MIT Team Designs Autonomous Vehicle

Imagine you are driving around town when you pull up to a stop sign. As you glance over at the car across the intersection, you are astonished to see that there is no driver. As the car makes a smooth right turn, you realize that the car is driving itself.

It sounds far-fetched, but such a scene may be closer than you think. A team consisting of faculty, researchers, and students from MIT, the Franklin W. Olin School of Engineering, and Wellesley College, as well as industry experts from around the area, is trying to build such a vehicle. The group, led by John J. Leonard, professor of mechanical and ocean engineering, entered into the latest Defense Advanced Research Projects Agency’s Grand Challenge. The goal is to build an autonomous car which can drive in an urban environment.

The difficulty of the goal becomes apparent when one considers the list of obstacles that DARPA is planning to throw in the paths of the vehicles. From large potholes and weak GPS signals to downed tree branches, obstructing construction equipment, and blocked roads, the cars will have to find their way from an initial position to a final position by sequentially visiting certain intersections, known only by their GPS coordinates. The car has to use a map of the course to navigate from one intersection to the next.

The challenge is so difficult that the MIT team has been working closely with many MIT departments, including Mechanical Engineering, Aeronautics and Astronautics, Electrical Engineering and Computer Science, as well as the School of Engineering and the Computer Science and Artificial Intelligence Laboratory. “It’s great to be part of the MIT community where there are so many experts and so many people willing to offer advice,” Professor Seth Teller of EECS, a faculty adviser for the challenge, said.

The level of difficulty involved in the Urban Challenge is what appealed to the members of the team, Teller said. Well, that and the potential for wide-ranging benefits to the environment, productivity, and safety. Troy Jones, an engineer at Draper Laboratory who has been working on the project since late July, said that although DARPA’s goal of creating autonomous vehicles for supplying the armed forces and performing reconnaissance is important, the project has the potential to save even more lives by making automobiles safer.

Autonomous cars could also boost productivity by allowing people to work during the time that they would normally have spent driving. Teller also described how future cars may communicate with each other in order to remain at speed when entering intersections, thus saving energy.

The competition, which began in May 2006, has thus far consisted of an application and a site visit. Coming up for the 36 remaining teams is the semi-final round to be held Oct. 26–31 in Victorville, Calif. Land Rover, the company that donated the MIT car, will also take care of transporting the modified vehicle to the former George Air Force Base. The top 20 teams in the semi-final round will move on to compete in the finals on Nov. 3.

Both undergraduate and graduate students have been driving forces in the project. Teller said that students approached the faculty with the idea of competing for the $2 million first prize. According to Gaston A. Fiore G, a graduate student in the Aeronautics and Astronautics Department, the students have learned not just about technical aspects but also about group dynamics.

The team has expectations beyond just qualifying for the final round. Robert D. Truax ‘08 said that the team “definitely has a chance at winning.” “Because the programmers are so ambitious and trying to make it the perfect vehicle, if something unexpected comes up, our car has a great chance of overcoming it,” Truax said. “We’re trying to go above and beyond the competition.”

The MIT team stands out from the crowd in that the team hopes to do more than simply win the Grand Challenge. The faculty hopes that, if the car proves successful, an autonomous vehicle laboratory could be created at MIT to continue studying this technology. Fiore said the students would like the car’s software to be open-source and available for public use. This would allow others to build on the progress made by the team at MIT. DARPA, in turn, would receive design solutions to the problem of creating autonomous urban vehicles.

When asked if the massive time commitment was worth it, Fiore replied that “it’s rewarding in a lot of aspects. You develop something and you get to test it. It’s your own little creation.”