The Intelligent Manipulator project concerns the development of a robot assistant consisting of a compliant manipulator and a supporting computer vision system.
A manipulator is compliant if the robot must actively maintain its posture and can be overcome without harming the platform.
Compliant manipulators are often designed for interaction with humans and are thus exceedingly safe.
The robot assistant is envisioned to be programmed through demonstration.
To load a new behaviour into the systemís vocabulary the user must guide the manipulator kinaesthetically through a few demonstrations.
Any objects important to the task will be localized by the system using the computer vision system and some user input.
Using the object information and these demonstrations a behaviour model is determined statistically. This model should ideally capture the userís Ďintentioní.
The system must then be able to reproduce the task with minimal human interference.
Our lab, Riegl imaging laser scanner and Barrett Whole Arm Manipulator
The research being conducted in the project targets three major problems
- J. Claassens, "An RRT-Based Path Planner for Use in Trajectory Imitation," IEEE International Conference on Robotics and Automation (ICRA), Anchorage, USA, 2010.
- G. K. Singh, J. Claassens, "An Analytical Solution for the Inverse Kinematics of a Redundant 7DoF Manipulator with Link Offsets," IEEE/RSJ International conference on Robots and Systems (IROS), Taipei, Taiwan, 2010.
- J. Claassens, Y. Demiris, "Generalizing Human Demonstration Data by Identifying Affordance Symmetries in Object Interaction Trajectories," to be presented at the IEEE/RSJ International conference on Robots and Systems (IROS), San Francisco, USA, 2011.