A LEO Satellite Network QoS Routing Scheme Based on Reinforcement-Learning

Abstract

Low earth orbit (LEO) satellite communication networks will provide a huge coverage network to carry the global communication traffic. However, the huge architecture brings a great challenge to the QoS routing. This paper integrates the inter-satellite transmission delay, the queuing delay and the jitter to design QoS factor, then, establishes a minimum cost function. Based on this model, this paper proposes a Q-Q routing scheme with Q-routing algorithm. Simulations show that Q-Q has a smaller routing cost and shorter delay compared with Dijkstra’s algorithm.