Research On The Performance Evaluation Method Of Information Fusion System And Simulation Of Distributed Fusion System

Multi-source information fusion system,which adopts accurate and effective rules to deal with information from different sources,gives a unified judgment and estimation results,and can be applied to many military and civil areas.How to evaluate the performance of information fusion systems,and feed results back to information fusion systems to optimize the design of information fusion systems,is one of key problems to be solved.According to this problem,the research of the thesis is done to improve performance of distant early warning system.The performance evaluation method of information fusion system is first proposed,and the improvement of the distributed simulation system of distant early warning system which has been built is then done.The main research work is summarized as followed:Firstly,research background and significance of the thesis are introduced,and evaluation index and method of information fusion system and simulation technology of distributed information fusion system at home and abroad are reviewed.Secondly,description method of battlefield information quality and metadata model are introduced.After combining the relevant domain knowledge,the evaluation indexes of information fusion system are summarized and analyzed on this basis.In order to evaluate performance of the information fusion system based on the qualitative and quantitative indexes,a new method which combines cloud theory and fuzzy model is proposed.The complex evaluation process is first divided into several levels using analytic hierarchy process method.The reasonable transformation of the qualitative indexes is then realized using cloud theory,and the model of quantitative indexes is built using fuzzy model.The evaluation results of cloud theory and fuzzy model are combined to realize the evaluation of information fusion system.Simulation results show that,this method can realize the comprehensive evaluation of qualitative and quantitative index and adjust the relative weights of various indexes and thus is more applicable.Thirdly,the function of the distant early warning simulation system having been built is improved: the distant early warning simulation system which can emulate single target is extended to realize the simulation of multi-target(ballistic missile,stealth aircraft,near space target).The dynamic configuration of the target,the number and the parameters of the sensors are accomplished.Then,the function simulation of the sensor management is achieved,and the function verification of the sensor management algorithm is realized.Finally,the algorithm integration of the simulation center module and sensor management module is introduced,which provides a platform support for the functional testing of the related algorithms.Finally,the main work and further research of the thesis is summarized.