Edge-Device Collaborative Semantic Communication Service for Distribution Network

Accurate data transmission and effective information representation significantly influence the smart management and service applications within distribution networks. To facilitate efficient and resilient information transmission under the constraints of limited and noisy communication conditions present in distribution networks, this study investigates the application of the semantic communication (SemCom) paradigm within the distribution network, emphasizing the construction of a Semantic Knowledge Base (SKB) for SemCom services in distribution networks, given that the SKB serves as a fundamental prerequisite for ensuring precise semantic comprehension and transmission. Specifically, we initially devise a SemCom system model incorporating edge-device collaboration to bolster SKB construction and SemCom services. Subsequently, we present the integration of Federated Learning (FL) to develop the SKB, leveraging the advantages of FL in feature aggregation and knowledge dissemination. Moreover, the personalized SKB construction methodology is realized, taking into account the data heterogeneity inherent in local devices, which facilitates the updating of the public SKB and the acquisition of the private SKB. Finally, we substantiate the efficacy of the proposed approach in terms of the private SKB and the performance of personalized tasks.