Power allocation strategies to maximize energy efficiency in cell-free massive MIMO with imperfect CSI

Yashuven, Sanmugam (2026) Power allocation strategies to maximize energy efficiency in cell-free massive MIMO with imperfect CSI. Final Year Project, UTAR.

PDF Download (4Mb)

Abstract

The rapid growth of global mobile data traffic and the increasing demand for sustainable wireless connectivity have elevated energy efficiency to a primary design objective in fifth generation and sixth-generation wireless communication systems. Cell-Free Massive Multiple-Input Multiple-Output has emerged as a promising architecture for next-generation networks, offering superior coverage uniformity and spectral efficiency through the cooperative transmission of geographically distributed Access Points. However, practical CFmMIMO deployments face three critical and interconnected challenges: pilot contamination causing imperfect Channel State Information, inter-user interference limiting signal quality, and suboptimal power allocation reducing energy efficiency. This project develops and evaluates an integrated simulation framework that addresses all three challenges progressively within a unified system model implemented in MATLAB R2022b. Four combinations of pilot assignment strategy and channel estimation method are implemented and compared, namely Least Squares, Matched Ratio, Minimum Mean Square Error, and the proposed Zero-Forcing precoding scheme, with random and greedy pilot assignment strategies respectively. Three power allocation strategies, Equal Power Allocation, Random Power Allocation, and the proposed Proximal Gradient algorithm, are evaluated under both ZF and MMSE precoding frameworks across 500 Monte Carlo realisations. The proposed ZF-PG framework achieves a peak Signal-to-Interference-plus-Noise Ratio of approximately 26 dB at 100 Access Points, a peak Spectral Efficiency of approximately 4.5 bit/s/Hz, and a peak Energy Efficiency of approximately 1.04 Mbit/Joule, representing improvements of 57 percent over MMSE-PG and over 200 percent over baseline methods. The results demonstrate that the integration of greedy pilot assignment, Zero-Forcing precoding, and Proximal Gradient power allocation delivers compounding and consistent energy efficiency gains across all evaluated operating dimensions, establishing the proposed framework as a practical and scalable solution for energy-efficient next-generation wireless network design.

Actions (login required)