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.