CommRad RF: A dataset of communication radio signals for detection, identification and classification

In this age of information, data is the most valuable commodity. In today's world it is easier to acquire millions of standard/known and even non- standard/unknown RF signals. Perhaps, having such huge amount of data can help people to develop novel deep learning techniques. This dataset focuses on communication radios baseband signals. It includes RF signals from multiple radios using different frequencies, in different environments. Over 2700 RF signals from 27 different radios are included in the dataset. An RF receiver was used to capture the dataset, which automatically detects RF signals and records them until they disappear. In order to create a signal library, the receiver is connected to a laptop that pre-processes and stores RF signatures from different emitters.

在信息化时代，数据已然成为最宝贵的商品。当下，获取海量标准化/已知乃至非标准化/未知的射频（Radio Frequency, RF）信号已愈发便捷。此类大规模数据集或可助力研究者开发创新性深度学习技术。本数据集聚焦于通信无线电基带信号，涵盖了不同环境下、不同工作频率的多台通信无线电所产生的射频信号，共包含来自27台不同通信无线电的逾2700条射频信号。数据采集采用射频接收机完成，该设备可自动识别射频信号并持续录制直至信号消失。为构建信号库，该接收机与笔记本电脑相连，由后者完成不同信号发射源的射频特征预处理与存储工作。

搜集汇总

背景与挑战

背景概述

该数据集包含2700多个来自27种不同无线电设备的通信无线电基带信号，覆盖多种频率和环境，适用于检测、识别和分类任务的深度学习研究。数据集通过RF接收器自动捕获并预处理，旨在构建信号库。