Bit-level Optimized Neural Network for Multi-antenna Channel Quantization

TL;DR

This paper introduces a deep learning-based method using JCResNet for efficient bit-level quantization of MIMO channel state information, significantly reducing feedback overhead and improving performance.

Contribution

It presents a novel joint convolutional residual network for bit-level CSI quantization, enhancing feature extraction and recovery in MIMO systems.

Findings

Substantially improved quantization performance

Reduced CSI feedback overhead

Effective deep learning-based approach

Abstract

Quantized channel state information (CSI) plays a critical role in precoding design which helps reap the merits of multiple-input multiple-output (MIMO) technology. In order to reduce the overhead of CSI feedback, we propose a deep learning based CSI quantization method by developing a joint convolutional residual network (JCResNet) which benefits MIMO channel feature extraction and recovery from the perspective of bit-level quantization performance. Experiments show that our proposed method substantially improves the performance.