Dodo Alive! - Resurrecting the Dodo with Robotics and AI: Simulation & Control

Lecturer (assistant)
  • Sami Haddadin [L]
  • Dennis Ossadnik
  • Abdalla Swikir
Number0000002337
TypePractical course
Duration5 SWS
TermSommersemester 2023
Language of instructionEnglish
Position within curriculaSee TUMonline

Admission information

Objectives

Upon completion of this laboratory, the students have acquired profound knowledge in planning, design and control of dynamically walking bipedal robots both methodologically and practically. In addition to their problem-solving and teamworking capabilities, the students expand their management and organization skills

Description

Imagine that you are at the natural history museum. You wander around and take a look at different parts of the museum. Suddenly, a robot in the form of a long extinct species – the Dodo – moves towards you, starts talking to you and tells you about itself. This course covers constructing an autonomous, intelligent, walking, Dodo-like robot, which requires a combination of a wide spectrum of fields, including design, control, navigation, locomotion, perception and AI. During this laboratory course, the students will develop the necessary skills to design and construct various components of “Robo-Dodo”, as part of a long-term, iterative project. The ”Simulation & Control” course will focus on the fundamentals of modeling, simulation and control of bipedal walking robots. After some introductory classes, a practical phase will follow in which the students work closely together in interdisciplinary groups. The introductory sessions are structured as follows: In the first part, the necessary theory is taught. Then, a live programming session follows in which the theory is applied in practice. The following topics are covered: Simulation: • Floating-base kinematics and dynamics • Contact modeling o Event-driven o Time-stepping methods (Stewart & Trinkle) Control: • Reduced-order models (templates) • Spring-Loaded Inverted Pendulum (SLIP) and derivations • Passive dynamic locomotion • Limit Cycles, Poincaré map analysis, stability • Optimal control methods • Contact-implicit optimal control

Prerequisites

We expect that the student is familiar with basic robot modeling (kinematics and dynamics) and programming (Matlab).

Teaching and learning methods

- Introductory classes - Independent student work - Team work (including supervised and unsupervised work in the laboratory)

Examination

The module mark is determined by a final project work, in which the students demonstrate their ability to work in a team and to put theory into practice. This includes a 10-minute presentation and a written report (10-15 pages) in which the students present their results. Here the functionality of the final software/hardware application is evaluated, as well as its development and testing process and the theoretical principles used for it.

Recommended literature

- M. W. Spong, S. Hutchinson, M. Vidyasagar, ‘Robot Modeling and Control’, vol. 3, New York: Wiley, 2006. - R. Alexander, ‘Locomotion of Animals’, Springer, 1982. - R. M. N. Alexander, R. McNeill, G. Goldspink, ‘Mechanics and energetics of animal locomotion’, Chapman and Hall, 1977. - H. Choset, K. Lynch, S. Hutchinson, G. Kantor, W. Burgard, L. Kavraki, S. Thrun, ‘Principles of Robot Motion: Theory, Algorithms, and Implementation’, MIT Press, 2005. - B. Siciliano, O. Khatib, ‘Springer Handbook of Robotics’, Springer, 2016. - M. W. Spong, S. Hutchinson, M. Vidyasagar, ‘Robot modeling and control’, vol. 3. New York: Wiley, 2006. - C. M. Bishop, ‘Neural networks for pattern recognition’, Oxford university press, 1995.

Links