The Application Of Fuzzy Neural Network In Mobile Communication Network Fault Diagnosis

With the development of science and technology, especially of wireless networking technology, the ways of communication have been changing greatly. Contemporary, mobile communication network has been infiltrating into people's working and lives. As the coming of 3G era, the size, speed, efficiency, capacity, intensity of mobile network will be greatly improved; in the other hand, the loss of system downtime or data loss caused by system failure will be even more serious. The traditional technology of fault diagnosis has been unable to meet the needs of such a large network, but along with the development of computer technology and artificial intelligence technology, the technology of fault diagnosis has entered a intelligent phase.Today, the theories fuzzy theory and BP neural network theory have been mature after decades of development, and hava some achiements in the area of fault diagnosis. In this paper, fuzzy theory and BP neural network in the application of fault diagnosis are discussed and a smart fault diagnosis structure which by BP neural network mainly and fuzzy theory as a useful complement is presented, so that the result of fault diagnosis will be more exact. First of all, fault diagnosis system receives all kinds of operating parameters which are as input for the BP neural network after the fuzzication processing by membership function, and then takes the trained neural network which has the capable of memory and association to identify the failures, and last get the final diagnosis results in the output layer.The experimental system model in this paper mainly contains two main parts:fuzzy systems and neural networks. Fuzzy system is used to fuzzifier actual fault diagnosis data, so that the data is in the range of [0,1], such data has a better convergence; neural network is used to train the network with the data which is fuzzifiered so as to get the fault identification network system. At last, the network is tested by test data.The result of the experiments has shown that the proposed fuzzy neural network structure has better convergence and faster convergence speed than BP neural network structure,...