Reinforcement Learning for Robotics  (Lecture with Project)

LecturerMatteo Saveriano
Allocation to curriculumSee TUMonline
Offered inWintersemester 2018/19
Semester weekly hours4  
Scheduled datesSee TUMonline
RegistrationSee “Course criteria & registration”
ContactMatteo Saveriano (matteo.saveriano@tum.de)

 

 

Content

The course will cover the following topics:

1. Fundamentals of dynamical system theory

2. Fundamentals of optimal control (LQR)

3. Nonlinear optimal control (Iterative LQR, introduction to model predictive control (MPC))

4. Introduction to Reinforcement Learning (terminology and definitions, Markov decision process, policy iteration, value iteration, Q-learning)

5. Model-free reinforcement learning (PI2 – POWER)

6. Model-based Reinforcement Learning (PILCO, PI-REM)

7. Approaches that combine nonlinear optimal control (ILQR, MPC) and Reinforcement Learning

8. Introduction to Deep reinforcement learning (End-to-end approaches)

Previous Knowledge Expected

Fundamentals of Linear Algebra, Probability and Statistics, Programming skills in Matlab/SImulink

Objective

(Expected Results of Study and Acquired Competences) At the end of this course, students are able to:

- Implement reinforcement learning algorithms for robots and autonomous systems

- Evaluate the performance of reinforcement learning algorithms Languages of Instruction English