Joint Source-Channel Coding for Channel-Adaptive Digital Semantic Communications

TL;DR

A new demodulation method is developed which assesses the uncertainty of the demodulation output to improve the robustness of the digital semantic communication system and a robust training strategy is devised which enhances the robustness and flexibility of the JSCC encoder and decoder against diverse channel conditions and modulation orders.

Abstract

In this paper, we propose a novel joint source-channel coding (JSCC) approach for channel-adaptive digital semantic communications. In semantic communication systems with digital modulation and demodulation, robust design of JSCC encoder and decoder becomes challenging not only due to the unpredictable dynamics of channel conditions but also due to diverse modulation orders. To address this challenge, we first develop a new demodulation method which assesses the uncertainty of the demodulation output to improve the robustness of the digital semantic communication system. We then devise a robust training strategy which enhances the robustness and flexibility of the JSCC encoder and decoder against diverse channel conditions and modulation orders. To this end, we model the relationship between the encoder's output and decoder's input using binary symmetric erasure channels and then sample the parameters of these channels from diverse distributions. We also develop a channel-adaptive modulation technique for an inference phase, in order to reduce the communication latency while maintaining task performance. In this technique, we adaptively determine modulation orders for the latent variables based on channel conditions. Using simulations, we demonstrate the superior performance of the proposed JSCC approach for image classification, reconstruction, and retrieval tasks compared to existing JSCC approaches.

Figures (8)

• Figure 1: Illustration of the digital semantic communication system with the BSEC modeling considered in our work.
• Figure 2: Visualization of the decision boundaries associated with the second bit of a 16-QAM symbol.
• Figure 3: Changes in the modulation orders assigned to the latent variables for various SNRs when employing our channel-adaptive modulation technique.
• Figure 4: Comparison of the classification accuracies of various JSCC and non-JSCC approaches with 4-QAM for image classification tasks using the MNIST, Fashion-MNIST, and CIFAR-10 datasets.
• Figure 5: Comparison of the BERs of various JSCC and non-JSCC approaches with 4-QAM for the image classification task using the CIFAR-10 dataset.