Energy-Aware Design For Mimo-Ofdm Network With Realistic Interference Model

Network densification is a promising technology to increase 5G system capacity, including large scale multiple-input multiple-output (MIMO) and ultra-dense network. Energy saving and resource utility are the key issues for the densified OFDM system. However, most of previous studies do not taking account of the effect of resource utility in the SINR model, e.g. the load factor. In this paper, we firstly formulate the SINR using more realistic model considering load factor in OFDM systems. Based on this model the energy saving problem is formulated as a joint optimization of beamforming and power subject to users' QoS and load constraint of each cell. To sovle this problem, we propose a two-stage scheme to decouple this problem into the beamforming and power control. In the beamforming stage, we resort to a suboptimal beamformer using the idea of best linear unbiased estimator (BLUE); For the power control stage, we try to minimize the transmit sum-power while fulfilling the user's QoS and load constraints based on theory of standard interference mapping. The convergence behavior of the proposed scheme is rigorously proved. Simulation results show that the proposed scheme converges faster in multiple-antenna cases comparing with the single-antenna counterpart. Moreover the proposed beamformer owns advantage against the traditional ones in term of power saving.