Study On Optimization Test And Analysis System Of CDMA Network Based On Fuzzy-neural Network

In CDMA network fault diagnosis and network optimization, it is extremely hard to obtain a precise mathematical model, because the relationship between fault symptom and fault cause is complicated and nonlinear. Consequently, it is very difficult to analyze, diagnose, and solve those problems for a CDMA network optimization engineer.According to the above-mentioned question, by means of the study on calculation principle of fuzzy-neural network (FNN) and on the base of the use of FNN in diagnosis technology, the fuzzy-neural network model for CDMA network fault diagnosis system is set up, in the thesis. The method preferably solves the difficult problem which general diagnosis arithmetic can hardly build model for CDMA network fault diagnosis system. The model is trained by MATLAB software and the synaptic weight value and fuzzy membership function of the network are determined. The model is simulated using MATLAB software, and the correctness of the model is validated. Based on the model and through the use of CDMA network fault diagnosis system, the system's hardware that can accomplish data collecting and fault diagnosis function, the hardware configures and the software plans are worked out.From the necessary of engineering realization, the main contributions of this paper include the following aspects:In the first chapter, the aim of thesis is put forward, firstly. Then, the domestic and overseas developments of CDMA optimization are summarized. At last, the purposes and meanings of this thesis are introduced.In chapter 2, the main characteristics and technical parameters of CDMA mobile communication network system are summarized, and the main flows and contents are also introduced. All the work above is the basis of fault diagnosis and analysis for CDMA network.In chapter 3, by analyzing the fault diagnosis methods of test and analysis systemfor CDMA network, the fuzzy neural network method for fault diagnosis is selected, which combines the fuzzy neural network with the CDMA expert database of network optimization experience.In chapter 4, the fault types and fault causes of CDMA network are obtained from some senior engineers. According to that information, the input variables of diagnosis system are induced, which can denote the states of a working CDMA network. As a result, an expert database of fault diagnosis is set up for a CDMA network based on fuzzy logic. An expert database of fault diagnosis model is proposed for a CDMA network based on fuzzy neural network. For the fuzzy neural network expert system, the number of neurons for input layer, output layer and hidden layer are determined based on the amount of fault type and the input variables needed for fault diagnosis. The simulation tests of training fuzzy neural network expert system are implemented using MATLAB software. As a result, membership function and synaptic weight coefficient (fuzzy relationship matrix) are determined. Through many simulation tests and comparisons, the results indicate the validity of the fuzzy neural network expert system proposed in this thesis.In chapter 5, the software and hardware designs of the optimization test and analysis system of CDMA network are researched based on research results, which are proposed in above all chapters.The conclusions and directions for future research work are discussed in the last chapter of this paper.The main innovative contributions of this paper include:1. CDMA network fault diagnosis method based on fuzzy-neural network is put forward for the first time in this paper.2. This paper is the first to summarize systemically the relationship between fault types and fault causes of CDMA network.3. From the engineering realization, a hardware system design of the intelligent test and analysis system of CDMA network is proposed.