PhD Student - Inverse problem based on reinforcement learning(13357)
Key Duties and Responsibilities
- Build or adapt a physics-based simulator of the cold-spray process.
- Validate simulation models against experimental measurements to ensure accuracy and reliability.
- Translate target part geometries into optimization constraints for spray-gun trajectory, including position, orientation, and speed.
- Implement cost functions that capture shape error, material usage, energy consumption, and mechanical constraints.
- Develop and apply classic optimization methods and advanced reinforcement learning algorithms.
- Design neural-network policies and value functions for high-dimensional robot-arm control problems.
- Define rigorous evaluation protocols, including shape-accuracy metrics, computation time, and robustness to uncertainties.
- Compare algorithmic performance using high-fidelity simulations and, where possible, conduct tests on a physical test-bench.
- Prepare monthly technical reports, draft conference/journal papers, and present results in seminars.



