Access network selection strategy using position prediction in heterogeneous wireless networks

Access network selection(ANS)strategy is one of the most important issues in future heterogeneous networks.The current solutions for this issue are not very efficient because they do not consider the motion scenarios and cannot predict the next location for mobile node.In this paper,an effective ANS strategy based on global positioning system(GPS)is proposed.Making use of information such as position coordinates and moving velocity acquired by GPS,the ANS proposed can predict the next point of attachment for mobile node(MN),assist existing ANS strategy to make more reasonable decision,and achieve better performance.

一种能量高效的基于主成分分析的多接入网络选择策略(英文)

This paper brings forward a novel dynamic multiple access network selection scheme(NDMAS),which could achieve less energy loss and improve the poor adaptive capability caused by the variable network parameters.Firstly,a multiple access network selection mathematical model based on information theory is presented.From the perspective of information theory,access selection is essentially a process to reduce the information entropy in the system.It can be found that the lower the information entropy is,the better the system performance fulfills.Therefore,this model is designed to reduce the information entropy by removing redundant parameters,and to avoid the computational cost as well.Secondly,for model implementation,the Principal Component Analysis(PCA) is employed to process the observation data to find out the related factors which affect the users most.As a result,the information entropy is decreased.Theoretical analysis proves that system loss and computational complexity have been decreased by using the proposed approach,while the network QoS and accuracy are guaranteed.Finally,simulation results show that our scheme achieves much better system performance in terms of packet delay,throughput and call blocking probability than other currently existing ones.