Skip to content
Technical University of Munich

Open Student Thesis Offers

When writing your thesis at TUM I6, please follow our thesis guidelines. A guide to writing good thesis can be found here. A collection of useful material for research can be found here.

Realizing Natural Language Scenario Descriptions in CARLA Using Large Language Models

We are looking for a self-motivated student to work on Code generation for Carla using Large Language Models:

AUTOtech.agil - Future of Autonomous Driving and Intelligent Traffic Infrastructure

CeCaS: Autonomous driving - Systems and Software Engineering

As part of the research project CeCaS, a group has come up to build a new system architecture for future vehicles
with a focus on autonomous driving.

Reinforcement Learning, Representation Learning, Meta-RL and Robotics

Three of our DEMOs can be found at https://sites.google.com/view/kuka-environment/ , https://sites.google.com/view/cemrl, and https://videoviewsite.wixsite.com/rlsnake.

  1. Meta-RL:
  2. RL:
  3. Language-conditioned Meta-RL:
  4. Safety Control for Robotic Manipulator

We have several other open topics in the domain of reinforcement learning in robotics and we also accept open proposals or ideas with yourselves. For more information, please contact Zhenshan Bing.

Embodied AI and Foundation Models

1. Embodied Lifelong Learning

The demo and recent publication related to this topic can be seen through this link: https://www.nature.com/articles/s42256-025-00983-2

New topic released at 2025.04:

[BA/MA]: Robotic Lifelong Reinforcement Learning

Keywords: Lifelong Reinforcement Learning, Cross-Embodiment, Shared Latent Representation, Transformer, Mixtrue of Experts

2. Inference Reuse for Meta-Reinforcement Learning

New topic released at 2025.04:

[BA/MA]: Enhancing Inference Reuse in Meta Reinforcement Learning for Non-Parametric Task Distributions (Contact: y.meng@tum.de)

This thesis project aims to advance current research in inference-based Meta Reinforcement Learning (MRL) by optimizing algorithmic frameworks for inference reuse and extending the agent’s meta-knowledge generalization capabilities to non-parametric task distributions. The work will focus on enhancing existing methods, developing novel optimization strategies, and conducting comprehensive experimental evaluations across diverse simulated tasks. Opportunities for transfer to real-world robotic platforms may also be explored.

Applicants are expected to have 

  • Strong proficiency in Python programming and a solid understanding of deep learning methodologies (completion of the introduction to deep learning course is required).
  • A strong academic performance (GPA < 2.3) is mandatory.
  • Availability for full-time work (at least 6 months).
  • Prior experience with reinforcement learning techniques is highly desirable and will be considered an advantage.

keywords: Meta-inference reutilization, Meta Reinforcement Learning, Mujoco  

3. Vision-Language-Action (VLA) foundation models

We are inviting highly motivated students to join our research on building robotic manipulation systems for unstructured environments, leveraging the power of foundation models. This open project direction focuses on developing Vision-Language-Action (VLA) models, integrating diffusion policies, transformers, and imitation learning techniques to enable robust, generalizable robotic skills.

Students interested in this direction should have:

  • Solid programming skills in Python and familiarity with ROS (Robot Operating System) and Linux OS.
  • A strong background in deep learning, particularly in vision or reinforcement learning models.
  • A strong academic performance (GPA < 2.3) is mandatory.
  • Availability for full-time work (at least 6 months).
  • Prior hands-on experience with robotic hardware and manipulation tasks is highly desirable and will be given priority.

Keywords: Foundation Models, Vision-Language-Action Models, Robotic Manipulation, Diffusion Policy, Transformer, Imitation Learning.

Outstanding candidates will have the opportunity to collaborate with us on publications in top-tier journals and premier academic conferences. If you are passionate about pushing the boundaries of embodied AI and large-scale robotic foundation models, we encourage you to reach out and apply!

