Modified fractional power allocation for downlink cell-free massive MIMO systems

Cell-free massive multiple-input multiple-output (mMIMO) significantly improves the spectral efficiency (SE) performance compared to conventional centralized mMIMO through its distributed antenna architecture. Fractional power allocation (FPA) algorithm is widely used for scalable power control with good performance in downlink (DL) of cell-free mMIMO. In this paper, we propose modified FPA (MFPA) and generalized FPA (GFPA) strategies for centralized and distributed precoding in the DL of cell-free networks, respectively. For the former, we abandon the traditional normalization of precoding vectors and introduce three adjustment parameters, which can dynamically adjust the power allocation of the DL according to the actual channel conditions. Regarding the latter, the GFPA strategy finds effective channel factors suitable for various distributed precoding schemes and correlates them with the power allocation coefficients of each user equipment (UE), enabling power allocation to adapt to multiple precoding schemes. Analysis and simulation results demonstrate that, under the MFPA strategy, UEs with poorer channel conditions can achieve higher SE, but at the expense of other UEs with better channel conditions. Under the GFPA strategy, UEs with better channel conditions experience significant SE improvements without sacrificing UEs performance with poorer channel conditions.