To mitigate the performance degradation caused by channel aging in massive multi-input multi-output (MIMO) systems, channel prediction is investigated in this paper. Based on the existing vector Prony method (VPM) and the Prony-based angular-delay domain (PAD) prediction, two novel channel prediction methods, referred to as the modified VPM (MVPM) and the modified PAD (MPAD), are proposed. In the proposed methods, we decouple the model size from the number of past channel estimates that are involved in the prediction of the future channels, allowing more flexible usage of channel estimates. Simulations demonstrate that when the number of past channel estimates becomes large, the proposed MVPM and MPAD significantly outperform VPM and PAD, respectively. Complexity analysis shows that this improvement in performance comes with a slight increase in computational complexity.
Novel Prony-Based Channel Prediction Methods for Time-Varying Massive MIMO Channels
24.06.2024
585157 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch