Abstract
Hearables are expected to evolve into high-performance devices offering functions such as voice assistants and vital sign monitoring. As these functions become integrated into daily life, users access more sensitive information, highlighting the need for secure authentication. While existing methods primarily utilize biometric traits, they often suffer from intra-user variability, environmental susceptibility, and concerns about the sharing of biometric information. To address these issues, we propose AudioPIN, a knowledge-based authentication system for hearables leveraging spatial audio. In AudioPIN, users authenticate by inputting a Personalized Audio Sequence (PAS) by selecting sound sources positioned spatially around them. This study focuses on the auditory interface design at the core of AudioPIN. Specifically, we optimized sound sets, spatial arrangements, and selection areas, and evaluated the system using a prototype. Results confirmed that with an arrangement offering over 10,000 combinations, users completed authentication with 91.0% accuracy in 20.6 seconds.
Funding Information
  • Japan Society for the Promotion of Science (JP25K03168,JP23KJ1884)
  • KGRI Challenge Grant (AY2024/AY2025)