Nontraditional teams work to develop future autonomous vehicle capabilities.
The race is on for super-advanced, beyond-next-generation technologies. Vying for a cash prize of $1 million, teams of engineers, software developers and car enthusiasts are taking on the challenge to create totally autonomous robotic ground vehicles that can travel from Los Angeles to Las Vegas on a designated course within a specified amount of time. The competition is part of a new program the military has developed to tap into the ingenuity of inventors throughout the United States who will design seemingly impossible capabilities that one day may be as commonplace in military operations as Predators.
Sponsored by the Defense Advanced Research Projects Agency (DARPA), Arlington, Virginia, the Grand Challenge program is an effort to reach beyond traditional defense contracting firms and military project offices to find innovative ideas and increase the speed of technology development. The autonomous ground vehicle contest is the first in what may be a series of events that explore solutions to a variety of problems the military faces. And, if the response to this first challenge is any indication, there is no shortage of clever inventors ready to demonstrate that today’s dream can be tomorrow’s reality.
The Grand Challenge idea developed last year. DARPA’s director, Dr. Anthony J. Tether, saw the need to increase both the speed of technology development and community outreach to people who do not work for the military or defense contractors but have different approaches to meeting military requirements. “He wanted to give those people who were apprehensive about dealing with the U.S. Defense Department acquisition bureaucracy the opportunity to spark their imaginations and come up with technology or ideas that would help the [autonomous vehicle] program along,” explains Col. Jose A. Negron, USAF, director, Grand Challenge, DARPA. One goal of the program is to reach out to people with a variety of skills, from grease monkeys to computer experts.
More than a dozen teams have already signed up to take on the first challenge, a number that came as a pleasant surprise to Grand Challenge organizers. “The question we asked ourselves was, ‘How many people would actually be interested in this type of competition?’ We didn’t realize there would be a big crowd,” Col. Negron says. When the challenge was announced, featuring a grand prize and winner-take-all format, the popularity of the competition became evident.
Originally, the program’s organizers anticipated that 50 people would be interested in attending an information conference. But as registrations for the competitors’ conference began coming in, Col. Negron knew that even the maximum the agency was preparing for, 250 attendees, would most likely be surpassed. In February, more than 400 people came to the conference just to obtain information about the race. So, Col. Negron admits, they were amazed by “the enthusiasm of the populace from academia to defense contractors to moms and dads to key scientists from throughout the world—some foreign nationals attended—coming to the event just to listen to ‘What is the Grand Challenge?’”
Grand Challenge Deputy Director Dr. Thomas M. Strat explains that the phenomenal interest in the challenge may be because it provides a focal point for a variety of organizations. “You have all these laboratories and universities and hobbyist clubs and some big companies that are working on robotics for various items—everything from battlebots to building systems for the Army. Many of these university labs and so forth have a lot of talent but are not focused on anything. What we’ve done is given them an application; we’ve given them a purpose,” Strat says. “So already, without spending a dime, the government has harnessed a whole lot of energy of innovative spirit out there by getting people to think about what it would take to build a vehicle to do this.”
By the end of May, 18 teams had registered, and the colonel says he could already see two or three strong competitors but adds that it is entirely possible that a registrant that is not well-known could win the event. “That happens. That’s exactly what this event is all about. From that perspective, it’s pretty exciting,” he relates.
Late this spring, the event’s organizers were designing the route that the vehicles will travel from Los Angeles to Las Vegas, a distance of approximately 300 miles. Col. Negron says that although it is a difficult course, it is possible. He believes that a team could win the first competition of the Grand Challenge program next March. If no one wins the first competition, the same challenge will take place in 18 to 24 months and continue to take place until either someone wins the contest or the congressional authority to award the cash prize expires in 2007.
“Congress granted DARPA that authority. I think we’re the only agency in the government that has the authority to issue cash prizes. The idea is to use that prize not to stimulate incremental progress on existing technologies but really to step out and do something very different, capture the imagination of people across the country,” Strat says. If a team wins during the first competition, Tether intends to conduct several challenges in various areas of technology. The agency currently is sifting through ideas about what other types of challenges could take place.
