Semantic-empowered integrated sensing and communication with resource slicing and bidirectional mapping

In integrated sensing and communication (ISAC) system, achieving signal-level integration is essential but challenging, as direct amalgamation of communication and sensing signals confronts inherent heterogeneity, engendering substantial complexity in both software and hardware. This paper presents a strategy for enhancing ISAC systems by implementing a semantic-level slicing approach to address key issues as resource utilization, real-time capability, and computational complexity in ISAC systems. A bidirectional mapping mechanism is introduced, combined with edge intelligence, that enhances system predictability, autonomy, and reduces task completion costs by slicing and allocating semantic resources in cloud-edge environments. Additionally, the proposed communication-sensing complementary strategy leverages semantic fusion to enable efficient and adaptable execution of communication and sensing tasks at the resource level, resulting in enhanced task flexibility. The simulation results show that the proposed bidirectional mapping and communication-sensing complementary method significantly improves the signal-to-interference-noise ratio of the traditional system by about 55% in medium and low dynamic environments.