Abstract
Cooperative sensing and computing are crucial for high-efficiency task completion in numerous Unmanned Aerial Vehicle (UAV) networks-based sensing applications. However, the frequent data exchange among UAVs strains communication bandwidth resources. This challenge can be further exacerbated in low signal-to-interference-plus-noise ratio (SINR) environments. Semantic communication, by extracting and transmitting task-relevant semantic information rather than raw bit streams, offers a novel approach to mitigate these issues by significantly reducing the transmission data volume while enhancing the system's robustness against noise-induced distortions. In this paper, we propose a cooperative sensing semantic communication framework with a variable semantic compression ratio in the scenario of UAVs-based multi-target cooperative sensing. We then design a resource scheduling strategy in order to efficiently integrate the limited resources of the UAVs to simultaneously enhance sensing accuracy and minimize latency of data processing. The simulation results demonstrate that the sensing accuracy of our semantic communication scheme outperforms the traditional BMP+LDPC+QPSK approach by 85.8 % at 0dB SINR, and our solution achieves an improvement up to 64.5 % in average task effectiveness score and up to 68.2 % reduction in latency, when compared to the non-semantic-assisted baseline.