Multi sensor data fusion method based on fuzzy neural network

Abstract

With the uncertainty of the multi sensor data of the fuzzy neural network fusion, the measure data from sensors is used to as the input of the fuzzy neural network and then is fuzzed. Next the data is analyzed and disposed by the neural network rule. Finally it is output after defuzzification. Confronting with the input fuzzification with uncertain membership function, we adopt the golden partition method to decide the initial center and width of membership functions of the fuzzification layer. The way of the model fuzzification and the improved BP network study rule is introduced to the network judging rule, and the judging result is output after defuzzification according to the weight rule The article gives a general method of the multi sensor data gaining based on fuzzy neural network. The structure of network is rational and has rather quick training speed. It also has good generalization ability.