Control for Robotics: from Optimal Control to Reinforcement Learning

What it is about

You learn fundamentals and practical methods from optimal control, model-based learning and reinforcement learning with a view to robotics applications. In the end you will derive optimal control approaches, implement them practically and design learning-based controllers that deal with disturbances and model inaccuracies.

What you will be able to do

• derive optimal control equations for mobile robots
• practically implement and analyze optimal controllers
• understand state-of-the-art approaches to learning-based control and reinforcement learning
• design learning-based controllers to cope with nonidealities such as model errors

What the module consists of

• LectureIntroduction to core concepts of optimal control, model-based learning and reinforcement learning
• In‑Class Exercises and Tutorialsjoint work through examples for deepening understanding
• Assignmentspractical application and hands-on implementation experience
• Optional Course Projectdeeper exploration and application in a larger practical project

Teaching method

• Lectures (presentation/Blackboard/slides)Conveying the theoretical foundations
• In‑Class Exercises and TutorialsApplying and practicing the concepts on examples
• Assignments (handouts/implementations)practice-oriented exercises and code implementation
• Online Videosadditional tutorials to support self-study