Robotics
ACME - The Active Measurement Facility
ACME is a 15 dof robotic system we have designed and built for Reality-based Modeling. ACME also explores new ways of teleprogramming complex robots on the Internet and multi-sensor active perception (ACME can simultaneously use 4 cameras, two range sensors, a force sensor, and several microphones). More information is available at the ACME website.D. K. Pai, J. Lang, J. E. Lloyd, and R. J. Woodham. ``ACME, A Telerobotic Active Measurement Facility.'' in Proceedings of the Sixth International Symposium on Experimental Robotics, Sydney, Australia, March 1999. [pdf 336k] [gziped postscript 1.4M]
Personal Desktop Robotics
I'm interested in how small personal robots can perceive and manipulate common objects like paper and pens found on desks, and interact with people. I have been collaborating with Mason (CMU), Rus (Dartmouth), and Erdmann (CMU) in this area. We built a robot with 4 independent wheels called the ``mobipulator'' which uses its wheels flexibly for both locomotion and manipulation of paper on a desktop (see also Platonic Beast).Videos: early prototype, more at Mason's...
M.T. Mason, D. K. Pai, D. Rus, L. R. Taylor, and M. A. Erdmann. A Mobile Manipulator. 1999 IEEE Int Conf Robotics and Automation. [pdf 360k] [.ps.gz 600K] .
M.T. Mason, D. K. Pai, D. Rus, J. Howell, L. R. Taylor, and M. A. Erdmann. Experiments with Desktop Mobile Manipulators. 1999 Int Symp Experimental Robotics. [pdf 584K], [ps.gz 1.1M].
Telerobotics with Contact
Performing contact-based manipulation over the Internet is very difficult because of latency and bandwidth limitations. We addressed this problem by providing the user with a local simulated environment which is kept synchronized with the remote real environment using a model-based vision system at the remote site, and recognizing the user's actions. This allows even untrained users to perform difficult contact tasks by simply interacting with the local simulated environment.Videos: overview, cornering task, dragging a block around a corner.
J. E. Lloyd, J. S. Beis, D. K. Pai, and D. G. Lowe, ``Programming Contact Tasks Using a Reality-based Virtual Environment Integrated with Vision.'' IEEE Transactions on Robotics and Automation, 15:3, June 1999, pp. 423--434. [pdf 455k], [ps.gz 1M].
Multiresolution Motion Planning
I'm interested in algorithms for multiresolution surface representation and related applications. For instance, we developed a motion planner which finds good paths on rough terrain using wavelets with good approximation properties and a new terrain roughness measure based on the multiresolution.D. K. Pai and L.-M. Reissell, ``Multiresolution Rough Terrain Motion Planning,'' IEEE Transactions on Robotics and Automation, 14 (1), February 1998, pp. 19--33. [ps.gz 640k].
The Platonic Beast
How do we make a robot walk? Legged robots have often been based on biological designs but robots do not have the same constraints or capabilities as biological systems. In the early 1990's I designed and built a spherically-symmetric robot called the Platonic Beast. The symmetry of the robot allows it to recover more easily from falls and to locomote with a novel ``rolling gait''. See video (sped up to 2.5 frames/second of real time). We also explored new ways of programming such complex robots (the Beast had 12 motors and 4 processors on board), for instance, using Least Constraint.D. K. Pai, R. Barman, and S. Ralph, ``Platonic Beasts: Spherically Symmetric Multilimbed Robots,'' Autonomous Robots, 2:4, pp. 191--201, December 1995. [pdf 360k]
Constraint Programming for Robots
I've proposed a framework called Least Constraint (LC), for programming robots by specifying constraints on the robot's actions. A constraint based, declarative approach appears to be more natural for programming robots in several situations, e.g., for high dof robots with multiple simultaneous objectives, for fault tolerant robots, and for dynamic and uncertain environments.D. K. Pai, ``Least Constraint: A Framework for the Control of Complex Mechanical Systems,'' in Proceedings of the American Control Conference, (Boston), pp.~1615--1621, June 1991.
R. J. Spiteri, D. K. Pai, and U. M. Ascher, ``Programming and Control of Robots by means of Differential Algebraic Inequalities.'' IEEE Transactions on Robotics and Automation, 16:2, April 2000, pp. 135--145. [pdf 330k]
S. K. Ralph and D. K. Pai ``Computing Fault Tolerant Motions for a Robot Manipulator,'' in Proceedings of 1999 IEEE Intl. Conf. on Robotics and Automation, Detroit, MI, May 1999.