The Flow Distribution Of Multi-mode Terminal Based On Artificial Neural Network

With the development of communication, a series of different kinds of wireless access networks have emerged, including2G,3G and WLAN wireless network, the coverage of these different networks overlap and form a heterogeneous network environment. How to select the most appropriate network to access for a multi-mode mobile terminal in heterogeneous network environment has become an urgent problem to be solved in wireless communications area.This thesis mainly studies centralized access selection algorithm in heterogeneous wireless network, simulation and analysis were done to these algorithms. Then the artificial neural network is used to solve the network select problem, the advantage of using artificial neural network is not necessary to calculate the weights of parameters, after the neural network training is completed, the best access network is obtained when the network parameters are inputted to the neural network, this is simpler, more adaptive compared to the traditional method. Another improved method in this thesis is the parameter weighting processing, there are subjective parameter weights and objective parameter weights in network selection, the previously mentioned algorithms only use one weight, but in this thesis, the subjective and objective weights are combined to get a new weight, which means that both the user’s subjective needs and the objective conditions of the actual network are considered, simulations result shows that the performance of using combined weight is between using subjective weight and objective weight.