DRL-Based Load-Balancing Routing Scheme for 6G Space–Air–Ground Integrated Networks

Due to the rapid development of space–air–ground integrated network (SAGIN), a satellite communication system has advantages wide coverage and low requirements for geographical environment is gradually becoming main competitive technology 6G. The low-earth-orbit (LEO) characteristics transmission delay, small propagation loss, global coverage, its exploration become research object contemporary communications. However, traditional routing algorithms cannot adapt high dynamics load-balancing LEO networks. In this paper, algorithm satellites based on Deep Q-Network (DQN-LLRA) proposed by using deep reinforcement learning. Making use model obtained DQN training, nodes can select best results according bandwidth, queue utilization surrounding nodes. simulation analysis show that path load low. Compared with Q-learning-based algorithm, reduces maximum rate 5%, average 13%, effectively balances in network.