AI in SAGIN: Building Deep Learning Service-Oriented Space-Air-Ground Integrated Networks
He Li, Kaoru Ota, Mianxiong Dong
Abstract
In next-generation mobile communications, space-air-ground integrated networks (SAGINs) is an emerging infrastructure in future wireless access networks. Since artificial intelligence (AI) applications become more and more important, it is essential to build a deep learning service-oriented SAGINs. In this article, we present a hierarchical intelligent computing structure focusing on processing deep learning tasks in future SAGINs. An optimization strategy is also proposed to improve the quality-of-service (QoS) of deep learning tasks in the proposed structure. We test our work in small testbed and simulations. The evaluation results show that the proposed work outperforms other offloading strategies in a SAGIN environment.