David Bethge, Dr.

Research

While I love working on cutting edge technologies, I am most interested in bringing them into products.

I have written over 15 research papers and have over 10 patents. Visit Google Scholar for a complete and up-to-date list.

Publications

HappyRouting: Learning Emotion-Aware Route Trajectories for Scalable In-The-Wild Navigation

David Bethge, Daniel Bulanda, Adam Kozlowski, Thomas Kosch, Albrecht Schmidt, Tobias Grosse-Puppendahl

arxiv, 2024

Paper

Novel navigation algorithm that finds the "happy" route. Using contextual information and machine learning to predict most likely happy route elements everywhere in the world.

Machine learning systems for human emotional states

David Bethge

Dissertation / PhD thesis, 2023

PhD thesis

This doctoral thesis explores various machine learning systems for understanding and predicting human emotional states using diverse sensor data, with a focus on real-world applications in driving contexts and physiological signals.

Technical design space analysis for unobtrusive driver emotion assessment using multi-domain context

David Bethge, Luis Falconeri Coelho, Thomas Kosch, Satiyabooshan Murugaboopathy, Ulrich von Zadow, Albrecht Schmidt, Tobias Grosse-Puppendahl

Ubicomp / IMWUT, 2023

Paper

Research explores non-intrusive prediction of driver emotions through contextual smartphone data, outperforming facial recognition by 7% in a study of 27 participants.

Interpretable Time-Dependent Convolutional Emotion Recognition with Contextual Data Streams

David Bethge, Constantin Patsch, Philipp Hallgarten, Thomas Kosch

CHI EA, 2023

Paper

Convolution-based neural network for emotion classification with interpretable time- and feature-aware model decisions, tested on a real-world driving dataset.

TS-MoCo: Time-Series Momentum Contrast for Self-Supervised Physiological Representation Learning

Philipp Hallgarten, David Bethge, Ozan Özdenizci, Tobias Grosse-Puppendahl, Enkelejda Kasneci

EUSIPCO, 2023

Paper/ Code

Self-supervised learning framework based on a transformer architecture for unlabeled physiological time-series, efficient for domain-agnostic systems in biomedical applications.

EEG2Vec: Learning Affective EEG Representations via Variational Autoencoders

David Bethge, Philipp Hallgarten, Tobias Grosse-Puppendahl, Mohamed Kari, Lewis L. Chuang, Ozan Özdenizci, Albrecht Schmidt

SMC, 2022

Paper

End-to-end representation learning framework for modeling user-specific affective representation from raw EEG.

Domain-Invariant Representation Learning from EEG with Private Encoders

David Bethge, Philipp Hallgarten, Tobias Grosse-Puppendahl, Mohamed Kari, Ralf Mikut, Albrecht Schmidt, Ozan Özdenizci

ICASSP, 2022

Paper

Multi-source deep learning network that is able to learn domain-invariant latent representation from multiple data-specific private encoders.

Exploiting Multiple EEG Data Domains with Adversarial Learning

David Bethge, Philipp Hallgarten, Ozan Özdenizci, Ralf Mikut, Albrecht Schmidt, Tobias Grosse-Puppendahl

EMBC, 2022

Paper / Code

Enabling multi-source learning for EEG-based brain-computer interfaces via adversarial representation learning.

VEmotion: Using Driving Context for Indirect Emotion Prediction in Real-Time

David Bethge, Thomas Kosch, Tobias Grosse-Puppendahl, Lewis L. Chuang, Mohamed Kari, Alexander Jagaciak, Albrecht Schmidt

UIST, 2021

Paper / Code

Remote sensing system that analyzes traffic dynamics, environmental factors, in-vehicle context, and road characteristics to implicitly classify driver emotions.

SoundsRide: Affordance-Synchronized Music Mixing for In-Car Audio Augmented Reality

Mohamed Kari, Tobias Grosse-Puppendahl, Alexander Jagaciak, David Bethge, Reinhard Schütte, Christian Holz

UIST, 2021 (Best Paper Award)

Paper

In-car audio augmented reality system that mixes music in real-time synchronized with sound affordances along the ride.

