Authors: Mehdi Sattari, Hao Guo, Deniz Gündüz, Ashkan Panahi, Tommy Svensson

Published on: February 06, 2024

Impact Score: 8.22

Arxiv code: Arxiv:2402.03886

Summary

• What is new: Introduces full-duplex transmissions for mmWave MIMO systems using neural networks for channel estimation, addressing self-interference and pilot overhead issues.
• Why this is important: The challenge of efficient channel estimation in full-duplex mmWave MIMO systems, hindered by self-interference and the need to reduce pilot overhead.
• What the research proposes: Employing neural networks to share pilot resources effectively between user equipments and transmit antennas, and to map channels between downlink UEs and the transmit/receive antenna arrays.
• Results: Achieved reduced pilot overhead in full-duplex systems to a level comparable to half-duplex systems and provided successful channel estimation despite non-linear distortions and varying channel conditions.

Technical Details

Technological frameworks used: Neural networks with varying architectures

Models used: Different numbers of hidden layers, introduction of non-linear distortion

Data used: Pilot signals, channel conditions (low-correlated and high-correlated channels)

Potential Impact

Telecommunications, IoT device manufacturers, companies involved in 5G/6G technology development

Want to implement this idea in a business?

We have generated a startup concept here: WaveAI Solutions.