Abstract: LEO satellite communication has the advantages of long transmission distance,wide coverage,and is not restricted by terrain.It has become an important communication method for the civil aviation transportation and general aviation industries.However,the low-orbit satellite network is a highly heterogeneous and dynamic environment.The mobility of satellite nodes,the complexity of communication links,uneven spatial and temporal distribution of aircraft,and the coexistence of multiple services make task offloading and resource allocation face many challenges.To this end,this paper proposes an aircraft task offloading method based on DDRL,with the purpose of maximizing the overall effectiveness of the system.Firstly,the system utility maximization problem is modeled as a joint optimization problem of task offloading and resource allocation,taking into account the computing power and coverage time of LEO satellites.Next,the problem is transformed into a Markov decision process,using DDQN algorithm to learn the optimal task offloading decision,and based on this,TD3 is used to obtain the optimal resource allocation strategy.Simulation experiments show that under different computing resources and communication resources,the proposed scheme is better than other benchmark schemes in terms of system utility,proving the usability of the proposed framework.

Key words: LEO satellite network, Task offloading, Resource allocation, Deep reinforcement learning