Proposal of a new PUF based on sensors for the identification of IoT smart mobile devices
Authors: Raúl Aparicio-Téllez (r.aparicio@unizar.es), Jorge Fernandez-Aragon (jorgefusti96@gmail.com), Abel Naya-Forcano (abeln@unizar.es) and Guillermo Diez-Señorans (gds@unizar.es)
Supervisors: Miguel Garcia-Bosque (mgbosque@unizar.es), Santiago Celma (scelma@unizar.es)
Group of Electronic Design (GDE), I3A, University of Zaragoza, Pedro Cerbuna 12, Zaragoza, Spain
Links
](https://github.com/abeln94/GDE-puf/blob/main/2023_CSAW_Acelerometros.pdf) [Conference poster ](https://github.com/abeln94/GDE-puf/blob/main/Poster_CSAW_23.pdf) [Play store app 
](https://play.google.com/store/apps/details?id=es.unizar.gde.puf)
Abstract
The increasing development of the Internet of Things (IoT) has suppose the exchange of large amounts of information among devices, that must be protected against cyberattacks. In this context, Physically Unclonable Functions (PUFs) arise as an optimal solution to this problem, as they exploit the stochastic variations occurring during the manufacturing process of the devices to generate a binary sequence that uniquely identifies each device. In this work, we propose a new PUF based on the signal measured by several sensors such as the accelerometer or the gyroscope of different smart mobile devices. This new proposal, designed exclusively for the participation in this competition, is especially interesting since using the sensors of the mobile devices does not require the implementation of additional electronic circuits, making the proposed PUF already integrated into an IoT system. The results obtained show that this PUF can be used for identification and authentication purposes, demonstrating that our solution can efficiently protect an IoT system.
Keywords
Authentication, hardware security, identification, internet of things, physically unclonable function.
Key ideas
Own-developed app to measure and obtain the binary key.
-
Justification
- Sensors of smart mobile devices provide an inherent non-zero acceleration, angular velocity and noise level when stationary.
-
Proposal
- Use the non-zero measured parameters obtained to generate a unique fingerprint.
-
PUF
- Results provide a valid identification and authentication system. Already integrated in IoT system.