Stud. Assistant (m/f/d) - Learning contact-rich tasks in simulation with GenAI/Foundation models/RL

Advertisement for the field of study such as: Automation Engineering, Electrical Engineering, Computer Science, Cybernetics, Mechanical Engineering, Mechatronics, Control Engineering, Software Design, Software Engineering, Computer Engineering, Robotics, Autonomous Systems or comparable. Contact-rich tasks in assembly often involve complex physical interactions and high sensitivity to parameter settings. With a large number of product variants, it is not feasible to manually tune or validate process parameters for every possible variant, which limits scalability and flexibility in industrial applications. Simulation environments offer a promising approach to address this challenge. By accurately modeling the underlying physics of assembly operations, it becomes possible to learn generalized representations of these processes. NVIDIA Isaac Sim and other simulation engines like MuJoCo enables high-fidelity physics simulation, sensor emulation for accurately simulating these assembly processes. Reinforcement learning (RL) and modern foundational models can then be leveraged to learn policies and predictive models that capture the dynamics of contact-rich interactions.