Additionally, we have several ongoing projects and welcome interested students to reach out at any time: Yuan Meng

Model Predictive Control for Quadruped Robots

Whisker-Inspired Tactile Sensor and Biomimetic Rodent Robot

Contour Extraction via Contiuous Contact
Whisker Sensing Mechanism Schematic
Search Area and Volume of the Whisker Array

We aim to build a flexiable, light-weight tactile sensor based on the biological structure of rodent's vibrissae, and apply them on the robot arm, biomimetic rodent robot and other platform to provide perception and self-estimate ability.

Sensor Design and Contact Estimate:

Others:

For more information about this topic, please contact Yixuan Dang.

Autonomous Driving, Robotic Grasping, and Dense Prediction (Seeking outstanding students all the time!)

The general research focus is as follows:

  • [GR/FP/IDP/SA/BA/MA/etc.] Topic 1: Vision language model for robotic grasp pose estimation (2D planar and 6D).
  • [GR/FP/IDP/SA/BA/MA/etc.] Topic 2: Object detection and semantic segmentation for autonomous driving.
  • [GR/FP/IDP/SA/BA/MA/etc.] Topic 3: Autonomous driving perception based on adverse weather image restoration.
  • [GR/FP/IDP/SA/BA/MA/etc.] Topic 4: Multi-modal or multi-sensor fusion for robust object perception (2D/3D detection, segmentation, depth estimation, end-to-end autonomous driving, etc.).
  • [GR/FP/IDP/SA/BA/MA/etc.] Topic 5: Correspondence-based pose estimation / simultaneous pose and correspondence registration (SPCR) (2D-3D / 3D-3D registration or globally optimal algorithms).
  • [GR/FP/IDP/SA/BA/MA/etc.] Topic 6: Large language model (LLM) and vision language model (VLM) for robotics (autonomous driving or robotic grasping).
  • [GR/FP/IDP/SA/BA/MA/etc.] Topic 7: Medical image segmentation.

I am also looking for highly self-motivated students to work on risk-based safety model for autonomous driving research. The topics include:

  • [GR/FP/IDP/SA/BA/MA/etc.] Topic 1: LLM for safety-critical scenario generation for autonomous driving.
  • [GR/FP/IDP/SA/BA/MA/etc.] Topic 2: Risk quantification estimation( Responsibility Sensitive Safety (RSS),  Safety Force Field (SFF), etc.).
  • [GR/FP/IDP/SA/BA/MA/etc.] Topic 3: Uncertainty estimation for autonomous driving perception (such as bayesian neural network).
  • [GR/FP/IDP/SA/BA/MA/etc.] Topic 4: BEV-based detection and map segmentation.
  • [GR/FP/IDP/SA/BA/MA/etc.] Topic 5: Behavioral intention prediction in driving scenes (Trajectory prediction, Behavioral intention prediction (BIP), BIP-aware trajectory prediction, etc.).

If you are interested in these research topics, please send me an email with your CV and transcript of records. We also have several other open topics in robotics and welcome proposals or ideas from your side. For more information, please contact Dr. rer. nat. Hu Cao.

Object Detection
Large Vision-Language Model for Robotic Grasping
Robotic Grasping

Systems and Software Engineering

Our group provides a range of topics related to systems and software engineering with applications in robotics and automotive.

Currently, the following thesis proposals are open:

Neural SLAM and Biomimetic Rodent Robot

For more information about this topic, please contact Zhenshan Bing and Florian Walter.

BMW - Source-Agnostic Cooperative Perception

As part of the research project from BMW, our focus is on the real-time infrastructure of the V2X communication and sensor fusion.

 

For more information about this topic, please contact Kuo-Yi Chao.

KI.FABRIK: Future AI & Robotic Factory

 1. Dual/Multi-Robot Collaboration for Cable Manipulation:

 2. Path Planning:

 3. Obstacle Avoidance:

 4. Visual Servoing Control:

 5. Knowledge Graphs with Foundation Models

 6. Dynamics Learning:

 7.  Underwater Salient Object Detection:

 8.  Shared Autonomy based Teleoperation for Tight-Clearance

DeepSLAM: Deep Learning based Localization and Mapping (Vision-based Perception and Navigation)

HAL4SDV - Systems Safety Security Software

HAL4SDV, or the Hardware Abstraction Layer for Software-Defined Vehicles, is a European research initiative aimed at revolutionizing the automotive industry by developing a standardized hardware-software integration framework.

