基于空天地网络的卫星边缘计算卸载方案研究

Research on Satellite Edge Computing Offloading Scheme Based on Space-Air-Ground Integrated Network

空天地一体化网络以其广覆盖、大容量、不受地域影响等特点,可支持实现全球无缝通信,满足丰富的业务需求。而基于空天地一体化网络的卫星边缘计算,与地面边缘计算协同完成泛在的边缘计算服务,在提高业务可靠性的同时也可满足时延敏感业务的需求。针对空天地一体化网络下卫星边缘计算技术应用需求的分析,对基于空天地一体化网络架构的典型计算卸载方法及其优缺点进行了对比分析,特别对基于强化学习的计算卸载方案进行了具体研究,并给出其卸载实现流程。最后对未来卫星边缘计算研究需面对的挑战进行了探讨,为后续研究提供参考。

With its wide coverage,large capacity,and unaffected geographical characteristics,the air-space-ground integrated nctwork can support scamless global communication and mcet rich business nceds.Satellite edge computing based on the space-air-ground integrated network works with terrestrial edge computing to complete ubiquitous edge computing services,which can improve service reliability and mcct the nceds of latency-sensitive services.In order to analyze the application requirements of satellite edge computing technology under the space-air-ground integrated network,the typical computing offloading methods based on the space-air-ground integrated network architecture and their advantages and disadvantages are compared and analyzed,especially the computing offloading scheme based on reinforcement learning is studied,and the implementation process of the offloading is given.Finally,the challenges to be faced by future satellite cdge computing rescarch are discussed,which provides reference for follow-up research.