Clustering of Remote Radio Heads in C-RAN to minimize the number of Base-Band Units

Cloud Radio Access Networks (C-RAN) offer the means for adopting the coverage and the capacity according to the changing user densities and the associated change of required capacity. C-RAN decouples the provisioning of virtual Base Band Unit (BBU) from the physical transmission provided by the Remote Radio Head (RRH). This decoupling provides a similar decoupling between capacity and coverage. C-RAN allows the activation of the number of RRHs needed for coverage and the number of BBUs needed to satisfy the capacity need. The ability to connect multiple RRHs to the same BBU allows for enhancing coverage while achieving rate requirements using the least number of BBUs. The aim of this paper is to minimize the number of BBUs used to offer wireless coverage by providing techniques to split the coverage into the smallest number of RRH clusters. Each cluster is then connected to a single BBU. The paper investigates the use of three RRH clustering techniques in indoor 4G networks: a modified K-means based technique and two heuristic algorithms using numerical simulations. Results show that the K-means technique produced lower activated BBUs and the RRH clusters were tighter in space providing better user coverage per cluster.