Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deep Joint Source Channel Coding With Attention Modules Over MIMO Channels

Weiran Jiang, Wei Chen, Bo Ai

TL;DR

Experimental results demonstrate the proposed DJSCC structures with attention modules have improved image transmission performance, and reveal the phenomenon via non-parameter entropy estimation that the learned DJSCC transceivers tend to transmit more information over better sub-channels.

Abstract

In this paper, we propose two deep joint source and channel coding (DJSCC) structures with attention modules for the multi-input multi-output (MIMO) channel, including a serial structure and a parallel structure. With singular value decomposition (SVD)-based precoding scheme, the MIMO channel can be decomposed into various sub-channels, and the feature outputs will experience sub-channels with different channel qualities. In the serial structure, one single network is used at both the transmitter and the receiver to jointly process data streams of all MIMO subchannels, while data steams of different MIMO subchannels are processed independently via multiple sub-networks in the parallel structure. The attention modules in both serial and parallel architectures enable the system to adapt to varying channel qualities and adjust the quantity of information outputs in accordance with the channel qualities. Experimental results demonstrate the proposed DJSCC structures have improved image transmission performance, and reveal the phenomenon via non-parameter entropy estimation that the learned DJSCC transceivers tend to transmit more information over better sub-channels.

Deep Joint Source Channel Coding With Attention Modules Over MIMO Channels

TL;DR

Experimental results demonstrate the proposed DJSCC structures with attention modules have improved image transmission performance, and reveal the phenomenon via non-parameter entropy estimation that the learned DJSCC transceivers tend to transmit more information over better sub-channels.

Abstract

In this paper, we propose two deep joint source and channel coding (DJSCC) structures with attention modules for the multi-input multi-output (MIMO) channel, including a serial structure and a parallel structure. With singular value decomposition (SVD)-based precoding scheme, the MIMO channel can be decomposed into various sub-channels, and the feature outputs will experience sub-channels with different channel qualities. In the serial structure, one single network is used at both the transmitter and the receiver to jointly process data streams of all MIMO subchannels, while data steams of different MIMO subchannels are processed independently via multiple sub-networks in the parallel structure. The attention modules in both serial and parallel architectures enable the system to adapt to varying channel qualities and adjust the quantity of information outputs in accordance with the channel qualities. Experimental results demonstrate the proposed DJSCC structures have improved image transmission performance, and reveal the phenomenon via non-parameter entropy estimation that the learned DJSCC transceivers tend to transmit more information over better sub-channels.
Paper Structure (8 sections, 9 equations, 6 figures)

This paper contains 8 sections, 9 equations, 6 figures.

Table of Contents

  1. Introduction
  2. System Model
  3. MIMO channel and SVD-based precoding
  4. Proposed MIMO-DJSCC Structures with Attention Modules
  5. Simulation Results
  6. Image reconstruction performance
  7. Entropy of DJSCC encoder output over sub-channels
  8. Conclusion

Figures (6)

  • Figure 1: The MIMO communication system with SVD-based precoding.
  • Figure 2: The proposed MIMO-DJSCC structures.
  • Figure 3: The attention module in both the serial and parallel structures.
  • Figure 4: Performance of DJSCC schemes over MIMO channel. The SVD-based precoding is used in our serial and parallel MIMO-DJSCC and ViT-based DJSCC, and no precoding is used in Multiplexing/Diversity-DJSCC.
  • Figure 5: Performance of our serial and parallel MIMO-DJSCC and Multiplexing/Diversity-DJSCC over MIMO channel with $3$ transmitting antennas.
  • ...and 1 more figures