A dataset of barometric readings for enhancing security and privacy of IoT

The security and privacy of wireless channels is typically enforced by leveraging cryptographic tools. However, there are scenarios where these methods are unfit, such as in resource-constrained environments, i.e., Internet of Things (IoT), or when an extra layer of security is needed. A promising solution involves correlating air pressure (barometric) readings to securely pair IoT devices while requiring zero-interaction. This paper presents an experimental dataset of real-world barometric measurements collected in open areas under different weather conditions. Specifically, our dataset includes readings recorded using the reference hardware platform BMP280. The experiments involve a reference scenario constituted by three Adafruit BMP280 barometric sensors connected to a Raspberry Pi 3 Model B board to collect barometric measurements. The three sensors represent two communicating parties (Alice and Bob) and an adversary (Eve), respectively. The dataset is constituted by three experiments characterized by different relative distances among Alice, Bob, and Eve. We considered 5cm and 2m between Alice and Bob while placing Eve at 2m and 8 meters, respectively. The second configuration, i.e., (Alice-Bob at 2m and Eve at 8m) has been replicated in a different scenario characterized by less air pressure fluctuations. The sampling frequency has been set to 70Hz while the measurements last for 50, 24 and 41 hours, respectively. Researchers can use this dataset in several ways, including: (i) Study the air pressure variation and correlation between devices separated by different distances, (ii) Develop a co-location verification extension for the Diffie-Hellman (DH) key agreement method that utilizes air pressure data streams, (iii) Study possible attacks against proximity-based authentication techniques that depend on pressure correlated variations.