Advanced Robot Learning and Decision-Making

What it is about

You will learn advanced methods for robotics that connect modeling, optimal control and learning-based methods (including reinforcement learning). In the end you will be able to model robotic systems, derive modern control algorithms and apply learned models in control architectures to handle uncertainties and model errors.

What you will be able to do

• Derive optimal controller design for mobile robots
• Develop learning-based approaches to compensate model errors
• Assess the strengths and limits of optimal control, learning-based control behavior and RL
• Implement state-of-the-art algorithms and apply them prototypically

What the module consists of

• VorlesungenIntroduction and conveyance of the theoretical foundations
• In-Class TutorialsWorking through examples and setups
• Praktische SessionsHands-on implementation of algorithms, partly in lectures, partly as take-home exercises

Teaching method

• VorlesungenConveying the core concepts
• In-Class TutorialsDeepening through practical examples and exercises
• Praktische Sessions / ProgrammieraufgabenImplementation and application of algorithms to solidify understanding