Mehr als 500 Jobs in und um Heimsheim

Advances in learning-based control and the increasing demand for human-machine interaction are driving the need for modular safety certifications in modern control systems. Prominent areas include surgical robotics, automated driving and smart factories. Recent research efforts are addressing this challenge through so-called predictive safety filter methods, which allow the safe integration of, for example, human-in-the-loop inputs or reinforcement learning algorithms into safety-critical systems. While these safety filters ensure persistent constraint satisfaction, additional safety and stability properties are often desired, which may vary depending on the specific use case or environmental context. To this end, recent developments aim at extending predictive safety filters with additional desirable closed-loop properties such as bounded convergence and asymptotic stability. During your Master thesis, you will review the existing literature on predictive safety and stability filters, as well as related concepts from Model Predictive Control. You will design and implement a new advanced driver assistance and autonomous driving safety function using a recently proposed stability filter methodology. Furthermore, you will verify safety and stability in simulation using an advanced vehicle simulation environment, considering challenging driving scenarios with desired inputs generated by humans or imitative learning algorithms. In addition, you will refine the methods to improve practicality and include possible additional targets depending on the driving situation and environment. Last but not least, you will support the implementation and testing on a real vehicle. Social counselling and intermediary service for care services Discounts for employees