Fronthaul Load Balancing in Energy Harvesting Powered Cloud Radio Access Networks

Abstract

Enhanced with wireless power transfer capability, cloud radio access network (C-RAN) enables energy-restrained mobile devices to function uninterruptedly. Beamforming of C-RAN has potential to improve the efficiency of wireless power transfer, in addition to transmission data rates. In this paper, we design the beamforming jointly for data transmission and energy transfer, under finite fronthaul capacity of C-RAN. A non-convex problem is formulated to balance the fronthaul requirements of different remote radio heads (RRHs). Norm approximations and relaxations are carried out to convexify the problem to second-order cone programming (SOCP). To improve the scalability of the design to large networks, we further decentralize the SOCP problem using the alternating direction multiplier method (ADMM). A series of reformulations and transformations are conducted, such that the resultant problem conforms to the state-of-the-art ADMM solver and can be efficiently solved in real time. Simulation results show that the distributed algorithm can remarkably reduce the time complexity without compromising the fronthaul load balancing of its centralized counterpart. The proposed algorithms can also reduce the fronthaul bandwidth requirements by 25% to 50%, compared with the prior art.