Pursuing the Dream of Cheap, Reliable, Useful Robots.
New Control Kids on the Block
Maudlin Auditorium (NSH 1305 )
Talk 3:30 pm
It is easy to make simple robots that are cheap and reliable, it is more difficult to make them also be useful. The idea of building robots with few moving parts, stripped-down computing, and basic sensors is appealing because there is less that can go wrong, but making these types of systems perform nontrivial tasks is a challenge.
I see two keys to making simple robots useful. The first is to design dynamic mechanisms and associated controllers generate novel forms of mobility and manipulation. The second is to develop estimation algorithms that enable a robot to determine its position over large, diverse environments with modest sensing and computing. Advances in these two areas will combine to yield cost-effective, reliable robots that have the mobility and perception necessary autonomously navigate across the assortment of environments in which humans live and work.
In the first part of this talk, I will review some of the work that I have been involved in here at CMU that has led me to this point of view. This includes work with an underwater vehicle (DEPTHX), a balancing robot (Ballbot), several variants of the SLAM problem, and sensor networks. In the second part of the talk, I will outline some specific research problems in the areas of mechanism design, control, and estimation that I intend to explore toward the goal of creating cheap, reliable, useful robots.
“New Control Kids on the Block”
As a recently-hired Robotics Institute faculty member I will "officially" present myself to the RI community. In this talk I will present my past and current work on control of nonholonomic systems and aerial vehicles, and discuss my technical work plans for the next five years. I will also briefly share the non-technical aspects of life as a researcher that I learned while working as an Innovation Manager for a private R&D institute.
George Kantor is a new Systems Scientist at the Robotics Institute at
. He received M.S. and Ph.D. degrees in Electrical and Carnegie Mellon University
Computer Engineering from the
in 1995 and 1999, respectively. He has been at RI since 2000, both as a Postdoctoral Universityof Maryland College Park
Research Fellow and a Project Scientist.
His research interests lie in the control of robots with interesting dynamics, estimation for localization and mapping, and distributed sensor networks.
Marcel Bergerman is a systems scientist with the Robotics Institute. Since December 2005 he works at the Autonomous Helicopter Laboratory, developing
navigation, guidance, and control methodologies for robotic helicopters. From June to November 2005 he worked at the RI's Tele-Supervised Autonomous Robotics Laboratory, on the NASA-funded "Wide Area Prospecting Using Supervised Autonomous Robots" project. Prior to joining the RI Dr. Bergerman was an Innovation Manager with Genius Institute of Technology, in Manaus, Brazil, where he was responsible for funding procurement, knowledge management,
institutional relationships, intellectual property, and official certifications and accreditations. Prior to that he was the co-principal investigator of
, a project financed by the Brazilian government focused on the development of a robotic airship for environmental monitoring. Dr. Bergerman received his Ph.D. from the Electrical and Computer Engineering Department at CMU in 1996, and M.Sc. and B.Sc. degrees in Electrical Engineering from the University of Sao Paulo, Brazil, in 1992 and 1990 respectively. AURORA
For appointments, please contact Peggy Martin (firstname.lastname@example.org)