Concurrent Load-Aware Adjustment of User Association and Antenna Tilts in Self-Organizing Radio Networks

Abstract

Network operators expect a coordinated handling of parameter changes submitted to the operating network's configuration management entity by closed-loop self-organizing network (SON) techniques. For this reason, a major research goal for emerging SON technologies is to achieve coordinated results out of a plethora of independently or even concurrently running use-case implementations. In this paper, we extend current frameworks to compute desirable user associations by an interference model that explicitly takes base-station loads into account. With the aid of this model, we are able to make considerably more accurate estimations and predictions of cell loads compared with established methods. Based on the ability to predict cell loads, we derive algorithms that jointly adapt user-association policies and antenna-tilt settings for multiple cells. We demonstrate by detailed numerical evaluations of realistic networks that these algorithms can be applied to capacity and coverage optimization, mobility load balancing, and cell outage compensation use cases. As a result, rather than performing any heading or tailing coordination, the joint technique inherently comprises all three use cases, making their coordination redundant. For all scenarios studied, the joint optimization of tilts and user association improves quality of service in terms of the fifth percentile of user throughput compared with state-of-the-art techniques. The proposed models and techniques can be straightforwardly extended to other physical and soft parameters.