Physiological sensing is currently one of the most important application areas for earable computing (Röddiger et al, 2022; Hu et al, 2025). Among the physiological signals that can be captured using ear-worn devices, heart rate is of particular interest due to its relevance for health monitoring, fitness applications, and biosignal-adaptive systems. Photoplethysmography (PPG), a sensing modality already widely adopted in smartwatches and fitness trackers (Nazarian et al., 2021; Sibomana et al., 2025), is especially promising for earables because of the rich vascularization of the ear region. Consequently, PPG has been employed in various earable systems for heart-rate estimation (Choi et al., 2022, Du et al., 2025) and has recently been integrated into state-of-the-art earable platforms such as Omnibuds (Montanari et al., 2024) and OpenEarable 2.0 (Röddiger et al., 2025).

While heart-rate estimation from in-ear PPG signals is generally feasible, reliable measurements during everyday activities remain challenging. In particular, motion artifacts caused by head movements, speaking, walking, running, or cycling can substantially distort the optical signal and lead to inaccurate heart-rate estimates (e.g., Zhang et al., Pollreisz et al., 2019). Although a wide range of artifact-reduction approaches have been proposed for wrist-worn devices and other wearable sensors, solutions developed for one device often do not transfer directly to another due to differences in sensor placement, hardware characteristics, and movement patterns.

The goal of this thesis is therefore to investigate how the multimodal sensing capabilities of OpenEarable 2.0 can be leveraged to improve heart-rate estimation under motion, with the goal of compensating for motion artifacts in in-ear PPG in real time.

1. Literature Review: Analyze and present the state-of-the-art in motion artifact detection and removal for PPG-based heart-rate estimation, covering approaches for wearable systems in general and earable systems in particular.
2. Data Collection: Design and conduct a user study that captures in-ear PPG data under different motion conditions, including activities such as speaking, walking, running, and cycling.
3. Algorithm Development: Develop at least one novel algorithmic approach for motion artifact removal and heart-rate estimation from in-ear PPG signals. The developed solution should be capable of real-time execution, either directly on OpenEarable 2.0 or on an accompanying smartphone. Particular attention should be paid to exploiting the multimodal sensing capabilities of OpenEarable 2.0 (for orientation, see Hummel & Burzer et al. (2026)).
4. Evaluation and Benchmarking: Compare the developed approach against an established baseline. The evaluation should be conducted both on the collected dataset and on publicly available datasets (e.g., Montanari et al., 2023) to assess generalizability and robustness.
5. Real-World Validation: Deploy the selected algorithm(s) in realistic day-to-day scenarios and evaluate their practical applicability and performance under everyday conditions.
6. Bonus: Investigate how reliably heart-rate variability (HRV) related metrics can be inferred after motion artifact removal.

• Interest in wearable, physiological, and multimodal sensing systems.
• Ability to build upon and extend existing approaches from the literature to develop novel solutions.
• Designing and conducting a user-study.
• Strong skills in Python (for signal processing and data analysis).
• Solid skills in Flutter (for building an in-app in OpenWearables).
• Optional: Experience with ZephyrOS (for firmware-level adaptations on OpenEarable 2.0).

• A paragraph explaining your motivation.
• Your study program (Bachelor/Master), current semester, and field of study.
• A transcript of records (courses and grades).
• Your programming experience.
• Any areas of interest relevant to the topic.
• Your CV (if available)