Robotics may be the obvious technology of the first challenge; however, Col. Negron points out that other capabilities will be equally important to completing the route. Mobility will be one area that teams will need to address. They will have to determine a vehicle type that can both survive and maneuver the course. Some of the teams are using existing vehicles as their platforms then developing the technologies to install on them. Others are starting from scratch and creating their own designs.
Sensing is the second technical area that must be addressed because each vehicle must be able to know its location and destination and identify obstacles it must negotiate.
The third area is behavior control. The vehicle must use reasoning to take the data it receives from sensors and determine what it is seeing, comprehend its objectives, plan its route and decide how fast it needs to go to complete the course in the designated amount of time.
Strat points out that a lot of robotics work has been done and capabilities have been demonstrated; however, this has generally taken place on a closed course for a distance of no more than five miles. The Grand Challenge route will be between 250 and 300 miles, and competitors will receive course information just two hours prior to the start of the race. These conditions will raise the bar on the required technologies, which likely will ensure that they have military applications, he adds.
To ensure that the competitors have a viable concept, each team is required to submit a technical paper that describes its approach. The day before the race, the teams will bring their vehicles to the California Motor Speedway where each will participate in a test run to demonstrate that the vehicle was built to the specifications provided.
John Nagle, the leader of Team Overbot, says that sensor myopia is the biggest technical issue the teams will have to address. “Existing sensors and processing algorithms aren’t good enough to drive off-road at any useful speed. This isn’t a fundamental problem; it’s a market size problem. Five companies have exited the scanning laser rangefinder field in the last decade due to lack of a market. What’s needed is a killer application for these things. That may come from the automotive industry, but it’s still a few years off. The Grand Challenge should accelerate the pro cess a bit,” he says.
Nagle’s team, which is made up of senior engineers in Silicon Valley, California, is considering submillimeter radar for terrain sensing. “That technology isn’t mainstream yet, but it’s coming. What’s needed for this job is a low-cost, three-dimensional, phased-array radar system with a range of perhaps 100 meters and which can image the ground at 5-centimeter resolution,” Nagle says.
Although these types of capabilities will be demonstrated in an autonomous ground vehicle, Col. Negron relates that they certainly could be applicable to other platforms as well. “I don’t look at it as just a ground vehicle. I’m looking at autonomy in the mobility, the behavior, the sensing capability. How would those technologies be applied in air technologies? How would an airman use it? How would a sailor use it? How would the ground forces use it?” the colonel states.
Concerning more far-reaching applications, Col. Negron points to recent operations and the area of logistics. In the supply convoys moving from Basra to Baghdad, for example, not all of the vehicles reached their destinations. If this procedure could be automated, it would greatly improve the process for the military, the colonel offers. In addition, autonomous vehicles can act as scouts, clear minefields or enter unstable areas after an explosion. “We don’t have that capability. We are just at the very beginning of that capability,” he says.
If the military chooses to adopt any of these technologies, they could move forward in several ways, Strat explains. “Technology transfer can and will happen at multiple levels. First and foremost, if somebody actually wins this, what it’s going to do is set a stake and say, ‘Hey, this level of autonomous navigation is possible.’ Just the proof that someone can do it will get others saying that they can do it too, and there will be lots of people replicating it,” he offers.
The accomplishment also will show military procurement managers who are skeptical about robotics that they should take another look at the potential of autonomous vehicles, Strat says. “So here’s a new set of resources that the various services or agencies within the services can draw from to satisfy their requirements,” he adds.
Col. Negron maintains that the challenge offers benefits that reach beyond the winning technology. “Not only will the winner accelerate the technology and put that stake down, but I call it the trickle down effect. What about the other competitors that did not finish first? The critical piece for me is the technologies that they may have in their vehicles that need to be explored that can be taken immediately off that vehicle and imparted into a military lab or right onto a military system. That would be fabulous,” the colonel says.
Col. Negron would be tremendously satisfied if at least one vehicle completes the entire route and happy if some of the entries make it only half to three-quarters of the distance. But he also would be impressed even if they can complete only one-quarter of the course. “Even to do 50 miles, you need some kind of intelligence. It’s not easy,” he says.
Although the colonel acknowledges that this first Grand Challenge may seem impossible to some, he notes that technologies that currently are in wide use began as ideas that many thought were impossible to achieve. “As an airman, I look at stealth technology, I look at GPS [global positioning system] today or precision-guided munitions. All of those were all dreams at one time. What really strikes me is the Predator. About 10 years ago, my predecessor was out here briefing the Predator and the concept of deployment, and they just about tarred and feathered the guy. Today, we use them all of the time. The search for new capabilities is a dream, but it’s also a possibility,” he says.
