"Designing an Autonomous Vehicle to Compete in DARPA's Urban Challenge"

Yiannis E. Papelis, PhD
Research Associate Professor
ODU - Virginia Modeling Analysis & Simulation Center

Prior DARPA-sponsored autonomous vehicle competitions focused on off-road
navigation and obstacle avoidance.  For 2007, DARPA sponsored the Urban
Challenge, a competition within which autonomous vehicles must navigate
roadways and interact with other manned and unmanned vehicles and obey
traffic laws, all while pursuing missions that involve reaching a series of
specific checkpoints in a finite amount of time.  This talk will outline
the efforts of a team that designed a vehicle that entered the competition
and was one of the 11 teams that reached the final stage.  Issues covered
will include the overall design approach, primary and secondary sensor
selection, AI strategy and implementation, control strategies, navigation
approach and finally, practical lessons learned both during the preparation
phase as well as the final week of the competition.

Yiannis Papelis earned a BSEE (with honors) from Southern Illinois
University in 1988, a MSEE from Purdue University in 1989 and a Ph.D.
degree in Electrical & Computer Engineering from the University of Iowa in
1993.  He has been involved in research on immersive virtual environments
used primarily in ground vehicle simulations.  While at the University of
Iowa, he conducted research in autonomous agents with applications in
microscopic traffic simulation, rapid development of virtual environments
used in ground vehicle simulation, design of deterministic scenarios within
stochastic simulations and effects of in-vehicle devices in transportation
safety.  Dr. Papelis spent one year as a visiting faculty at the University
of Central Florida where, among other things, he applied earlier research
in driver modeling to an autonomous vehicle that competed in the DARPA
Urban Challenge competition.  Dr. Papelis is currently conducting research
on autonomous agent modeling issues as applied to a wide range of topics,
including simulation of critical infrastructure components, realistic
simulation of crowds and control of autonomous ground and aerial vehicles.