         
|
GRASP Lab Seminar 2003-2004
December 12, 11:00 AM, Levine Hall 307, hosted by Kostas Daniildis.
Stergios Roumeliotis
University of Minnesota
Analysis of Positioning Uncertainty in Reconfigurable Networks of Heterogeneous Mobile Robots
Abstract: Recent trends in sensor technologies have expanded our sensing capabilities in terms of scale, quality and content of information. The availability of large and diverse amounts of sensory data has introduced a set of unprecedented challenges when attempting to infer causal relationships or model the behavior of complex systems. Groups of autonomous robots represent an exemplary case of such systems.
Numerous robotic applications require that robots work in close collaboration. Communication and sensor sharing among groups of robots can significantly increase the accuracy and robustness of the performed tasks. In order to coordinate with the rest of the team, each robot needs to know its position and orientation, i.e. it has to localize. We present a novel formulation of the multi-robot localization problem and its solution by employing a Kalman filter based estimator. This estimator is able to compensate for the correlations between relative position measurements and previous state estimates while optimally combining all available positioning information. We provide an analytical expression for the upper bounds in positioning accuracy as a function of a weighted connectivity graph for the network of relative position measurements in the robot group. This network has a time-varying topology determined by the availability of relative position measurements between pairs of robots. The weights of the connectivity graph depend on (i) the odometric and orientation accuracy of each robot, and (ii) the accuracy of the robot tracker on each member of the team that measures its relative position with respect to other robots in the group. During my presentation I will also discuss several cases of sensor based probabilistic algorithms recently developed or under investigation for various types of autonomous vehicle navigation: autonomous stair climbing, safe and precise planetary landing, and Mars rover navigation on steep terrain.Biography: Dr. Stergios I. Roumeliotis received his diploma in Electrical Engineering from the National Technical University of Athens (NTUA), Greece in 1995, and his M.Sc. and Ph.D. degrees in Electrical Engineering from the University of Southern California (USC), Los Angeles, CA in 1999 and 2000, respectively. Between 2000 and 2002, he was a Postdoctoral Fellow at the Division of Engineering and Applied Science at the California Institute of Technology, Pasadena, CA. Currently, he is an Assistant Professor at the Department of Computer Science and Engineering and a faculty affiliate with the Digital Technology Center at the University of Minnesota. His research has focused on inertial navigation, sensing and estimation for distributed autonomous systems, sensor networks, and fault detection and identification. He has authored or coauthored more than 30 journal and conference papers in the above areas. He was the recipient of the Myronis fellowship (USC, 1998-2000). Finally, he has received research grants from the Jet Propulsion Laboratory (JPL, NASA) and the National Science Foundation (NSF) for work pertinent to autonomous vehicle state estimation.
