Energy Efficiency of Downlink Cell-Free Massive MIMO System with Full-Pilot Zero-Forcing

摘要

This paper investigates the downlink energy efficiency (EE) of cell-free massive multiple-input multiple-output (MIMO) system, where all the geographically distributed access points (APs) with large numbers of antennas coherently serve a number of independent users. In contrast to conventional zero-forcing (ZF) precoding scheme, this paper aims at suppressing the inter-cell interference by leveraging on the full-pilot ZF(fpZF) precoding strategy. In particular, the fpZF precoding can be implemented at the APs–in a distributed manner, there is no channel state information (CSI) exchange between the APs and the central processing unit (CPU). In this work, based on the obtained EE, a downlink EE maximization problem is explored under each user’s quality-of-service (QoS) requirement and each AP’s power constraint. Considering the proposed EE maximization problem is non-convex, a sub-optimal sequential convex approximation (SCA) algorithm is presented. Simulation results demonstrate that the proposed SCA algorithm can converge to a stable value after a few iterations and outperform the full power transmission counterpart.