VASC Seminar Announcement
            =========================

Date: Monday, 12/13/99
Time: 3:30-4:30pm
Place: Smith Hall 2nd Floor Common Area

Speaker:  Iwan Ulrich
          CMU Robotics Institute
          http://www.cs.cmu.edu/~iwan/

Title:  Appearance-Based Place Recognition for Topological Localization

Localization is a fundamental problem for most mobile robots. Although research in 
localization has been very active and has made recent improvements, we are not 
aware of any system that works well both indoors and outdoors. In this talk, I will 
present a novel topological localization method that performs well in a variety of 
man-made indoor and outdoor environments. Our localization system uses color 
vision, works in real-time, and can easily be trained in new environments. Furthermore, 
our system does not require any modifications of the environment. To sense the 
environment, the localization method uses a panoramic vision system. The acquired 
color images are classified in real-time based on nearest-neighbor learning, image 
histogram matching, and a simple voting scheme.

The localization system has been evaluated in a variety of large-scale environments. 
I will present the results of eight cross-sequence tests in four unmodified environments, 
three indoors and one outdoors. In all eight cases, the system successfully tracked the 
mobile robot's position. The system correctly classified between 87% and 98% of the 
input color images. For the remaining images, the system was either momentarily confused 
or uncertain, but never classified an image incorrectly.

I will also briefly talk about our experience with IEEE-1394 digital cameras (Firewire). 
We recently wrote a device driver for Windows that communicates with these cameras. 
The driver and a software library will soon be made publicly available on the web. 
This driver allows real-time acquisition of high quality images on a regular laptop 
computer, which is appealing for many mobile robot applications.