Thesis Flagship Project - RoboRacer Autonomous Racing
- Institut
- Professur für autonome Fahrzeugsysteme (TUM-ED)
- Typ
- Bachelorarbeit Semesterarbeit Masterarbeit
- Inhalt
- experimentell theoretisch konstruktiv
- Beschreibung
We are proud to announce our new Autonomous Racing Thesis Flagship Project at the Technical University of Munich. Over the last few years, we have developed our RoboRacer (formerly F1Tenth) autonomous racing platform as a central pillar for student thesis projects, offering a hands-on environment where students can master coding and AI challenges on a real-world embedded platform.
From day one, you will gain valuable skills for your academic and professional life by pushing the limits of autonomous systems. Unlike traditional theoretical theses, we provide you with your own RoboRacer vehicle from the very first day, ensuring your research is grounded in reality and tested on the track.
Motivation & Research Topics
Autonomous racing serves as a critical test for researchers, acting as a vital building block for general autonomous vehicles and AI. By offering a real-world 1:10 scale autonomous vehicle platform to our students, we enable you to focus entirely on algorithmic design and high-speed multi-agent racing.
We are looking for motivated students who want to work at the intersection of robotics, autonomous driving, and high-performance racing. We value students who take the initiative, bring in their own ideas, and truly possess passion for the topic. You can shape your thesis within any part of our cutting-edge software stack, whether it is perception, motion planning, control, or topics such as end-to-end learning algorithms, including foundation models or high-fidelity autonomous racing simulators. We also provide the following explicit example of what your thesis topic could look like:
- Differentiable, Gradient-Based Motion Planners: Move beyond static racelines by developing planners capable of gradient-based optimization. You’ll evaluate sampled trajectories through complex cost functions to bridge classical planning and ML, creating highly adaptive, real-time racing and overtaking strategies.
- Adaptive Control - Merging RL with MPC: Combine MPC’s reliability with RL’s adaptability. You will develop algorithms that allow vehicles to learn optimal racing policies and adapt to dynamic track conditions in milliseconds while maintaining strict safety guarantees.
- High-Fidelity Sim2Real Frameworks: Close the "reality gap" by enhancing the RoboRacer simulation with HD track models and realistic sensor suites. You’ll implement interfaces for learning agents and validate performance through extensive real-vehicle vs. simulation benchmarking.
- Cognitive Racing with Foundation Models (LLM/VLM): Integrate LLMs and VLMs into the software stack to add a reasoning layer. You’ll develop a logic for strategic overtakes and behavioral analysis, providing the car with a cognitive-level understanding of racing strategy.
Your work won't stay on a screen. The algorithms you develop will be deployed on the actual RoboRacer racing car, demonstrating full autonomous driving at the dynamical limits. You will demonstrate how your code makes our vehicles fast, more intelligent, and safer. These capabilities prepare you in the best hands-on way for your future career, whether in academia or the automotive industry.
- Voraussetzungen
Requirements:
- Coding experience (Python, PyTorch)
- Framework experience (ROS 2, Git)
- Knowledge in Autonomous Driving or Robotics
- Curious mindset
- Independent and organized work attitude
- Open-minded team player
Your Benefits:
- Fascinating, future-oriented field of research
- Young and dynamic team
- Academic and professional support beyond the thesis
- Organized and structured thesis project
- If suitable, publication as a scientific paper
- Thesis project in German or English
The thesis can be started immediately. Please fill out the application form under the following link: Google Forms - Thesis Application
- Tags
- AVS Jahncke
- Möglicher Beginn
- sofort
- Kontakt
- Felix Jahncke