"Simulated CSI Dataset for Robust Vehicular Federated Learning via AirComp"
收藏DataCite Commons2026-03-27 更新2026-05-03 收录
下载链接：
https://ieee-dataport.org/documents/simulated-csi-dataset-robust-vehicular-federated-learning-aircomp
下载链接
链接失效反馈官方服务：
资源简介：
"This dataset contains simulated high-mobility channel traces for evaluating robust Over-the-Air Computation (AirComp) in Vehicular Federated Learning (VFL). Due to severe Doppler shifts in V2X networks, perfect Channel State Information (CSI) is unattainable. To model this, the dataset leverages the 3GPP TR 37.885 path loss model, Jakes Doppler spectrum, and AR(1) channel aging to simulate a 5.9 GHz highway scenario with varying vehicle densities and speeds (e.g., 110 km\/h).The provided data structures include true instantaneous channels, nominal channel estimates, nominal error covariance matrices, and corresponding Frobenius-norm uncertainty bounds. It is specifically designed to facilitate research on worst-case robust transceiver design, beamforming optimization, and collaborative intelligence aggregation under composite CSI uncertainty."
提供机构：
IEEE DataPort
创建时间：
2026-03-27
