Low-Complexity OFDM Deep Neural Receivers

2025-12-04

AI Summary

To address the high computational complexity and slow training convergence of OFDM deep neural receivers (NeuralRx), this paper proposes a low-complexity residual network architecture. The method introduces a lightweight channel-splitting and shuffling module featuring small convolutional kernels, controllable dilation rates, and uniform channel dimensions, while eliminating element-wise addition in residual connections—thereby significantly reducing FLOPs and memory access overhead. Built upon the ResNet framework, the model incorporates GELU activation and a structured channel-splitting mechanism. Experimental results on standard OFDM tasks demonstrate that the proposed approach maintains or even improves decoding accuracy while reducing FLOPs by 37.2% and accelerating training convergence by 2.1× compared to baseline NeuralRx models, validating its efficiency and practicality for real-time OFDM signal processing.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Natural Language Processing: Learning & Optimization for NLP
• Search and Optimization: Learning to Search

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Deep neural receivers (NeuralRxs) for Orthogonal Frequency Division Multiplexing (OFDM) signals are proposed for enhanced decoding performance compared to their signal-processing based counterparts. However, the existing architectures ignore the required number of epochs for training convergence and floating-point operations (FLOPs), which increase significantly with improving performance. To tackle these challenges, we propose a new residual network (ResNet) block design for OFDM NeuralRx. Specifically, we leverage small kernel sizes and dilation rates to lower the number of FLOPs (NFLOPs) and uniform channel sizes to reduce the memory access cost (MAC). The ResNet block is designed with novel channel split and shuffle blocks, element-wise additions are removed, with Gaussian error linear unit (GELU) activations. Extensive simulations show that our proposed NeuralRx reduces NFLOPs and improves training convergence while improving the decoding accuracy.

Problem

• Reduces FLOPs in OFDM neural receivers
• Improves training convergence for deep neural receivers
• Enhances decoding accuracy with low-complexity design

Innovation

• ResNet block with small kernels and dilation
• Uniform channel sizes reduce memory access cost
• Channel split and shuffle with GELU activations