This project supports the transition to Software-Defined Vehicles (SDVs), where software plays a central role in vehicle functionality, allowing for over-the-air updates, modular design, and enhanced customization.

Key goals of HAL4SDV include:

  • Creating a unified interface between vehicle hardware and software to simplify development processes.
  • Enabling faster deployment of new features and ensuring high safety and performance standards.
  • Supporting both safety-critical and general-purpose vehicle applications.
  • Enhancing collaboration across European countries to build a comprehensive SDV ecosystem.

We are looking for eager and motivated students to work on the following topics:

Application of Retrieval-Augmented Generation (RAG) Systems in Standard Compliance

We are looking for a self-motivated student to work on extracting test scenario for autonomous driving research using RAG system:

We have other open topics in the domain of Large Language Models in autonomous driving software development and we also accept open proposals or ideas with yourselves. For more information, please contact Vahid Zolfaghari.

Spiking Neural Networks - Next Generation AI for Autonomous Driving

At the KI-ASIC project we are researching about the application of bio-inspired neural networks to real-world applications.

If you are interested in learning about neuroscience and how neuromorphic engineering is trying to narrow the gap between biology an technology, do not hesitate to contact us.

Available:

Previous topics:

Simulation-Based Learning Control for Real-World Robotic Manipulation and Navigation

For more information about other available topics in reinforcement learning, data-driven control and sim2real transfer, please contact Hossein Malmir.

Autonomous Vehicles

The Cyber Physical Systems group is pursuing a wide range of research directions related to safe decision making, motion planning and control for autonomous vehicles, involving both formal methods, sampling- and optimization-based methods as well as deep learning-based methods.

Currently, the following thesis proposals are open:

Safe Reinforcement Learning in Single Robot and Multi-Robot Systems

We have several topics about Reinforcement Learning, Robotics,  Autonomous Driving and AI Safety, for more information, please contact Shangding Gu (Remote).

Safe Reinforcement Learning in Robotics

Currently open positions:

Additionally, we will have open topics in safe reinforcement learning for manipulators and mobile platforms in the future.
If you are interested in these topics, you can contact: jakob (dot) thumm (at) tum.de

Modular Robotics

For interest in a BA/MA thesis in machine learning for modular robotics, please contact Jonathan Kuelz or Julian Balletshofer.

Currently open:

Robust and Nonlinear Motion Planning & Control

Deep Learning for Computer Vision

Point cloud-based 3D object detection.
Generated with DALL·E mini. Text input: "chair of robotics, artificial intelligence and real-time systems".

If you are interested in researching trending topics in the field of computer vision and pattern recognition:

  • Object detection
  • Segmentation
  • Tracking
  • Surface reconstruction
  • Mesh generation
  • Text-to-image generation
  • Domain adaptation of synthetic data

then please contact: Bare Luka Žagar

Formal Methods and Reachability Analysis

Offline Reinforcement Learning

[Research internship] Mastering the game of Skat using decision transformers

Safe Reinforcement Learning, Multi-Agent Reinforcement Learning

Reinforcement Learning for Safe and Efficient Combustion Engine Control

Neurorobotics in the Human Brain Project

SPEED-CARGO Project

Machine Learning Algorithms for Hybrid Vehicle Data

Please see this page for the available topics about 3D Object Detection and Tracking.

Autonomous Robot & Visual Servo & Deep Learning & Robot Design & Medical Robotics

For more information, please visit my homepage Mingchuan Zhou or contact me via email (zhoum@in.tum.de).

OSBORNE (Future Automotive E/E Architectures for Autonomous Cars)

We have a set of open topics in the domain of affective computing and multimodal emotion recognition, within the context of OSBORNE project, for more information please contact Sina.

HORSE / REMODEL

We have several open topics related to computer vision in the HORSE and REMODEL projects. Please contact Arne for more details.

Collaboration with Chair for Product Development and Lightweight Design

External thesis proposals

Low-level vision

For more information, please visit my homepage (Yuning Cui) or contact me via email (yuning.cui@in.tum.de).

To top