Description

In this role, you will validate controllers for nonlinear dynamical systems using a data-driven approach. Your work will involve conducting literature reviews, selecting models, and integrating them into simulation environments, ultimately optimizing controller performance.

Requirements

• •Master's degree in Engineering, Computer Science, Robotics, Mathematics, or comparable
• •Experience with dynamic systems, simulation, machine learning, system identification, and control theory
• •Motivated self-management, clear communication of complex issues, and independent task structuring
• •On-site presence required
• •Fluency in English
• •Enrollment at university

Tasks

• •Focus on validating controllers for nonlinear dynamical systems
• •Explore a data-driven validation approach
• •Identify requirements for closed-loop simulation of dynamic systems
• •Conduct a literature review on state-of-the-art foundation model architectures
• •Compare foundation models for modeling dynamic systems
• •Design and implement a data-driven workflow
• •Develop a strategy for data selection and preparation
• •Fine-tune selected foundation models
• •Generate additional measurements for data when needed
• •Systematically evaluate the accuracy of fine-tuned models
• •Compare fine-tuned models with traditional physics-based and data-based models
• •Analyze model performance, focusing on corner cases
• •Integrate fine-tuned models into a closed-loop simulation environment
• •Validate performance of existing controllers
• •Explore optimization possibilities for controllers
• •Analyze trade-offs between model accuracy, simulation speed, and computational resources
• •Provide recommendations for practical applications