IMPROVING SPECTRAL EFFICIENCY IN MASSIVE MIMO FOR 5G

The growing use of data-hungry applications and the rapid rise in connected devices have placed
stringent requirements on modern wireless networks, including higher data rates, lower latency,
and improved energy efficiency. Fifth-generation (5G) communication systems address these
demands through advanced technologies such as Massive Multiple-Input Multiple-Output
(Massive MIMO), beamforming, and millimetre - wave transmission. Among these, Massive
MIMO plays a vital role by leveraging large antenna arrays at the base station to enhance spectral
efficiency (SE) and system capacity through spatial multiplexing. However, the effectiveness of
Massive MIMO largely depends on accurate Channel State Information (CSI) and efficient precoding strategies. Conventional linear pre-coding methods, particularly Zero Forcing (ZF) and
Minimum Mean Square Error (MMSE), are widely used due to their simplicity and strong
interference mitigation capability. Nonetheless, their performance degrades under imperfect CSI,
noise uncertainty, and increasing computational complexity in realistic channel conditions. Hence,
evaluating these techniques under practical 5G scenarios is necessary to establish reliable
performance benchmarks. This work investigates spectral efficiency enhancement in a multi-user
Massive MIMO downlink system using classical signal processing techniques. MATLAB R2022a
simulations are carried out over a Rayleigh fading channel with additive white Gaussian noise.
MMSE-based channel estimation and ZF pre-coding are employed, and performance is evaluated
with respect to SNR and antenna scaling. The results confirm that increasing antenna count
significantly improves SE, while highlighting the performance gap caused by CSI imperfections.
This study forms a baseline for future exploration of Quantum-Inspired Deep Learning-based
precoding solutions.