Living Las Vegas

At Home in the City of Lights

sponsored resources

Top Vegas Hotels: Bellagio, Luxor, Rio

Roads from the Ashes: An Odyssey in Real Life on the Virtual Frontier, by Megan Edwards

Standing on the Corner, Watching All the ‘Bots Go By

For the third time in four years, the Defense Advanced Research Projects Agency (DARPA) held a robot race and invited the public to come and watch the spectacle. This time, the event was billed as the “Urban Challenge,” because, unlike the first two events, which were held in the howling desert, this contest required the ‘bots to navigate city streets and interact with traffic. They even had to share the road with each other, something the organizers claimed had never happened on earth before. Unable to resist the opportunity to see autonomous vehicles yield to each other at four-way stop signs, I headed out last Friday on a road trip in search of geeks, ‘bots, and the people who love them.

My destination was Victorville, California, a former watering hole on Route 66 and the former home of the Roy Rogers Museum. What DARPA officials liked, however, is that Victorville is also home of the former George Air Force Base, which was shut down in 1992 but still boasts a large ghost town of decaying houses, dead trees, and a network of obsolete streets. To anyone else, the enclave might look like a good set for a low-budget remake of Night of the Living Dead, but to DARPA, the site was the perfect venue for a select group of robotic SUVs, pick-up trucks, cars, and one big Oshkosh truck to test their artificial intelligence without killing people. To ensure that the ‘bots could do no harm even if they went berserk, untold miles of chain link fence and thousands of concrete K-rails were erected around the race course. Other improvements included a big grandstand next to the start/finish line, a huge tent equipped with hundreds of chairs and giant video monitors, team “pit” areas, and an infinite number of portable toilets.

I drove out to the Urban Challenge site the night before the big race. My experience at the other DARPA events had been that team members would be burning the midnight oil with their ‘bots, making final adjustments to hardware and tinkering with software code. It would be a good chance to get an idea of what the teams thought about the race and how the qualifying events had gone.

During the previous week, the field had been reduced from 38 teams to eleven. While a few eliminated teams had stuck around for the final event, most had departed, leaving noticeable – and for me, sad – gaps in the expanse of asphalt in the “pit area.” Missing, for example, was The Golem Group, a team I had admired greatly at the previous two races. CajunBot, the University of Louisiana’s entry, was another vehicle that didn’t make it to the finals, but the team had stayed to celebrate with their traditional night-before-the-race crawfish feast anyway.

The teams still “in residence” ranged from the top performers in the 2005 race to several first-time entrants. Dominating the scene with a big white tent was Stanford University, whose ‘bot took first place in the 2005 race. Boasting a slate of sponsors including Volkswagen, Google, and Red Bull, Stanford’s enclave looked like something at a NASCAR race. Inside, Junior, a 2006 Passat station wagon named in honor of Stanford’s founder, glistened under generator-powered lights as uniformed team members chatted nearby. No last-minute tinkering here, just an aura of subdued confidence.

Nearby, a similar tent housed the team from Carnegie-Mellon University. Inside lurked Boss, a 2007 Chevy Tahoe named for inventor Charles F. “Boss” Kettering. No tinkering was going on here, either, but fortunately team member Jim Nikolaou was willing to chat. Jim works for GM, one of Carnegie-Mellon’s major sponsors. He’s also a Carnegie-Mellon alum.

Although DARPA’s goal in developing autonomous vehicles is to meet a congressional mandate that a third of military vehicles be autonomous by the year 2015, the technology is obviously valuable for improving the safety and performance of ordinary cars. This potential is in part what has drawn companies like GM and Volkswagen to sponsor DARPA Challenge teams. “We’re interested in ‘active safety,’” Jim explained as he described systems that can help keep cars on track, improve stability, and detect hazards. But, as if to underscore the unique nerd-as-jock nature of the event, he also said, “We’re in this to win.” There was no denying the determination in his tone. Carnegie-Mallon came in second in 2005, losing to Stanford’s Volkswagen Touareg. This year, apparently, it was going to be a VW-GM face off.

