Work

Performance analysis of up-link pilot-based channel estimation in massive MIMO using convolutional neural networks

Year: 2026
Type: article
Abstract: Massive multiple-input multiple-output (MIMO) is a key technology in 5G, enabling a base station (BS) to support hundreds of ports and thereby improve spectral efficiency (SE). Despite its high potential throughput, channel estimation remains a fundamental challenge. We need to address factors such as computational complexity, channel reciprocity, and pilot contamination that impair performance. Accurate channel estimation is crucial for the performance of massive MIMO systems, and this paper explores convolutional neural networks (CNNs) in pilot-based channel estimation techniques, focusing on key parameters like computational complexity, mean square error (MSE), and SE. Simulation results show that pilot-based channel estimation without CNN enhancement has lower computational complexity, revealing a trade-off between improved accuracy and increased complexity. Conversely, the CNN-enhanced MMSE estimator demonstrates significant improvements, achieving the lowest NMSE and highest SE when combined with the maximum ratio combining (MRC) scheme. This suggests that although CNNs increase computational demands, they substantially boost accuracy and SE in massive MIMO channel estimation. Therefore, CNNs offer a promising approach to balancing performance trade-offs for reliable, high-throughput wireless communication.
Institution: Woldia University
Language: English
Open Access status: diamond
OpenAlex ID: W7236342561
FWCI:
Cites:
Cited by:
Related to: 10