6G enabled seamless integration of satellite-terrestrial networks

The random traffic distribution of users for Low Earth Orbit (LEO) satellites reduces the chances of securing the envisaged low latency of the Sixth Generation (6G) networks during the short period of connection. The distance from the Earth, Inter-Satellite Links (ISL) routing, and processing time are the main contributors to increasing the latency. In this paper, a deep Q-learning-based decision-making algorithm is proposed to allow proactive trajectory optimization of satellites links between origin and destinations after considering the channel estimation of adjacent satellites. The algorithm deployment is supposed to be done at the 6G Distributed Units (DUs) close to users to utilize the powerful processing capabilities for the route selection by considering the current and future locations of satellites within the constellation. Thus, seamless connectivity can be accomplished in real-time with minimum latency. Obtained simulation results will demonstrate the performance enhancement by the proposed method to achieve longer connections between satellites and UTs.