Research On Access Selection Schemes For Heterogeneous Radio Networks

Heterogeneous network access selection is a key technology in the future ABC network and it should achieve two goals, one of which is selecting a best network for terminal users to offer "always be connected" service, and the other is improve network resource utilization. The traditional access selection algorithm often cannot pay attention to both of these two goals, so we need to seek better access selection algorithm on the basis of previous studies. Therefore our research is very significance.In this paper, author mainly studied heterogeneous wireless network access selection algorithm, include the access selection algorithm based on the effective capacity and its improved algorithm, the network prediction algorithm and dynamic access selection algorithm based on network parameter prediction. The paper’s main content is as follows:First, The author analyzed the channel capacity and effective capacity theoretically, then according the analysis and comparison results we concluded that effective capacity is a quantitative description of service ability about the physical layer service process. The channel capacity can be considered as a special case of effective capacity because of considing the QoS parameters of transport service. In link layer when the time-delay resulted in business Queuing is greater than the permissible maximum time-delay. the QoS will drop. The proposed access selection algorithm is based on the effective capacity and can make the throughput of ABC network biggest. Though it contributed for the network throughput, it ignores other network elements, such as signal-strength. Therefore in its improved algorithm, MADM process is added when making network selection. This improved algorithm first do a MADM to confirm a target network for users, the second is to determine the user’s access priority using the original algorithm. According to analyzing the simulation results we can see that this algorithm can increase network throughput and reduce service time-delay.Because of the variability character of network’s parameter, usually we cannot get the accuracy value of this parameter. In addition, the decision results will be affected by the influence of the parameters changing. So in this paper, we proposed a dynamic access selection algorithm based on the parameter prediction. This algorithm use kalman filter to predict the future network parameters and using the predicting results to change original MADM. The simulation results show that kalman filter can effectively predict network parameters changing trend, and the improved algorithm can effectively reduce the filling business blocking rate.