A LTE/WLAN selection method based on a novel throughput estimation method
Takayoshi Nakayama, Senoo Hiroaki, Yun Wen, Yoshiharu Tajima, Dai Kimura · 2017
In this paper, we propose a novel radio access technology (RAT) selection method that can maximize total throughput by simultaneously selecting optimal RATs for a group of user equipments (UEs). To perform simultaneous selection, we also propose a novel throughput estimation method based on not wireless information collected from the existing network equipment but the accurate estimation of the amount of radio resources available for the UEs. Throughputs of all the UEs are estimated for all RAT selection patterns, by using the estimated available radio resources. The RAT selection method then selects a RAT selection pattern that gives the maximum total throughput. We conduct both simulations with a network simulator and field experiments with a commercially operating LTE network. The field experiments confirm that the proposed throughput estimation method can achieve an average estimation error of about 10%. The simulations confirm that the proposed method with a 10% throughput estimation error can achieve 96% throughput compared to that with perfect estimation. A conventional successive RAT selection method takes 1–10 minutes to achieve the same throughput as the proposed method, while the proposed simultaneous RAT selection method only takes about 1–10 seconds. Moreover, the conventional method achieves only 71% throughput of the proposed method with same processing time as the proposed method.