Sparse Channel Estimation Aided Federated Learning Assisted Non-Terrestral Network
Hongming Zhang, Xi Meng, Fei Qi, Yuqing Chai, Pengpeng Li, Xiaoyong Liu · 2024
The space-air-ground-sea integrated network is expected to provide seamless data services in 6G. As an important enabler, non-terrestrial network (NTN) should be carefully designed so that a satisfactory quality of experience service can be provided for ground users and maritime users. In this paper, a sparse channel estimation aided federated learning (FL) assisted NTN is proposed for 6G. Firstly, a sparse channel estimation scheme is proposed for providing accurate channel estimation results both in uplink model parameter transmission and in downlink model parameter transmission for FL training. Then, the system performance of the FL assisted NTN is evaluated in terms of spectral efficiency and energy efficiency. Finally, simulation results are provided for evaluating the attainable system performance of the proposed scheme, showing that the proposed FL assisted NTN is capable of attaining a limited level of co-channel interference performance. Furthermore, a better system performance can be attained by the proposed sparse channel estimation aided FL assisted NTN system, in comparison to the erroneous counterparts.