Utility-Efficient Edge-Cloud Collaborative Space-Air-Ground Integrated Internet of Things

As the demand for tasks requiring low latency and high computational intensity increases, interest in the study of integrated space-air-ground networks (SAGINs) is growing. Meanwhile, the multi-layer task data processing for Internet of thing (IoT) devices continues to be a topic worthy of investigation. In this paper, we propose a multi-layer data processing method in edge-cloud collaborative SAGINs for IoT systems. The IoT devices may select the most suitable unmanned aerial vehicle, and low earth orbit satellite with the optimized data splitting portion to assist data processing based on their residual energy, task latency requirements, communication conditions, and multi-layer computing capabilities. Based on the system model, we formulate an utility efficiency maximization problem. Due to the non-convex nature of this problem and the coupling between variables, we introduce a deep reinforcement learning-based method to acquire multi-layer data processing decisions. Simulation results validate that our proposed method exhibits effectiveness and superiority compared to other benchmarks.