Abstract: Researchers have addressed the localization problem for mobile robots using many different kinds of sensors, including rangefinders, cameras, and odometers. In this paper, we consider localization using a robot that is virtually ``blind'', having only a clock and contact sensor at its disposal. This represents a drastic reduction in sensing requirements, even in light of existing work that considers localization with limited sensing. We present probabilistic techniques that represent and update the robot's position uncertainty and algorithms to reduce this uncertainty. We demonstrate the experimental effectiveness of these methods using a Roomba autonomous vacuum cleaner robot in laboratory environments.
| 12-06-06-loc1.ogg 12-06-06-loc2.ogg 12-06-06-loc3.ogg |
December 6, 2006 | The same localization plan from three different starting locations. |
| 02-26-07-loc1.ogg 02-26-07-loc2.ogg 02-26-07-loc3.ogg |
February 26, 2007 | A longer localization plan in a non-simple and less rectilinear environment. The same plan from two distinct starting locations. |
| 03-07-07-wireless.ogg 03-08-07-wireless1.ogg 03-08-07-wireless2.ogg 03-08-07-wireless3.ogg |
March 7-8, 2007 | The Roomba no longer needs to be connected to the desktop computer to perform its localization tasks. Onboard processing is done by a gumstix computer mounted next to the power button. |
| roomba-driver.tgz | October 5, 2006 | A C++ library for UNIX-like OSes that takes care of the details of serial communications and the Roomba API. In principle, this should allow you start controlling your Roomba very quickly. We expect it to be useful in research and educational settings. |