An Ubiquitous Two-factor Authentication (2FA) Platform for IoT Applications

IoT-enabled smart wearables have become ubiquitous, offering continuous sensing and real-time data exchange through embedded sensors. Utilizing these capabilities, we propose a robust two-factor authentication (2FA) system that combines fingerprint and ear-canal-based biometrics. The fingerprint modality uses a capacitive sensor to address challenges posed by wet or oily fingers, while the ear-canal modality exploits the unique anatomical and acoustic properties of the ear-canal captured using standard earphones with built-in microphones. The system was evaluated with 30 participants under varied usage conditions. An inertial measurement unit (IMU) sensor placed near the earphone established a reference orientation during enrollment. Identity matching performed using the sum of squared distances (SSD) achieved a 100% recall for earphone orientation changes up to ±25°. Our system recorded a best area under the curve (AUC) of 99.68% and an equal error rate (EER) of 0.7% for authentication. Offset and drift correction were applied to the IMU data to ensure stable and reliable performance over time, which resulted in a recall improvement from 97% to 98.4% over one month, highlighting the system's robustness.