Pairing Wi-Fi and Bluetooth MAC addresses through passive packet capture

The majority of smart devices used nowadays (e.g., smartphones, laptops, tablets) is capable of both Wi-Fi and Bluetooth wireless communications. Both network interfaces are identified by a unique 48-bits MAC address, assigned during the manufacturing process and unique worldwide. Such addresses, fundamental for link-layer communications and contained in every frame transmitted by the device, can be easily collected through packet sniffing and later used to perform higher level analysis tasks (user tracking, crowd density estimation, etc.). In this work we propose a system to pair the Wi-Fi and Bluetooth MAC addresses belonging to a physical unique device, starting from packets captured through a network of wireless sniffers. We propose several algorithms to perform such a pairing and we evaluate their performance through experiments in a controlled scenario. We show that the proposed algorithms can pair the MAC addresses with good accuracy. The findings of this paper may be useful to improve the precision of indoor localization and crowd density estimation systems and open some questions on the privacy issues of Wi-Fi and Bluetooth enabled devices.