BA/MA: Multimodal Authentification with Earables

Background

Earables are earbuds that are equipped with multiple sensors. Authentication is one of the four big application areas of earables as defined by Röddiger et al. (2022). Although it has been the smallest when the review was published, there was a surge in publications thereafter. This points to the significance and grown interest in authentication.

However, most existing approaches employ only unimodal sensing, thus relying on data from a single sensor. With OpenEarable 2.0 (Röddiger et al., 2025), which provides extensive sensing capabilities, new opportunities emerge for multimodal authentication. The goal of this thesis is to explore which sensing modalities of OpenEarable 2.0 can be leveraged for user authentication, and how combining multiple signals can improve robustness and performance.

Tasks

  1. Literature Review: Analyze the current state of the art in earable-based authentication. Identify at least two apt sensing approaches that have been employed (or propose novel ones).
  2. Feasibility Analysis: Evaluate how these sensing approaches can be implemented using OpenEarable 2.0’s hardware.
  3. Data Collection: Design and conduct a data collection study with multiple participants to record multimodal authentication data in base and additional contexts (for MA: more and additionally naturalistic settings required).
  4. Data Analysis: Evaluate the performance of each sensing modality individually and in combination.
    • For BA: Focus on lightweight algorithms suitable for running on a companion phone or directly on the earable.
    • For MA: Focus both on a high-performance (“full compute”) solution and a lightweight version deployable on OpenEarable 2.0.
  5. Deployment (MA only): Deploy the final algorithm to OpenEarable 2.0 to demonstrate real-time authentication.

Requirements

  • Interest in Human-Computer-Interaction (HCI) and the real-world application of new devices
  • Good Python skills (for data analysis)
  • Optional
    • Good C/C++ skills, Zephyr OS (for programming OpenEarable 2.0). Alternatively, there is the option to use a prototyped firmware offering a micropython runtime
    • Flutter (for adapting the OpenWearables App)

If you are interested in this topic, please contact Jonas Hummel (hummel@teco.edu).