Spectral Temporal Graph Neural Network for Massive MIMO CSI Prediction
Sharan Mourya, M. Pavan Reddy, SaiDhiraj Amuru, Kiran Kuchi
Abstract
In the realm of 5G communication systems, the accuracy of Channel State Information (CSI) prediction is vital for optimizing performance. This letter introduces a pioneering approach: the Spectral-Temporal Graph Neural Network (STEM GNN), which fuses spatial relationships and temporal dynamics of the wireless channel using the Graph Fourier Transform (GFT). We compare the STEM GNN approach with conventional Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM) and Transformer models for CSI prediction. Our findings reveal a significant enhancement in overall communication system performance through STEM GNNs. For instance, in one scenario, STEM GNN achieved a spectral efficiency of 4.683 bps/Hz which is 16.5% higher than that of a transformer, 63% higher than an LSTM and 198% higher than that of an RNN. The spectral-temporal analysis capabilities of STEM GNNs capture intricate patterns often overlooked by traditional models, offering improvements in beamforming and interference mitigation.