This project focuses on using multimodal AI methods to analyze videos of teacher–parent counseling interactions. It combines speech, facial expressions, gaze behavior, posture, and vocal dynamics to automatically evaluate counseling competence and communication quality. The work includes multimodal feature extraction, temporal alignment of behavioral signals, and the development of AI-based feedback systems for teacher training and educational simulations.

Der Brandschutz ist technisch und rechtlich durch Schutzziele und Normen definiert. Kommt es dennoch zu Brandereignissen, spielt Wahrnehmung durch die Bevölkerung eine Maßgebende Rolle. Dabei rücken gesellschaftliche Akzeptanz und Risikotoleranz in den Vordergrund.

This thesis explores the application of SAM3 to mobile video data for object classification and tracking. The focus is on evaluating SAM3’s performance, adaptability, and limitations in real-world mobile scenarios such as dynamic scenes, occlusions, and changing viewpoints.

This thesis aims to extend our current provably safe adaptive RL framework toward autonomous driving scenarios.

You are passionate about applying cutting-edge information technology to solve the energy and climate crisis and would like to work in a vibrant research environment? Then let’s design the energy systems of the future together!

In dieser Arbeit soll untersucht werden, wie sich die Düngung (N) auf die Entwicklung und die Attraktivität (für Insekten) einer Blühwiese auswirkt.

We are looking for Master's students who would like to write their thesis on a topic at the intersection of information systems and healthcare.

Learning and imitating scrub nurses helps context-aware robotic grasping for surgical tasks. Based on hand and object poses from existing dataset, the extracted contact heatmap can help guide efficient robotic grasping and subsequent handover tasks. Evaluation on state-of-art methods on real dataset and robotic grasping experiments can validate the effectiveness of contact map guided grasping approach.

Are you interested in soil properties and how soil supports biodiversity? Then this thesis is for you! We look for a motivated student to join our projects and be part of a soil campaign across different sites in Germany. As part of the project you will be able to use existing data and produce new one to link questions on how land use change affects soils and the biodiversity that inhabits them. The project can be tailored for bachelor, master thesis and mandatory internships.

Development of a optical ultrasound detector for optoacoustic sensing towards non-invasive glucose monitoring.

Blühflächen sind wichtige Habitate zur ökologischen Aufwertung von Agrarlandschaften. Ziel der ausgeschriebenen Arbeit ist, die Auswirkung unterschiedlicher Saatmischungen und Mahdregime auf die vorkommenden Tagfalterarten zu erfassen. Mit Hilfe der jeweils blühenden Pflanzenarten und auf den Flächen vorkommenden Futterpflanzen für Raupen können weitere Rückschlüsse gezogen werden.

The Chair of Robotics, Artificial Intelligence, and Real-Time Systems offers a Master’s thesis focusing on adaptive cooperative perception for ETSI-based Vehicle-to-Infrastructure systems using covariance-aware movement clustering and track-to-track fusion. The thesis investigates how object information from CAM and CPM messages, including covariances, velocities, and object parameters, can improve multi-agent object association and global tracking in urban traffic scenarios.

This thesis investigates the integration of Reinforcement Learning (RL) with end-to-end (E2E) autonomous driving approaches. While E2E methods rely on large amounts of expert data, RL enables learning through interaction in simulation. The goal is to explore how RL-based simulation and feedback mechanisms can improve the robustness and performance of state-of-the-art E2E driving policies.

Solve a 50-year old Mystery! Why do jewel wasp females say “yes” only when males nod on their antennae with open mouthparts? Join us to uncover the hidden pheromone behind wasp courtship using cutting-edge analytical chemical detection methods and behavioral assays. No chemistry experience needed—just curiosity and precision! Flexible start, thesis/internship options.

This thesis focuses on using mobile LiDAR and mobile video to generate a real-time 3D eagle-eye visualization, enabling users to view their surroundings from a top-down perspective.

This thesis investigates whether autonomous robots can realistically perform large scale crop monitoring in operational greenhouse environments. The work is part of a collaboration between vGreens Holding GmbH, an ag-tech startup developing plant physiology sensing technologies, and the Chair of Robotics, Artificial Intelligence and Real-Time Systems at TUM.

Are you knowleadgeable in GIS and landscape mapping? We offer the chance to develop your skills in an interdisciplinary project mapping the hydrology of the Biodiversity Exploratories of Germany, as part of the project WaterWeb.

Modern computer-assisted surgery relies increasingly on robust visual understanding of the operating room. While most existing approaches for surgical scene understanding and phase recognition are based on narrow-field-of-view endoscopic or monocular camera data, real-world operating rooms often provide significantly richer visual contexts. This thesis aims to investigate surgical scene understanding using multiple images for surgical phase recognition, and higher-level scene representations such as surgical scene graphs. The goal is to analyze how VLM/multi-view VLMs can use the visual context can improve robustness, generalization, and interpretability. The project will commence with a comprehensive review of the state-of-the-art in surgical activity recognition, surgical phase classification, and scene graph–based modeling. Based on this review, existing methods will be categorized and experimentally evaluated, with adaptations made where necessary to support panoramic or wide-field-of-view imagery. Particular emphasis will be placed on testing and validating approaches on recordings beyond classic benchmark datasets, including less constrained or previously unseen surgical recordings. The student is encouraged to bring in their own ideas and creativity. The outlined approach serves as a guiding framework rather than a strict implementation plan. Your own creativity is welcome to solve the problem; the upper proposal merely presents an idea. Please read the application details: https://hex-lab.io/vacancies/01_students/ Key research areas include: Reviewing the state-of-the-art in surgical activity / phase classification techniques and surgical scene graph techniques Categorize and experiment with existing approaches Test on recordings aside from classic datasets Recommended background (or motivation in learning): Basic knowledge of computer vision Experience with deep learning model training and application Some experience with C++ and Python Medical Context: - https://github.com/egeozsoy/ORacle Multi-View VLMs: - InternVL https://github.com/OpenGVLab/InternVL - LLaVA-NeXT as well: https://arxiv.org/abs/2407.07895; https://github.com/LLaVA-VL/LLaVA-NeXT Fine-tuning VLMs for 3D vision tasks: - https://github.com/openvla/openvla - https://roboflamingo.github.io/ - https://physx-anything.github.io/

As modern Autonomous Mobile Robots (AMRs) move away from fixed paths and begin to navigate freely in shared spaces, their intended actions can become ambiguous to human workers. To bridge this communication gap, this thesis focuses on the improvement of the already existing concept of the Anthropomorphic Robotic Mock Driver (ARMoD) an interface character designed to communicate a robot's intent to its surroundings.

The Human-Centered Computing and Extended Reality Lab of the Professorship for Machine Intelligence in Orthopedics seeks applicants for Bachelor/Master Thesis for the Summer Semester 2026.