identified and corrected.Finally,the revised evidence information is fused based on evidence theory.Property analysis and numerical simulation show that the proposed method can effectively deal with conflicting evidence and improve the sensitivity of evidence theory to interference information and the decision-making ability of fusion results.