Robot Task and Motion Planning (R-TAMP)

Course Type

Lab Course

Study Programs

• Master Computer Science
• Master Data Science
• Master Media Informatics
• Master Software Systems Engineering
• Master Simulation Sciences
• Master Computational Engineering Sciences

Content

Task and Motion Planning (TAMP) deals with combining high-level decision making with low-level manipulation in robotics. In the lab course, we implement and evaluate multiple state-of-the-art TAMP approaches in AI domestic service robot scenario. You will learn about key concepts of robotics such as task planning with PDDL, motion planning algorithms, as well as integrated task and motion planning (TAMP).

You will work in groups to implement recent approaches to integrated TAMP using the ROS2 middleware, first in simulation, then on a HelloRobot Stretch 3 domestic service robot. By the end of the course, you will have a solid understanding of ROS2, TAMP, and practical skills for robotic software development both in simulation and with a real robot.

Examples

Examples of TAMP in Practice:

The HelloRobot Stretch 3 in action:

Objectives

• Introduction to TAMP (Task Planning, Motion Planning, Integrated TAMP)
• Introduction to ROS2 with PyRoboSim
  • Understand core concepts of TAMP, ROS2, and practical aspects of robot programming
• Implementation of a TAMP solver within ROS2
  • Understand a state-of-the-art approach to TAMP
  • Implement the approach with ROS2 in the context of a household scenario
  • Test your ideas in a 3D Simulation of the HelloRobot Stretch 3
• Transfer your results to a real Stretch 3

Recommended Prior Knowledge

Bachelor degree in CS or equivalent. Experience with machine learning, AI techniques, robotics.

Registration

Please register via the SUPRA platform offered by the department of CS. If you have any questions regarding the registration, please contact Daniel Swoboda.