ICRA 2007 Tutorial
Nonlinear Control of Flexible Joint Robots
Date and Venue
The tutorial will be held from 8:30-18:00 on Saturday 14. April 2007 in conjunction with ICRA'07 (10-14 April) in Rome, Italy. It will take place at the School of Engineering of the Universitā di Roma "La Sapienza", Via Eudossiana 18. The School is in the area of S. Pietro in Vincoli, near the Colosseum.
To sign up for the tutorial, please mark it during the (online) registration procedure for ICRA 2007.
22.02.2007: We are happy to announce that finally the availability of the robot for the hands-on experiments is ensured!
Scope of the Tutorial
The tutorial presents the state of the art on nonlinear control of flexible joint robots. It is primarily intended for PhD students and researchers working on control issues of robots with non-negligible joint elasticity. But since many different approaches from nonlinear control will be treated it is also of interest for researchers who work on other aspects of robot control.
The control of robots with non-negligible joint elasticity is a topic that has drawn the attention of the robot control community since more than twenty years. Nowadays a huge amount of literature on this problem is available, mainly in form of conference and journal papers. However, there is no standard work available describing in a compact way the state of the art. Researchers instead must dig through a vast amount of papers in order to get a good overview of the known control solutions. This tutorial, therefore, aims at summarizing the nonlinear control methods developed for flexible joint robots. Presently, the control issues related to flexible joint robots come again into focus of robotics research when considering those newly developed robot technologies like antagonistically driven joints, variable stiffness actuation, serial elastic actuators, all motivated by efficiency and safety issues for robots directly interacting with unknown environments and with humans.
The tutorial will treat a large amount of methods from nonlinear control theory and detail their application to robot models with flexible joints. Apart from describing the state of the art, also missing links and open problems for further research will be identified.
Since performance of physical human-robot interaction can be assessed only in the real experiment, we will present at the end of the tutorial some selected experiments highlighting key issues of the covered material with a DLR and/or a KUKA/DLR light-weight arm.
- Dynamic Modeling of Flexible Joint Robots
- Regulation Control
- PD control, Torque Feedback, State Feedback
- Compliance Control
- Observer Design
- State Observer
- Disturbance Observer
- Tracking Control
- Singular Perturbation
- Static/Dynamic Feedback Linearization (slides will be available soon)
- Triangular Systems, Backsteppin, 'Passivity-Based' Tracking
- Experiments with the DLR-KUKA arms (CONFIRMED)
The presentation slides of the tutorial will be available here after ICRA'07.
Session 1: 8:30-10:20
Dynamic Modeling of Flexible Joint Robots [De Luca]
Regulation Control I: PD Control, Torque Feedback,
Full State Feedback [Albu-Schäffer]
Coffee Break: 10:20-10:40
Session 2: 10:40-12:30
Regulation Control II: Compliance Control [Ott]
Observer I: State Observer [De Luca]
Observer II: Disturbance Observer [Albu-Schäffer]
Lunch Break: 12:30-14:00
Session 3: 14:00-15:50
Experiments with the DLR-KUKA arm
Tracking Control I: Singular Perturbation Approach [Albu-Schäffer]
Coffee Break: 15:50-16:10
Session 4: 16:10-18:00
Tracking Control II: Feedback Linearization [De Luca]
Tracking Control III: Decoupling, Backstepping,
"Passivity-Based" Tracking [Ott]
Summary and Discussion
If you have any further questions on this tutorial, feel free to contact
christian.ott<at>dlr.de or alin.albu-schaeffer<at>dlr.de.
Last Modification: 08.03.2007