Additional information on the Grand Challenge is available on the World Wide Web at http://www.darpa.mil/grandchallenge.
Inventors Strive to Survive the Challenge
One goal of the Grand Challenge program is to reach out to individuals and organizations that are not traditional technology providers for the U.S. Defense Department. A look at some of the registered teams proves that objective has been met. The teams range from colleges that will use the challenge as a class project to seasoned experts who have been in the engineering business for years.
The Defense Advanced Research Projects Agency (DARPA) is sponsoring the event and intentionally kept the rules as open as possible so that teams of any size and with a range of resources could participate. Some of the teams are backed by large organizations, while others are searching for sponsors who will support their efforts in return for advertising.
The Red Team is led by William “Red” Whittaker, robotics professor, Carnegie Mellon University, who says that the challenges of this competition fall into three categories: teamwork, technology and resources. Whittaker’s team members reside throughout the United States, so coordinating the effort across time zones is one challenge. Detection of negative obstacles, such as ditches, holes and ruts in the road, is a technical problem that teams will have to address. “Speculation is rampant about what we might encounter. Creative approaches rage, but no silver bullet is apparent. The Red Team operates in reality, not dreams, and this is an area where technical magic isn’t coming,” he relates.
In terms of resources, the challenge is twofold: personnel and funding. Whittaker is interested in getting students involved in the project. “Students bring a dimension of enthusiasm and energy to complement the wisdom of veterans,” he states. The funding plan targets technical, supply, sponsor, foundation and individual members. Founding partners who are sponsoring the Red Team include Caterpillar, SAIC, Chip Ganassi Racing, Seagate, Applanix, Mobots and Carnegie Mellon University in Pittsburgh.
If youth is one of the keys to success in this competition, a team being organized by Charles Reinholtz, alumni distinguished professor in mechanical engineering at Virginia Tech, Blacksburg, may lead the pack. Reinholtz is using the challenge as a teaching tool. Virginia Tech has been competing in the Society of Automotive Engineers Mini Baja competition and the Association for Unmanned Vehicle Systems International Intelligent Ground Vehicle competition for several years. Approximately 20 undergraduates will work on the Grand Challenge project, which Reinholtz says is a high-level combination of those two events.
The competition also has sparked interest from recent college graduates. Seth Cabe, Team LoGHIQ captain, is a 23-year-old mechanical engineer who recently graduated from Rensselaer Polytechnic Institute, New York. Rather than modifying an existing vehicle, his team is designing its entry from the ground up. To fund the effort, the five-member team is looking for sponsors to donate funds or materials in return for sign space on the race vehicle, Web site and uniforms. The team also is negotiating with several television production studios.
But the challenge is not only for newcomers to technology; it also is an attractive project for veterans. Paul F. Grayson, chief engineer, American Industrial Magic LLC, describes himself as “an old Cold War warrior” who can still fight the enemy, even if it is with a T-square and slide rule. A lieutenant commander in the Merchant Marine Reserve and U.S. Naval Reserve, Grayson says that, as an engineer, he believes that all technical challenges can be overcome given enough time, inventive skill and money. “Looking into this project, it looks like the integration of systems is going to be the biggest challenge, but it is also my specialty,” he states.
James Hayter, who is leading a team called IRobotFactory, is a senior programmer/analyst at an Internet company. He has been interested in machines, electronics and physics all of his life and co-authored the first automated hotel check-in/check-out computer system in the United States. Hayter currently is writing automated media monitoring software for TVEyes Incorporated; however, his dream is to start a robotics company. He says he would most likely use the $1 million prize to fulfill that dream.
Leaders of teams that do not have corporate sponsorship have expressed some concern that they may be at a distinct disadvantage. Although DARPA has left the design somewhat open-ended, some of the other requirements, such as producing detailed designs five months prior to the race, could handicap smaller teams. Large schools and corporations have been examining some of the challenges for many years and can pull together designs and afford expensive inertial navigation systems much easier than a team that has been put together solely for the Grand Challenge. Despite this, many say that their techniques may still meet the challenge and win the race.