Application Of Fuzzy Neural Network In Multi-source Information Fusion

The multi-source information fusion is also called sensor information fusion,it is the integration of multiple channels and multiple aspects of the local environment of incomplete information to be integrated,The process of eliminating the redundant and contradictory information in the multi-source information,which is complementing each other and reducing its uncertainty,so as to form a relatively complete and consistent description of the system environment,In order to improving the speed and correctness of the decision-making,planning,reflection and reduce the risk of decision-making.In this paper,fuzzy neural network is applied to multi-sensor information fusion,Compared with the traditional fusion method based on probability theory,it does not need any priori information,It overcomes the defects that the evidence is difficult to obtain and the calculation is large.Not only broadens the ability of neural network information processing,so that it can not only deal with accurate information,but also can deal with imprecise information and fuzzy information,it will be contained in the fuzzy inference mechanism of neural network structure;moreover,the trained fuzzy neural network without additional information can solve the multi sensor information fusion,function extraction the theory of fuzzy rules and fuzzy fusion of multi sensor information in determining the system fusion,improving the ability and accuracy of fusion.Secondly,on the basis of the data fusion of T-S fuzzy neural network,the algorithm and network structure are improved based on the time and space complexity of the standard T-S fuzzy fusion algorithm.The rationality,stability and accuracy of the improved algorithm are verified by simulation experiments.Through the Kohonen network clustering algorithm to select the appropriate number of membership function and the experimental verification of the clustering algorithm in the T-S fuzzy neural network data fusion algorithm used in therationality and effectiveness.Then,based on the fuzzy neural network,the recognition of radar target echo is studied and tested,and two practical filtering algorithms are introduced to optimize the radar echo data.The simulation results show that the filtering effect is very significant.Finally,the analysis method for feature extraction of radar target echo data with the field collection of principal components.The experimental results show that this method has high accuracy and good generalization performance.