Throughput and Delay Driven Access Point Placement

With higher rates and increased diversity, distributed massive MIMO systems show promise. In this work, we look at such a system with no collaboration between Access Points (APs). We address the AP placement problem with two criteria in mind: throughput and delay. While the former deals with optimal AP locations given a user distribution, the latter adds a constraint of equal user access to the APs. In this regard, we propose the Cell Equalized Lloyd Algorithm (CELA), which in addition to the steps of the Lloyd algorithm, moves users from cells with more users to cells with fewer users in order to equalize the number of users in each cell without significantly affecting user throughput. Normalized user rate and user access count metrics are employed to evaluate the system characteristics. The performance of CELA is compared with that of the traditional Lloyd algorithm, and while a slight reduction in throughput is observed, the increase in user access outweighs this reduction.