TransforMR: Pose-Aware Object Substitution for Composing Alternate Mixed Realities

Mohamed Kari, Tobias Grosse-Puppendahl, Luis Falconeri Coelho, Andreas Fender, David Bethge, Reinhard Schütte, Christian Holz

ISMAR, 2021

Paper

Mixed reality system for mobile devices that performs 3D-pose-aware object substitution to create meaningful mixed reality scenes.

HMInference: Inferring Multimodal HMI Interactions in Automotive Screens

Jannik Wolf, Marco Wiedner, Mohamed Kari, David Bethge

AutoUI, 2021

Paper

System to predict interactions in the car HMI based on contextual CAN-BUS data.

Prognostication of Neurological Recovery by Analyzing Structural Breaks in EEG Data

David Bethge, Jieshi Chen, Oliver Grothe, Jonathan Elmer, Artur Dubrawski

ICDM Workshop, 2019

Paper

Unsupervised, multivariate yet interpretable structural break testing for prognostication of neurological recovery after cardiac arrest.

Patents

• P1: DE/102022120257A1, “Verfahren, Wearable, Vorrichtung und Fahrzeug zur Emotionalisierung eines Fahrererlebnisses im Fahrzeug” (2024)
• P2: US/20240027215A1, “Computer-implemented method for determining a navigation route” (2024)
• P3: DE/102022113585A1, “Computerimplementiertes Verfahren zum Verbessern einer Performance einer Cockpit-Nutzerschnittstelle“(2023)
• P4: DE/102022110349A1, “Verfahren und Vorrichtung zur Überwachung von Objekten” (2023)
• P5: DE/102022106797B3, “Verfahren zur automatischen Einstellung zumindest eines Rückspiegels eines Kraftfahrzeugs” (2023)
• P6: DE/102022106812B4, “Computerimplementiertes Verfahren zur Ermittlung eines Emotionszustandes einer Person in einem Kraftfahrzeug” (2024)
• P7: DE/102022104325A1, “Verfahren, System und Computerprogrammprodukt zur Erstellung einer Datenstruktur für auf künstlicher Intelligenz basierende Projekte” (2023)
• P8: DE102022104322A1, “Vorrichtung und Verfahren zur Fahrerassistenz für ein Kraftfahrzeug, die Vorrichtung umfassendes Kraftfahrzeug” (2023)
• P9: DE/102021130939B3, “Vorrichtungen und Verfahren zur gemeinsamen Routenführung” (2023)
• P10: US/20230147024A1, “Method and system for robust identification of a vehicle occupant” (2023)
• P11: DE/102021129108A1, “Verfahren, System und Computerprogrammprodukt zum Ausgeben von Ausgabewerten zur Analyse und/oder Bewertung und/oder Prozesssteuerung einer Entität” (2023)
• P12: US/20230146013A1, “Method for producing a model for automated prediction of interactions of a user with a user interface of a motor vehicle” (2023)
• P13: DE/102021129094A1, “Verfahren zum Ermitteln von Ersatzteilbedarf” (2023)
• P14: US/20230020786A1 “System for a motor vehicle and method for assessing the emotions of a driver of a motor vehicle” (2023)
• P15: DE/102021116641A1, “Verfahren, System und Computerprogramm zur interaktiven Auswahl und Wiedergabe von in Echtzeit erzeugten Audio- und/oder Videosequenzen in einem Kraftfahrzeug” (2022)
• P16: DE/102021110268A1, “Verfahren und System zur szenensynchronen Auswahl und Wiedergabe von Audiosequenzen für ein Kraftfahrzeug” (2022)

Talks / Lectures / Podcasts

I am regularly giving talks and keynotes about technology in general. Feel free to reach out.

AI-Driven DoA and Interference Mitigation on a 4T4R Radar SoC

Webinar @ IWPC: The International Wireless Industry Consortium

October 22, 2025

Recording Link

Radar AI - Deep Neural Networks for Enhanced Radar Point Cloud Generation

Webinar @ IWPC: The International Wireless Industry Consortium

November 13, 2024

Recording Link

Data Science and Machine Learning Fundamentals

Cooperative State University Stuttgart / Karlsruhe (DHBW)

Visiting Lecturer: 2018/2019/2020/2021

Lecture Materials