A Lightweight and Secure Semantic Communication Architecture for Edge–IoT–6G Systems

Semantic communication has emerged as a viable paradigm for enhancing efficiency and reliability in next-generation edge-enabled IoT, as well as 6G networks. The paper presents a proposal for a lightweight secure semantic communication (LSSC) architecture, which combines semantic encoding and security validation at the edge. It has been shown that the scheme proposed has 0.91 semantic similarity at 5 dB SNR and semantic similarity of 0.86 at 0 dB, the experimental results being better than the non-secure semantic and conventional bit-level transmission. The results of reported performance improvements are achieved through the controlled edge-IoT in AWGN and Rayleigh fading channel with SNR of 0-20 dB employing publicly available sensing and short-text data. Findings as such are hotspot-based semantic inference performance under limited deployment constraints, as opposed to universal network behaviors. The results show that LSSC has the potential to provide secure and efficient semantic communication within edge-enhanced 6G environments.