Team Tartan takes DARPA Urban Challenge win
November 4, 2007 The man-most-likely finally got to stand atop the victory dias in the DARPA Urban Challenge today when Tartan Racing’s Chevy Tahoe Boss gave Red Whittaker the victory everyone thought would be his in the 2005 Grand Challenge. Tartan Racing won the US$2 million prize for first, while Stanford Racing’s VW Junior won the $1 million second place prize, reversing the order from the last DARPA event. Third was Team Victor Tango’s Odin. The event was a massive triumph for the educational system of Pennsylvania which provided the dominant winner of the event (from Carnegie Mellon) and the most ingenious and successful of the underfunded “Track B” teams which came from University of Pennsylvania and Lehigh University (Ben Franklin Racing Team).
Tartan Racing's win was no surprise - the team's Chevy Taho named BOSS looked formidable from the first day that all the vehicles got on the track together, running cleaner than all the other robots and “much more aggressively” as other team leaders were want to describe it.
Tartan looked the winner in every respect. It;'s an old racing adage that to finish first, you must first finish and Boss's speed came not from brute force, but from confidence in its vision and processing powers. As with most form's of racing, anyone can go fast for a lap or two, but to maintain that speed until the end of the race requires the skills to cope with problems at speed and to make continuously infallible decisions about what to do to correct them. Boss's speed came from confidence that its systems would enable it to make good decisions and NOT crash or break the rules.
But speed and cleanliness of its manouvres and decision-making were not the only aspects which made Team Tartan look unbeatable all week.
Its corporate sponsors all turned up in force for the press conference and put on one of the most impressive showings of the entire week (behind Boss's display in the final and the monumental logistical nightmare overcome by DARPA in staging the event) at the team press conference on Friday.
Just how do you assemble a team of sponsors which includes General Motors AND Caterpillar AND Continental AND Target AND Google AND a bunch of other household names. Putting together the dream team of sponsors deserves recognition in itself and it seemed to fit with the general perception that the overall package was complete from every angle.
Friday's press conference was opened by Whittaker but once it was underway, he left the floor to a stellar cast which included GM’s Larry Burns (VP of R&D and Strategic Planning), Caterpillar’s Tana Utley (VP and CTO) and Continental Automotive System’s President Karl-Thomas Neumann. Burns had GM’s CTO in tow and together with Team Tartan’s CTO Dr Chris Urmson, the four stood eye-to-eye with the world’s most enthusiastic technology media amid the dust and made no bones about the fact they had come to learn but also they had come to win, and that they’d sought out the best in order to do so.
Burns said GM was already beginning to incorporate camera-based and radar-based sensors in its 2007 cars and one of the key focuses of the week for the corporation was to examine the technologies and what they might offer the consumer of the future. He mentioned the company was already working on a “Virtual Valet” function which would see the company’s cars given the technology to park themselves in the not-too-distant future and he mentioned the inevitability that computers would be increasingly incorporated into road cars within the next few years for accident mitigation purposes and would eventually produce “a car that doesn’t crash.”
Utley said that semi-autonomous functionality had been at market for some time already in the Caterpillar B10 and B11 machinery, but that because of the high cost and value proposition of machinery which commonly cost more than a million dollar apiece, the technology was migrating downwards faster in her industry and had already reached the D6 dozer line.
Technology always starts with the most expensive high-end equipment first, she said, but that the time to trickle down was “getting quicker.”
Like all three of the technology companies on stage as sponsors, all had a long-standing research relationship with Carnegie Mellon but all were clearly dipping deeply into the corporate funds to help finance the team they saw as the logical winner.
As with the Team's ability to have confidence in its sensors and vision enough to run at a speed well beyond other autonomous vehciles, Professor Whittaker seemed incredibly confident in the outcome of the event. Before he passed the stage to his sponsors, he confidently invited the gathered media throng back to the same tent 24 hours hence to “enjoy the outcome.”
In robotic racing, Whittaker’s Tartan Racing is the equivalent of Formula One’s Ferrari – a team with resources equal to any in all of the fields that count - sponsorship development, technology development, logistical management under extreme duress, communication with all stakeholders, ad infinitum.
Stanford with its own charismatic leader in Dr Sebastian Thrun is not far behind, though it clearly did not have the money or resources to apply to the task.
Though its corporate sponsor list was also impressive, it never looked to be running at Tartan's speeds. Stanford looked to be the most capable of the team's shooting for podiums as Tartan was in a different race to the rest of the teams from this humble scribe’s perspective.
Some other team leaders estimated Tartan’s full-time work force to be “more than 100” compared to the “less than 20” of all the other teams (not counting Stanford). When pressed for the size of his budget, Whittaker quickly deflected the question with a gruff laugh – “we’re priceless, that’s what we are – PRICELESS.”
The parallels with automotive racing grew much stronger in this third DARPA event. Like most free market systems, there are a few haves and many have-nots and the DARPA Urban Challenge shaped up similarly.
The most logical distinction between those with a generous budget and those with mortgages, race-tape, fencing wire and camaraderie bordering on insanity was the Track A US$1,000,000 funding grant from DARPA. Awarded just over twelve months ago, eleven track A teams each received US$1 million in technology development funds from DARPA and seven of those teams reached the final held yesterday.
Like Grand Prix racing, those teams can be likened to “factory teams” and the “non-factory” or privateer teams that made the final were really in a race of their own. Three of those teams, Team Annie Way, CarOLO and Team UCF fell out of the competition in the first stage of the final, but the Students and Faculty of the University of Pennsylvania and Lehigh University should be immensely proud of the efforts of their team. They gave the best brains in existence a million dollar head start and nearly ran them down in the finish.
The analogies with racing are so many that I’m not going to bore you with any more of them in what is supposed to be a short article, bar this last one.
DARPA’s logistical efforts in creating this event have to be seen to be appreciated. Whereas all the contestants and the organizers have equivalents in the world of Formula One (an awesomely-sized enterprise in its own right), none comes close to the need to build a racetrack longer than any in existence elsewhere. The former George Air Force Base in Victorville, is used by the U.S. military to train for urban operations. The network of roads on the site effectively simulated the type of terrain American forces operate in when deployed overseas, but in turning it into an arena safe for spectators to watch robots run at road speeds in safety, it effectively built a one-off venue of a size rarely attempted.
So the big winners of the week were Tartan, Stanford, Team Victor Tango, Ben Franklin Racing and DARPA. Everyone who competed was a winner and we suspect that the continued growth of this remarkable community is assured, with some great stories still to be filed on the things we found among the dust at Victorville.