Abstract: In one embodiment, a device obtains radio frequency (RF) characteristic data for a wireless client. The device inputs the RF characteristic data for the wireless client to a deep learning-based encoder. The device learns a latent space representation of the RF characteristic data from the encoder. The device uses the learned latent space representation as a unique signature to identify the wireless client in a wireless network.

Abstract: Modulated radio frequency (RF) packets are received from a wireless device, and converted to modulated baseband packets. Baseband parameters are derived from the modulated baseband packets. A baseband profile including the baseband parameters is created for the wireless device. A database including baseband profiles of wireless devices is accessed. The baseband profiles are classified under known device types based on baseband parameters included in the baseband profiles. The baseband parameters of the baseband profile are compared to corresponding baseband parameters of the baseband profiles in the database. If the comparing indicates a match between the baseband profile and one of the baseband profiles, the wireless device is classified under the known device type of the one of the baseband profiles, and the baseband profile is stored in the database under the known device type.