Multimodal-Oriented Interactive Joint Source-Channel Coding for Lightweight Semantic Communication

As a pivotal technology for sixth-generation (6G) mobile networks, semantic communication (SC) aims to enable low-energy, high-efficiency intelligent communication through joint source-channel coding (JSCC). However, current multimodal SC systems experience semantic divergence and conflicts due to imbalances in the representation capacity of individual encoders. Furthermore, high computational demands significantly elevate the deployment and operational costs of SC. In this letter, we propose a multimodal-oriented lightweight SC system based on an interactive JSCC scheme, promoting the evolution of 6G towards a multi-source collaboration paradigm. Additionally, a dynamic resource allocation method is introduced, which hierarchically adjusts computational costs to improve efficiency. Simulation results demonstrate that the proposed system reduces computational overhead by 60<inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\%$</tex-math></inline-formula> compared to six benchmarks while maintaining high performance and robustness in multiple tasks.