CarSim(TM) Helps Race Team Find Its Way on Robotic Race Course

Carnegie Mellon Vehicles To Start In First & Third Position In Oct. 8, 2005 Race

ANN ARBOR, Mich., Oct. 6 -- Every long distance commuter has the same dream: to program their car to do the driving for them. The realization of that dream may now be on the horizon, as 23 finalist teams are preparing to compete in a road race of driverless vehicles Oct. 8 in the desert near Primm, Nev.

Ann Arbor, Mich.-based Mechanical Simulation Corporation has played a role in the 2005 Grand Challenge, applying its CarSim(TM) vehicle dynamics technology for the two vehicles representing Carnegie Mellon University, Pittsburgh. CMU's H1ghlander and Sandstorm vehicles earned the pole and third positions in the Oct. 8 road course, to travel a 165-mile course in 10 hours or less without human intervention.

Forty university and industry teams have designed and built electronically-controlled autonomous vehicles for the Grand Challenge, sponsored by the U.S. military's Defense Advanced Research Projects Agency (DARPA). CMU's teams and the other 21 finalists are competing for a $2 million first prize. But more important, DARPA will use this technology to program future vehicles to perform vital functions during wartime and in other military situations - without endangering humans.

Over the past year a team of roboticists, automotive professionals and students from CMU has equipped their vehicles with the latest drive-by-wire technology, electronic controls and computer assistance programs for suspension, steering and braking. Mechanical Simulation has worked closely with the CMU teams, using CarSim to predict vehicle behavior over a wide variety of road conditions. CarSim simulates the dynamic vehicle behavior of cars and light trucks, generating over 500 output variables that are automatically plotted and animated for ease of analysis.

According to Daniel Bartz, a member of Red Team and a CMU specialist in automated vehicles and road data collection, the team collected over 2,000 miles of three-dimensional road information. These data were entered into CarSim, along with the functional design details of the vehicles. CarSim can predict how the vehicles would perform through a range of course obstacles and weather conditions.

"CarSim is designed to be extremely flexible with the type of data it receives," explained Bartz. "The benefits of simulation are significant - accurate vehicle test results using experimental systems without spending large amounts of time in road testing. And you save the vehicle from potential damage on a rough course. We developed algorithms based on our models, utilizing different speeds, soil and other road conditions, including natural impediments such as rocks or trees. CarSim helped us see how the vehicle would react to every condition at various speeds. We were able to better understand when the vehicle will be stressed so we can avoid accidents that would eliminate us from the race."

"Being able to predict vehicle reaction under actual road conditions is of paramount importance to success in this event," Bartz said. "But the real- world rewards are in terms of new vehicle design and military technology. We're confident that much of what you see utilized in the DARPA Challenge will be used on vehicles in a few years."

Who knows, perhaps that programmable drive to work is not that far off in the future after all.