Distortion-Aware Power Allocation for Multi-Stream Distributed Massive MIMO System With Nonlinear Power Amplifier

Distributed MIMO systems leverage multiple access points (APs) distributed across a geographical area to enhance system performance and provide robust connectivity to users by transmitting independent data streams simultaneously, thus improving coverage and capacity. The nonlinear power amplifiers (PAs) employed at the APs can significantly distort the transmitted signal and degrade the spectral efficiency (SE), as PAs are more efficient when operating close to saturation. In this paper, we present an efficient power allocation framework in a multi-stream distributed massive MIMO system with nonlinear PAs, where APs send multiple spatial streams to the user equipment (UE). With the objective of maximizing the SE while adhering to total input power constraints, we first formulate the power allocation problem as nonlinear programming, which is first solved using the multiplier punitive function-based method. Moreover, we focus on the asymptotic analysis of the SE when multi-stream interference among APs is negligible. This can be achieved through ideal spatial multiplexing with interference nulling, assuming a sufficiently large antenna at the UE, or through interference-free frequency multiplexing. We examine the optimality of the water-filling method in power allocation and extend it by considering the PA's nonlinearity. Simulation results show that the proposed power allocation method can maximize the achievable rate while considering the PA nonlinearity, thereby preventing degradation in performance.