I did find one team burning the candle at both ends. Way off in a corner created by the back fence and a gravel pile the size of the Great Pyramid, lights were on and geeks were busy. This was the team from the University of Central Florida in Orlando. Knight Rider, a 1996 Subaru Outback, once belonged to the wife of one of the team members. It competed in the 2005 DARPA Challenge and still bore some scrapes from a close encounter with a tunnel wall to prove it. UCF’s team had only six members and the leanest budget of any group that made it to the finals. In talking to lead engineer Don Harper, I began to understand why most of the ‘bot enclaves were dark. In previous races, hardware issues kept teams working until the last possible minute. This time, with the emphasis on the bots’ abilities to analyze information and make decisions, the biggest challenges lay in programming. “People might be working in their hotel rooms,” Don said. “With computer modeling, they can run scenarios as if their ‘bot were out on the road.”

I couldn’t help hoping that a team like UCF’s might pull off a surprise victory in the morning. Not only did UCF’s team manage to get to the main event with a used car and practically no money, they had to test their ‘bot surreptitiously in parking lots in the middle of the night. In contrast, Carnegie-Mellon had the use of “Robot City,” a private testing facility near their campus, and when winter weather got too severe, they moved to one of GM’s proving grounds in Arizona. If the guys from Florida were good enough to compete with that on a soapbox derby budget, what might they accomplish with some serious funding?

The next morning, after getting up at zero-dark-hundred, I was back out at the racecourse for a press conference at 6:00. Norm Whittaker, the DARPA official in charge of the Urban Challenge, fielded questions in a big tent full of reporters and photographers. He explained that the winner would be the ’bot that completed the prescribed course and three “missions” within six hours while following traffic rules as outlined in the California DMV Driver’s Handbook. Not only would the vehicles be on the road with over sixty human-piloted cars, they’d be interacting with each other. They’d have to park, cross two lanes of moving traffic, deal with stop signs, and pass. If they broke laws or behaved dangerously, DARPA officials stationed all over the course would write tickets. When the race was over, all the data would be tallied up, and the winners would be announced the next morning.

Norm didn’t smile much, but he did light up when a reporter asked a question about how collegial the teams were. He seemed very pleased that the DARPA Challenge events have resulted in increased enrollments in engineering and computer science programs that focus on robotics, and it was then I began to realize that – once you got past the labels – there really was no “them and us” at this gathering. DARPA dude or ‘bot boy, they were all nerds at heart. Even though the event was dressed up to look like the Indy 500, there was a collegial spirit that I’ve never detected at an ordinary car race. Scientists can be competitive, of course, but when they share a larger goal, their interest in and respect for other people’s work shows through. The more I talked to participants and organizers at the DARPA Urban Challenge, the more I felt a sense of camaraderie along with the competition. Carnegie-Mellon’s Jim Nicolaou, for example, told me how his team had repaired the University of Florida’s ‘bot when it had a damaging interaction with a wall during the qualifying events. “We brought it in and rebuilt the whole front end,” he said. That’s not a story you’d hear at the Indy 500.

The race was set to start at 8:00, and the grandstand was full of spectators. A JumboTron video monitor next to the starting chutes provided a magnified view of the officials on a high platform. After a patriotic ceremony featuring Marines on wild Mustangs and a rendition of the Star Spangled Banner, sixty specially-outfitted Ford Tauruses driven by professional drivers paraded onto the race course. Then DARPA big cheese Tony Tether kicked the challenge off.

Or at least he tried to. Carnegie-Mellon’s Boss had earned the right during the qualifying events to hit the course first. But when the moment arrived for the ‘bot to roll, nothing happened. Then, nothing continued to happen. Finally, as nothing kept on happening, the number two ‘bot, Stanford’s Junior, rolled out. This was my first chance to get a feel for the personality of each ‘bot. I’d been hearing that, Junior was “smooth,” Boss was “aggressive,” and MIT’s vehicle was “schizoid,” but now I could anthropomorphize them for myself.

