AI in Sustainable Energy and Environment

en Deep Learning for Massive MIMO: AI-Driven Channel Estimation with Reduced Pilot Usage عمومى General پژوهشي Research This article investigates uplink massive MIMO system using 1-bit analog-to-digital converters (ADCs) and introduces a deep-learning-based framework for channel estimation. The proposed approach utilizes prior channel estimation observations alongside deep neural networks to establish a sophisticated mapping between quantized received measurements and the corresponding channels. To support this, the necessary pilot sequence length and structure are determined to ensure the feasibility of such a mapping function. An exciting and somewhat surprising finding is that increasing the number of base station antennas enhances the performance of the deep learning-based channel estimation for a fixed pilot sequence length. Alternatively, for a preferred channel estimation performance, smaller number of pilot sequences is desirable as the number of antennas increases. This observable fact is analytically demonstrated for specific channel models. Simulation results validate these findings, revealing that a high number of antennas improve channel estimation performance in terms of predicted signal to noise ratio per antenna and normalized mean squared error.   Massive MIMO, AI-Driven Channel Estimation, 1-bit ADCs, Pilot Reduction, Deep Learning in Wireless Communication 123 129 http://aisesjournal.com/browse.php?a_code=A-10-351-4&slc_lang=en&sid=1 Rishika Chauhan rishikaster@gmail.com 1003194753284600201 1003194753284600201 No Jaypee University of Engineering and Technology Shefali Sharma shefali.sharma@juetguna.in 1003194753284600202 1003194753284600202 No Jaypee University of Engineering and Technology Rahul Pachauri rahul.pachauri@juetguna.in 1003194753284600203 1003194753284600203 No Jaypee University of Engineering and Technology Pankaj Dumka p.dumka.ipec@gmail.com 1003194753284600204 1003194753284600204 Yes Jaypee University of Engineering and Technology