Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

CoNet-Rx: Collaborative Neural Networks for OFDM Receivers

Mohanad Obeed, Ming Jian

TL;DR

The paper tackles the inefficiency of DL-based OFDM receivers that reuse CV-inspired architectures by introducing CoNet-Rx, a collaborative multi-subnetwork receiver that fuses features from distinct perspectives via element-wise interactions. By employing a main and a support network with MRB blocks, CoNet-Rx achieves BER improvements over the DeepRx baseline and reduces inference latency by about 42% without increasing parameter count. The approach demonstrates robustness across realistic channels and interference scenarios, highlighting its suitability for real-time, resource-constrained wireless systems. Overall, CoNet-Rx advances practical neural receivers by balancing high detection performance with low latency and modest resource demands, enabling deployment on embedded and edge devices.

Abstract

Deep learning (DL) based methods for orthogonal frequency division multiplexing (OFDM) radio receivers demonstrated higher signal detection performance compared to the traditional receivers. However, the existing DL-based models, usually adapted from computer vision, aren't well suited for wireless communications. These models require high computational resources and memory, and have significant inference delays, limiting their use in resource-constrained settings. Additionally, reducing network size to ease resource demands often leads to notable performance degradation. This paper introduces collaborative networks (CoNet), a novel neural network (NN) architecture designed for OFDM receivers. CoNet uses multiple small ResNet or CNN subnetworks to simultaneously process signal features from different perspectives like capturing channel correlations and interference patterns. These subnetworks fuse their outputs through interaction operations (e.g., element-wise multiplication), significantly enhancing detection performance. Simulation results show CoNet significantly outperforms traditional architectures like residual networks (ResNets) in bit error rate (BER) and reduces inference delay when both nets have the same size and the same computational complexity.

CoNet-Rx: Collaborative Neural Networks for OFDM Receivers

TL;DR

The paper tackles the inefficiency of DL-based OFDM receivers that reuse CV-inspired architectures by introducing CoNet-Rx, a collaborative multi-subnetwork receiver that fuses features from distinct perspectives via element-wise interactions. By employing a main and a support network with MRB blocks, CoNet-Rx achieves BER improvements over the DeepRx baseline and reduces inference latency by about 42% without increasing parameter count. The approach demonstrates robustness across realistic channels and interference scenarios, highlighting its suitability for real-time, resource-constrained wireless systems. Overall, CoNet-Rx advances practical neural receivers by balancing high detection performance with low latency and modest resource demands, enabling deployment on embedded and edge devices.

Abstract

Deep learning (DL) based methods for orthogonal frequency division multiplexing (OFDM) radio receivers demonstrated higher signal detection performance compared to the traditional receivers. However, the existing DL-based models, usually adapted from computer vision, aren't well suited for wireless communications. These models require high computational resources and memory, and have significant inference delays, limiting their use in resource-constrained settings. Additionally, reducing network size to ease resource demands often leads to notable performance degradation. This paper introduces collaborative networks (CoNet), a novel neural network (NN) architecture designed for OFDM receivers. CoNet uses multiple small ResNet or CNN subnetworks to simultaneously process signal features from different perspectives like capturing channel correlations and interference patterns. These subnetworks fuse their outputs through interaction operations (e.g., element-wise multiplication), significantly enhancing detection performance. Simulation results show CoNet significantly outperforms traditional architectures like residual networks (ResNets) in bit error rate (BER) and reduces inference delay when both nets have the same size and the same computational complexity.
Paper Structure (6 sections, 1 equation, 9 figures)

This paper contains 6 sections, 1 equation, 9 figures.

Table of Contents

  1. Introduction
  2. System Model
  3. ResNets
  4. The Proposed Neural Receiver (CoNet-Rx)
  5. Simulation Results
  6. Conclusion

Figures (9)

  • Figure 1: Residual block (RB)
  • Figure 2: Example I of CoNet architecture, RB and MRB blocks are illustrated in figures \ref{['fig:RB']} and \ref{['fig:MRB']}, respectively.
  • Figure 3: Modified residual block (MRB)
  • Figure 4: Main component of CoNet
  • Figure 5: Example II of CoNet architecture
  • ...and 4 more figures