Abstract
In this work, we consider a heterogeneous cellular network (HCN) with a macrocell and multiple femtocells. In each femtocell, the femto base station (FBS) serves a femto user (FU), which is capable of simultaneous wireless information and power transfer (SWIPT) energy harvesting (EH) in a power splitting manner. All FBSs suppress their interference to macro users (MUs) via a interference power constraint. In such a HCN, we aim to maximize the energy efficiency (EE) of all femtocells via the joint design of beamforming and power splitting ratios under partial channel state information (CSI) - imperfect CSI and channel distribution information (CDI) at FBSs. For the imperfect CSI case, we consider the robust beamforming design that maximizes the worst-case EE for all femtocells under the quality-of-service (QoS) constraints and EH constraints. For the CDI case, we consider the problem of maximizing EE under rate outage constraints. The non-convex objective functions and constraints make the problem difficult to solve. To resolve this issue, We proposed successive convex approximation (SCA) algorithms (EE-SCA-1 for the imperfect CSI case and EE-SCA-2 for the CDI case) that provide high-quality approximate solutions. Simulations demonstrated that our proposed EE-SCA-1 and EE-SCA-2 achieve near-optimal performance for both the imperfect CSI and CDI cases. Furthermore, for the imperfect CSI case, our EE-SCA-1 improved EE by as much as 23.1% and 267.6% over the zero-forcing (ZF) scheme and a power minimization scheme. For the CDI case, our EE-SCA-2 improved EE by as much as 116.7% and 21.1% over the ZF scheme and a rate maximization scheme. It indicates that careful design of energy efficient beamforming scheme is essential for the next-generation wireless communication systems.