FDD大规模MIMO信道反馈实测数据集

本数据集具有三个数据文件，其中FDD大规模MIMO接收机上下行接收信号为典型场景下32发射天线单接收天线的上下行接收信号数据集；FDD大规模MIMO信道数据集为典型场景下32发射天线单接收天线的上下行信道状态信息数据集；基于AI的FDD大规模MIMO信道反馈数据集为典型场景下32发射天线单接收天线的上下行信道反馈数据集。数据类型为文本文件和数组文件，数据格式为txt和mat
采集时间与地点：2025年6月，电子科技大学清水河校区。
FDD大规模MIMO接收机上下行接收信号数据集采集的射频段，天线接收信号经射频板下变频至440MHz中频，通过增益放大器获取接收增益；在基带端，经过的处理步骤为时域同步模块实现符号定时与帧同步、加窗处理配合匹配滤波器完成波形整形、采样率转换模块将信号降采样至30.72MHz基带速率、频偏估计与补偿单元完成频偏校正、FFT模块将时域信号转换为频域资源网格并保存为txt文件；FDD大规模MIMO信道数据集采集使用个人电脑DESKTOP-D411OKA，使用过软件MATLAB，版本号MATLAB2019b完成了数据处理；基于AI的FDD大规模MIMO信道反馈数据集采集使用个人电脑DESKTOP-D411OKA，使用过软件Pycharm完成了数据分析。FDD大规模MIMO接收机上下行接收信号通过MATLAB分析数据，版本为2019b，具体采用了信道估计算法估计导频信号的信道，并通过MATLAB对信道进行SVD分解；FDD大规模MIMO信道数据集通过MATLAB分析数据，版本为2019b，具体采用了信道估计算法估计导频信号的信道，并通过SVD分解将原始信道数据转化为角度延时域，并保存为信道数据；基于AI的FDD大规模MIMO信道反馈数据集通过Pycharm分析数据，版本为2021.1.1。
数据量：407.4MB

This dataset includes three data files. The FDD massive MIMO transceiver uplink and downlink received signal dataset is a dataset of uplink and downlink received signals for a 32-transmit-antenna, single-receive-antenna scenario under typical operating conditions; the FDD massive MIMO channel dataset is a dataset of uplink and downlink channel state information (CSI) for a 32-transmit-antenna, single-receive-antenna scenario under typical operating conditions; the AI-based FDD massive MIMO channel feedback dataset is a dataset of uplink and downlink channel feedback for a 32-transmit-antenna, single-receive-antenna scenario under typical operating conditions.

The data types are text files and array files, with formats .txt and .mat.

Collection time and location: June 2025, Qingshuihe Campus, University of Electronic Science and Technology of China.

For the FDD massive MIMO transceiver uplink and downlink received signal dataset: during the radio frequency band collection phase, the antenna received signal was downconverted to a 440 MHz intermediate frequency (IF) via an RF board, and receiving gain was obtained through a gain amplifier. At the baseband stage, the processing steps are as follows: symbol timing and frame synchronization are implemented via the time-domain synchronization module; waveform shaping is completed via windowing combined with a matched filter; the signal is downsampled to a 30.72 MHz baseband rate via the sampling rate conversion module; frequency offset correction is performed via the frequency offset estimation and compensation unit; the time-domain signal is converted to a frequency-domain resource grid via the FFT module, which is then saved as a .txt file.

The FDD massive MIMO channel dataset was collected using the personal computer DESKTOP-D411OKA, with data processing completed using MATLAB (version 2019b).

The AI-based FDD massive MIMO channel feedback dataset was collected using the personal computer DESKTOP-D411OKA, with data analysis completed using PyCharm.

For the FDD massive MIMO transceiver uplink and downlink received signal dataset, data analysis was conducted using MATLAB (version 2019b). Specifically, channel estimation algorithms were adopted to estimate the channel of pilot signals, and singular value decomposition (SVD) was performed on the estimated channel via MATLAB.

For the FDD massive MIMO channel dataset, data analysis was conducted using MATLAB (version 2019b). Specifically, channel estimation algorithms were adopted to estimate the channel of pilot signals, and the original channel data was converted to the angle-delay domain via SVD decomposition before being saved as channel data files.

For the AI-based FDD massive MIMO channel feedback dataset, data analysis was conducted using PyCharm (version 2021.1.1).

Total data volume: 407.4 MB

搜集汇总

背景与挑战

背景概述

该数据集为FDD大规模MIMO信道反馈的实测数据，包含三个数据文件，涉及32发射天线单接收天线的上下行接收信号、信道状态信息及基于AI的信道反馈，数据格式为txt和mat，总数据量为407.4MB。