In the AVS lab, we are constantly innovating the way autonomous systems are developed, tested, and optimized. With access to 1/10th scale race cars and full-scale autonomous vehicles, our lab offers a unique platform for cutting-edge research in autonomous driving. This thesis focuses on implementing a training system for a fully differentiable software stack used in F1TENTH autonomous race cars.
Traditional training approaches often involve training individual models sequentially. However, with the rise of fully differentiable stacks, where every module can be trained based on its own loss function, this method becomes inefficient. Instead, a holistic approach where all modules are trained simultaneously is necessary to ensure optimal performance. This approach enables real-time policy adjustments during autonomous races, ensuring fast and adaptive decision-making.
Key Work Packages:
Requirements:
Your Benefits:
The thesis can be started immediately. Please send your resume, recent grade report, and a brief description of your fields of interest and why you are the perfect fit for this thesis to felix.jahncketum.de.
I am always on the lookout for motivated and committed students and look forward to receiving your application so we can work together on the mobility of tomorrow.