Patrick Deegan

Computer Science Department
University of Massachusetts Amherst
Office: CS254, (413) 545-3039
pdeegan(at)cs.umass.edu

B.S. Engineering, 1997, Harvey Mudd College. Claremont, CA

I'm a graduate student in the UMASS Laboratory for Perceptual Robotics.

Research Projects

  • The Balancing uBot
  • Robot Learning, Control, and Skill Acquisition
  • Distributed Control in Multi-Robot Systems
    		•

    Current Research

    I am currently completing my Ph.D. Thesis proposal.

    Interests

    Bimanual mobile manipulation tasks require whole-body control to create interaction forces while preserving postural stability. Thus, a human being will lean into a heavy object to push it. Moreover, the marriage of mobility and manipulation affords the ability to create momentum in powerful, massive structures and to transform it into kinetic energy in smaller limbs and grasped objects. My research explores the representations and control for exploiting dynamics in mobility and manipulation. I plan to show that dynamically stable robots have new opportunities to control the impedance of manual interactions as well. A series of robots have been constructed to this end - the UMASS uBot series.

    History of the uBot Platform

    uBot-0.5

    The uBot-0.5 platform was originally known as Oscar due to its dependence on spare parts and lab junk. Oscar had a rotating sonar and pyroelectric mast actuated by an R/C servo as well as differential steering powered by more R/C servos. Computation was provided by the MIT Handyboard and odometry was maintained by hand-made position encoders on each wheel (IR reflective). The uBot-0.5 could search for warm objects and avoid obstacles using a reactive control framework.

    uBot-1

    uBot-1 was deployed in teams of as many as 10's of reconfigurable copies in a distributed sensor network. Each platform utilized a Motorola 68332 processor for low-level motor control and IR proximity sensing. A StrongARM based SBC was employed for high-level programming and wireless connectivity. The platform had differential steering and casters at the front and rear for static stability. The IR proximity sensors were placed on each of the 8 edges to form an entire ring of coverage and a color camera could be mounted to accomodate either a standard machine vision lens or an omni-directional lens. A team of uBot-1s could maintain line-of-sight constraints while exploring unknown maze-like environments. See more at: uBot-1

    uBot-2

    After removing uBot-1's casters, adding an IMU consisting of a 1-axis accelerometer and a 1-axis rate gyro, and relocating the batteries to the top of the platfrom, uBot-2 emerged as a dynamically balancing robot. The control followed textbook state space notation and design for a Linear Quadratic Regulator. See more at: uBot-2.

    uBot-3

    uBot-3 demonstrated force transfer to the environment using kinematic prototypes of 3-DOF arms. The arms were machined from aluminum and actuated by R/C hobby servos.

    uBot-4

    Coming soon!


    Publications

  • Thibodeau, B., Deegan, P., Grupen, R. Static Analysis of Contact Forces With a Mobile Manipulator, IEEE International Conference on Robotics and Automation. Orlando, Florida. May, 2006.

  • Rosenstein, Michael, Robert Platt, Patrick Deegan, John Sweeney, Oliver Brock, Andrew Fagg, and Roderic Grupen. Haptic Coupling of Dexterous Manipulation and Dynamic Mobility. In Video Proceedings of the IEEE-RAS/RSJ International Conference on Humanoid Robots, Los Angeles, USA, November 2004. ( humanoids 04.avi; Mobile Manipulation.avi)

  • Araujo, E., Karuppiah, D., Yang, Y., Grupen, R., Deegan, P., Lerner, B., Riseman, E., Zhu, Z., Software Mode Changes for Continuous Motion Tracking, International Workshop on Self Adaptive Software, Oxford, England, April, 2000. (.ps.gz: 1,061,132 bytes; .pdf: 909,006 bytes)

    • updated 20-May-2006

    pdeegan(at)cs.umass.edu