AI-Based Cloud-Edge-Device Collaboration in 6G Space-Air-Ground Integrated Power IoT
Zhao Wang, Zhenyu Zhou, Hui Zhang, Geng Zhang, Huixia Ding, Ahmed Farouk
Abstract
Space-air-ground integrated communication networks and artificial intelligence (AI) play critical roles in future 6G wireless communication. The Power Internet of Things (PIoT) assisted by the space-air-ground integrated communication network, that is, space-air-ground integrated PIoT (SAG-PIoT), has emerged as a promising solution to provide seamless communication and computing services for PIoT devices with stringent QoS requirements. Cloud-edge-device collaboration is leveraged to enable intelligent resource management in SAG-PIoT. This article proposes an AI-based cloud-edge-device collaboration framework for SAG-PIoT to cope with multi-dimensional resource heterogeneity and network dynamics. Specifically, the SAG-PIoT application scenarios and research challenges are first introduced. Then we develop a hybrid and hierarchical cloud-edge-device collaboration architecture to adapt with different scenarios and domains, followed by the detailed implementation procedures. Next, a queue-aware deep actor-critic (Q-DAC)-based task offloading algorithm is proposed to facilitate decision making optimization under incomplete information. A case study is provided to validate the energy consumption and end-to-end queuing delay performance of Q-DAC.