Architectures, abstractions, and algorithms for large teams of robots

Abstract

Networked robots represent the convergence of robotics, sensor networks and mobile ad-hoc networks, with many applications and a growing market projected to be $200B in 2013. This talk will focus on some fundamental problems and practical issues underlying the deployment of large numbers of autonomously functioning robots. The central problem is the so-called inverse problem of deriving individual robot behaviors for a desired group behavior. There are numerous examples of group behavior in biology which suggest that analysis of swarming behaviors in biology may provide insight for the synthesis of collective behaviors for engineered systems. The author presents a methodology for modeling and analyzing such collective behaviors and discuss architectures, abstractions and algorithms for the control of large networks of robots.