A Novel Approach for Centralized 3D Radio Resource Allocation and Scheduling in Dense HetNets for 5G Control-/User-plane Separation Architectures

This paper presents a centralized 3-dimensional radio resources (namely, time, frequency, and power) allocation and scheduling approach for control-plane and user-plane (C-/U-plane) separation architectures for fifth generation mobile networks. A central station is considered where schedulers of all base stations (BSs) are located. We consider a multi-tier network that comprises of a macrocell BS (MCBS), several outdoor picocell BSs, and a number of indoor femtocell BSs (FCBSs) deployed in a number of multi-storage buildings. The system bandwidth is reused in FCBSs within each building orthogonally. In contrast to the conventional almost blank subframe, we consider a fully blank subframe based time-domain enhanced intercell interference coordination to split completely C-/U-plane traffic such that the controlplane can be served only by the MCBS and the user-plane of user equipments by their respective BSs. We propose two power management schemes for FCBSs based on whether or not the coordinated multi-point communication with joint transmission (JT CoMP) is employed during off-state of a FCBS and develop a power control mechanism for both a single user and multi-user per FCBS scenarios. An optimal value of average activation factor (OAF) for a FCBS is derived to trade-off its serving capacity and transmit power saving factor. It is shown that in order to improve the network capacity, a FCBS needs to operate at an average activation factor (AAF) greater than its OAF using JT CoMP to serve neighboring on-state FCBSs during its normal off-state, whereas at an AAF less than the OAF to improve the energy efficiency. With a system level simulation, we show that the capacity of a FCBS increases, whereas its power saving factor decreases linearly with an increase in its AAF because of serving increased traffic, and an OAF of 0.5 for the capacity scaling factor 2 1 and greater than 0.5 for 1 are found.