Research On Edge Computing Architecture And Task Strategy Of Satellite Communication Network

Satellite Integrated Networks(Satellite-Terrestrial Integrated Networks,STIN)through the effective integration of Satellite and ground communication network respective advantages,can expand network coverage,improve network flexibility and enhance the utilization of network resources.However,the existing satellite-earth integration network architecture still cannot have good network performance,which is mainly due to some problems and challenges in satellite communication network.Specifically,due to the limited satellite resources,it lacks dynamic and flexible on-board task processing and ground-based terminal equipment unloading schemes.Therefore,the research work of this paper is mainly carried out from the following three aspects:1?Aiming at the space-based part of STIN network,a two-layer space-based edge computing architecture is innovatively proposed based on hierarchical thought,which is specifically divided into two parts: SATELLITE control layer SCL and satellite data access,forwarding and processing layer SDAFPL.SCL is composed of satellite nodes with a lightweight space-based core network architecture,while SDAFPL is composed of satellite nodes with base station functions and integrated on-board edge computing platform.And the space-based core network in SCL layer is designed in detail.At the same time,three diverting methods are introduced to describe how SCL layer manages data from different terminal devices in SDAFPL layer.In the ground part of STIN network,a new network element for space-based control in the ground 5G core network--ground control satellite group network element is designed.Combined with the characteristics of this paper,the control function of the network element is specifically positioned in the scene of satellite-ground cooperative task unloading,and the processing process of satellite-ground cooperative task unloading is designed.2?In the space-based part,aiming at the problem of limited on-board computing capacity in SDAFPL layer,a multi-SECS multi-computing task computing model is established based on inter-satellite link.The delay and energy consumption of processing tasks on the current satellite node and that of processing tasks on other satellite nodes are considered in a unified manner to minimize system overhead.Model the assignment problem of multiple SECS and multiple computing tasks.Then,combining the Hungarian algorithm with the assignment problem,and improving the cost matrix of the Hungarian algorithm,a task assignment algorithm for satellite cooperative edge computing is proposed.Compared with other algorithms,it is proved that the algorithm can reduce the overall delay and energy consumption of the system through task migration.3?Aiming at the satellite to earth task unloading scenario,a multi-terminal device multi-channel unloading model of the on-board edge computing platform was established,and the system overhead of the terminal device under different unloading decisions was obtained.Then,based on game theory,the system model is transformed into a multi-participant non-cooperative model,and a double-cycle distributed task unloading algorithm is proposed,and the optimal unloading strategy set of all terminal devices in the system is obtained.Compared with other algorithms,it is proved that the algorithm has a certain stability and can reduce the system overhead in the unloading process of computing tasks.