Semantic Coding NOMA Downlink System for Satellite-Integrated Internet under AoI Minimization

Zhou Xu,Jian Jiao,Heqing Zhang,Lirong An,Ye Wang,Qinyu Zhang
DOI: https://doi.org/10.1109/iccc62479.2024.10682027
2024-01-01
Abstract:Satellite-integrated Internet is regarded to provide pervasive intelligent services for ubiquitous terrestrial user equipments (UEs) in the 5G advanced (5GA) and future 6 G network. To timely transmit goal-relevant data for the terrestrial UEs, we propose a semantic coding non-orthogonal multiple access (NOMA) downlink system for satellite-integrated Internet to deliver semantic information in this paper. First, we introduce a goal-relevant semantic coding system for the remote monitoring in rural and extreme environment. Then, we formulate an age of information (AoI) minimization problem subject to power constraints on satellite and model it as a Markov decision process (MDP). Furthermore, we propose an optimal power allocation policy for AoI leveraging double-network deep reinforcement learning for the above non-convex AoI minimization problem, and further reduce the AoI by utilizing the long short-term memory (LSTM) to capitalize on the high temporal correlation of UEs’ AoI. Simulation results demonstrate the superiority of our proposed scheme over the benchmark schemes.
What problem does this paper attempt to address?