Broad Area Colloquium For AI-Geometry-Graphics-Robotics-Vision

Made for Each Other: Metric Localization and Distributed Topological Mapping

Gaurav S. Sukhatme
Robotics Research Lab
University of Southern California

Wednesday, May 17, 2000
refreshments 4:05PM, talk begins 4:15PM
TCseq201, Lecture Hall B


Localization and mapping have received considerable attention recently. I will present a Kalman filter-based approach to precision robot localization which optimally combines local rate sensing with global landmark information. Using the resulting location estimates, I will show how robots may easily build topological representations of their surroundings. Such representations are lightweight yet useful, and they scale very well. As an example I will show how multiple robots can efficiently combine individual topological maps without a priori knowledge of each others' coordinate systems. I will conclude with a discussion of open problems and extensions to a new domain (ubiquitous intelligent embedded systems).

About the Speaker

Gaurav Sukhatme is a Research Assistant Professor in the Computer Science Department at the University of Southern California (USC) and the Associate Director of the Robotics Research Laboratory. He was an undergraduate at IIT Bombay before receiving a M.S. and Ph.D. in Computer Science from USC. His research interests and previous work include sensor fusion for robot fault tolerance, robot localization and mapping, and human-robot interfaces. He has recently begun a new research effort in algorithms for distributed, intelligent, embedded systems design. He is a member AAAI and IEEE. For further information please see
Back to the Colloquium Page
Last modified: Mon May 8 12:41:03 PDT 2000