Location-Aware Node Management Solution for Multi-Radio Dual Connectivity Scenarios

Abstract

The location of user equipments (UEs) allows application developers to customize the services for users to perceive an enhanced experience. In addition, this UE location enables network operators to develop location-aware solutions to optimize network resource management. Moreover, the combination of location-aware approaches and new network features introduced by 5G enables to further improve the network performance. In this sense, dual connectivity (DC) allows users to simultaneously communicate with two nodes. The basic strategy proposed by 3GPP to select these nodes is based only on the power received by the users. However, the network performance could be enhanced if an alternative methodology is proposed to make this decision. This paper proposes, instead of power-based selection, to choose the nodes that provide the highest quality of experience (QoE) to the user. With this purpose, the proposed system uses the UE location as well as multiple network metrics as inputs. A dense urban scenario is assumed to test the solution in a system-level simulation tool. The results show that the optimal selection varies depending on the UE location, as well as the increase in the QoE perceived by users of different services.

Crowdsourcing Based Performance Analysis of Mobile User Heterogeneous Services

“…Concerning the three MOS scores prediction, the results are given in Table 12. We report in this figure the Mean Absolute Error (MAE) [31] and Mean Absolute Percentage Error (MAPE) [32] and Pearson correlation rate (r) [8]. From Table 12, we notice that all the considered models, except the DT method, performed reasonably well on the task of MOS score prediction and showed high degrees of accuracy with at least 81% in the case of buffer-based MOS, 73% in the case of bitratebased MOS and 63% in the case of buffer-based MOS.…”