On the performance of multiple association with mobility support in 5G ultra‐dense networks: Realistic network simulation

Abstract

AbstractA user equipment (UE) in ultra‐dense networks (UDN) has a high chance of being in close proximity to many cells. multiple association (MA) can be used to serve users with high data rate demands in such a deployment scenario. However, because of the small footprint of the low power node (LPN), mobile users may experience highly frequent cell re‐association issues. This paper focuses on solutions for UEs with high data rate demands as well as mobility management. User‐centric clustering, control, and user‐plane split usages are designed and investigated in coordination with the MA scheme. The clustering involves two stages: a master clustering and user‐centric clustering. This approach goes with mobility management in long term evolution advanced (LTE‐A)/fifth generation (5G) and can easily be integrated with the existing system. By setting some predefined threshold, the scheme attempts to treat UEs separately, based on their speed. In addition, a clustering approach was developed that generates virtual cells with which UEs are associated. The combination of MA with clustering and mobility support improves cooperation between the most appropriate cells to serve a specific UE. To validate the performance gains, synthetic system level and realistic network deployment simulations were used. The results show that the proposed scheme solves the frequent re‐association issues and improves the user throughput performance. In the realistic network evaluation, up to 65.7% decrease in re‐association frequency was obtained for the UEs average speed of 90 Km/h with cluster size . In addition, the percentile UE throughput when showed about 5.5% increase compared with the case when for pedestrian and low speed users.