Junior was indeed smooth. He zipped out of his starting chute and headed over the start line giving no clue that there was nobody at the wheel. Other ‘bots followed smoothly until one of them turned a bit too sharply and smacked into a K-rail. A “do-over” was granted, and it finally pulled out onto the course.

Meanwhile, back at the first chute, Boss was still inert. The JumboTron monitor went dark, and red-clad team members scurried about. At last, about twenty minutes after the first ‘bot rolled, Boss stepped on his own gas and headed out onto the course. I couldn’t tell whether he deserved his “aggressive” tag yet, but he was going to need a trait like that to catch up with Junior.

DARPA had done a good job of setting up observation points around the ghost town that formed the course. Bleachers had even been erected at the “traffic circle,” a long curved stretch that allowed spectators to see the robots on the road together. A few other access points offered views of a four-way stop sign, the parking test area, and a passing lane. The best view, however, was probably from the helicopter that circled all day. Video monitors in the big tent displayed aerial views and shots from cameras posted all over the course. Even though you couldn’t be everywhere at once, it was reasonably easy to keep track of the ‘bots as they attempted to carry out their required three “missions.”

For the next six hours or so, I watched the ‘bots from several vantage points, but the most fascinating for me was the four-way stop sign. Watching driver-free vehicles pull to a stop, appear to look both ways, wait for others to proceed, and then pull out themselves was excellent drama, especially, of course, when things didn’t go quite as planned. Knight Rider, the Subaru from the University of Central Florida, stopped for fifteen minutes at the stop sign. My fellow spectators and I had begun to assume the creature was dead, even though its antennae, lights, sensors, and required noise emitter were all still wiggling, flashing, turning, and beeping. Then, for no apparent reason, the car pulled out safely, made a perfect left turn, and went on its way. From the same spot, I saw Team AnnieWay get eliminated from the competition. The car, a Volkswagen Passat sponsored by a German team, came to a stop. DARPA officials ferried team members to the ‘bot in another car, and they drove it off the course. Later, Knight Rider, the Subaru from Florida, took a wrong turn into a driveway and never came back out. The ‘bot sat out the race nosed into an old house, its sensors turning and lights flashing the whole time.

A few hours into the race, hunger got the better of me. Thinking I might head to a Starbucks in Victorville, I drove out of the old Air Force base. My path put me on old Route 66, where I noticed a vintage diner across from a cement factory. Its parking lot was full of cars, and I found a place among them. I had discovered Emma Jean’s Truck Stop, “Home of the Brian Burger.” It’s been in operation since 1947, one of only a handful of Route 66 businesses still thriving. An excellent BLT and much-needed cup of coffee later, I headed back up the hill to find the ‘bots.

This time, I stopped at another viewpoint where a dirt road the ‘bots had to navigate connected with a four-lane stretch of highway. This was where they were supposed to show off their passing abilities. I saw only one ‘bot in this area, MIT’s Land Rover. It had stopped at the white line at the intersection of the dirt road and the highway. We waited, but so did the ‘bot. Eventually, team members arrived in a DARPA car. They ran to the ‘bot, spent a few minutes inside, then ran back to the DARPA-mobile and sped away. The ‘bot’s brain, I learned later, had gotten a quick reboot, enabling it to head out on its way again.

I encountered MIT’s Land Rover one other time, not far from the four-way stop sign. It was here I began to understand why it had earned the adjective “schizoid.” I was standing behind a barricade of K-rail, but it was still unnerving to see a large SUV suddenly turn its wheels in my direction and begin to accelerate. I half expected it to laugh after it scared me.

The race ended rather unspectacularly, because we all knew the first ‘bot over the finish line wasn’t necessarily the winner. Not too surprisingly, since he had started first and performed consistently, Stanford’s Junior was the first to get a checkered flag. Carnegie-Mellon’s Boss was in hot pursuit, which was impressive because of the 20-minute delay at the start line. Four other ‘bots succeeded in completing the whole course: Virginia Tech’s hybrid Ford Escape called Odin, the Ben Franklin Racing Team’s Toyota Prius named Little Ben, MIT’s Land Rover, and Cornell University’s black Chevy Tahoe.

