A Four-Dimensional Matrix Pencil-Based Channel Prediction Method for FTTR C-WAN in Mobile Environments

This paper investigates the problem of channel aging in the fiber-to-the-room (FTTR) centralized Wi-Fi access network (C-WAN) architecture in mobile scenarios. We propose a novel channel prediction method based on the fourdimensional matrix pencil (FDMP), leveraging the angular-timefrequency domain structure of broadband channels. The method achieves high-precision estimation of key channel parameters by exploiting the superior predictive capabilities of the matrix pencil (MP) framework. Specifically, by utilizing the multiple antennas at the transmitter and receiver, we construct a 4-D MP matrix from channel sample data in the angular-time-frequency domain. This matrix enables an accurate estimation of the angle of arrival (AoA), delay, and Doppler shift, which are critical for effective channel prediction. Simulation results demonstrate that our proposed FDMP method offers a robust solution to channel aging in FTTR C-WAN systems, and the impact of station (STA) mobility within a specific speed range on channel prediction can be effectively mitigated with only two time-domain samples.