The ‘bots and their humans headed back to their pit areas, and the audience dispersed. The DARPA officials responsible for coming up with a slate of winners had a long night ahead of them. Unlike simple races where crossing a finish line ends all discussion, coming up with Urban Challenge winners required the consideration of a huge pile of data. It may have been good planning on their part that this was the night Daylight Savings Time would end, giving them an extra hour. They also bought themselves a bit more time by moving the recognition and awards ceremony to 10:00 a.m. instead of 8:00.

The next morning, on a high platform near the finish line, three trophies stood on display. All of them were eagles, each more tremendous than the last. Word floating around the spectator crowd was that the largest one weighed more than a hundred pounds. I’d been taking a straw poll about who the winner of the biggest bird might be, and people seemed to be split 50/50 between Stanford and Carnegie-Mellon.

Carnegie-Mellon, whose team was headed up by longtime robotics whiz Red Whitaker, ended up receiving the $2 million prize. Stanford got $1 million, and Virginia Tech got $500,000. Everybody who made it to the finals did a head bow to Tony Tether or Norm Whittaker to receive an Olympics-style medal.

Although I was of course not privy to all the data and discussion that went into deciding who won the DARPA Urban Challenge, I couldn’t help getting an Animal Farm feel about how the teams were treated. Over the course of the weekend, it seemed obvious that while all the teams were nominally equal, some were considerably more equal than others. Stanford, for example, was not assigned to a pit area right next to a giant pile of dirt. The dust storm generated by every slight breeze can’t have been good for UCF’s computers, and they had no tent. Also, during the race, I observed MIT and Cornell both get the opportunity to restart their ‘bots when they stopped performing, but when Team AnnieWay’s car stopped for a minute or two, they were immediately cut from the competition. And then there was the issue of Carnegie-Mellon’s delayed start. Not only was the team allowed to keep on trying, they were allowed to replace hardware on their car. At the awards ceremony, Tony Tether placed blame for the ‘bot’s troubles on radio frequencies generated by the JumboTron nearby. While I was not the eye of God observing from on high, from my lowly spot on the sidelines it looked as though the teams from select major American research institutions were given numerous special indulgences.

Which of course is just a reminder that the DARPA Urban Challenge was not really a car race. In actuality, it was a clever way to spend tax money to get difficult and expensive research done as quickly and cheaply as possible. If, as I’m sure is true, Carnegie-Mellon’s research capabilities and likely results exceed those of the University of Central Florida, well, I guess it really does make more sense to throw cash and acclaim their way. As for the German teams, I find it easy to believe that DARPA found it in their own best interest to avoid giving any of them the purse, even though they had the required American sponsorship. Otherwise, they might have to explain why $2 million of U.S. taxpayer money earmarked for research here in the homeland was winging its way to Germany. As for the Stanford/Carnegie-Mellon rivalry, Stanford won the big check last time, and anyway, isn’t it just better to give the biggest eagle to a good old all-American Chevy rather than a Volkswagen? I couldn’t help wondering about all this once again when I learned that the officially published difference between the finish times of Stanford and Carnegie-Mellon was one second. Given all the times the ‘bots had to pause and all the judgment calls about whose fault any mishaps were, a lot of subjectivity went into coming up with that one second.

In the end, of course, as in the beginning, it’s all about the technology. DARPA isn’t trying to be NASCAR, and they really have no incentive to be democratic or fair. What they have is a mandate to meet their looming deadline, and even with the progress that’s been made, that may well be an impossible dream. But even if the Urban Challenge wasn’t a genuine sporting event, I for one am grateful that DARPA decided to open it to the public. Watching those robots and interacting with the minds that are designing them was like getting a peek into the future. It was a tantalizing glimpse that left me very eager for more. And if it inspires more bright students to study engineering and computer science, that’s a good thing, too.

Planning | Advice | Tips | Routes | Books | Fun!