New York City's infamous taxicab drivers are up in arms. In the past few weeks they've gone on strike twice, protesting the installment of Global Positioning System satellite receivers and transponders in their cars. The Taxi Workers Alliance, representing about 20% of the city's 40,000-plus cab drivers, doesn't want GPS equipment in its cars.
New York City fathers, on the other hand, want every last cab to be equipped with interactive GPS units, video systems, and credit card readers. Indeed, under the city's new tech-savvy rules, taxis will be required to have this gear on board to pass vehicle inspection.
If the cabbies think New York City's new rules are onerous, wait until they encounter DARPA. If the Defense Department's Advanced Research Project Agency has anything to do with it, legalized GPS tracking is just the beginning of their problems.
Dull, Dirty, and Dangerous
In the 2001 National Defense Authorization Act (Public Law 106-398), the US Congress mandated (under Section 220) that the Armed Forces achieve widespread fielding of unmanned remote-controlled technology. By 2015, Congress expects one-third of the military's ground combat vehicles will be unmanned. They'll be able to perform dull, dirty, and dangerous tasks, carried out using machines instead of humans.
Even as New York's cabbies protest, DARPA pushes ahead. As a famous New York haberdashery says in its ad slogan, "Money talks, and nobody walks." Money does indeed talk, and this time it's coming from DARPA's Urban Challenge contest, an annual autonomous vehicle R&D race about to get under way (November 2007) about the time you read this.
To date, two DARPA Grand Challenge autonomous vehicle competitions have been run. The first, held in March 2004, featured a 142-mile course out in the desert. That's a far cry from the canyons of Manhattan, but the course was grueling nonetheless. Fifteen vehicles started the race. Not one finished. In the 2005 Grand Challenge, however, four autonomous vehicles successfully completed the route.
This year, DARPA's Urban Challenge features self-driving vehicles rolling and maneuvering in a mock city environment. The vehicles in DARPA's contest, although executing simulated military supply missions, must merge into moving traffic, navigate traffic circles, negotiate intersections, and avoid obstacles. That sounds like the New York City driving experience to me.
Big Money
The program is being conducted as a series of qualification steps leading to a competitive final event scheduled for November 3rd.
DARPA is offering a cool $2 million prize for the fastest qualifying vehicle, with $1 million and $500,000 for second and third places.
Check out the criteria for judging success, and see if you agree that the contest rules would make for a better experience in the back seat of a New York City taxicab. Keep in mind the objective of the DARPA program is safe and correct autonomous driving capability, in traffic, at 20 mph.
Overall, DARPA Grand Challenge autonomous vehicles must tour a series of checkpoints over a complex route (entering vehicles have five minutes to process a mission description before attempting the course). Cars must also interpret lane markings, such as white and yellow lines, and behave in accordance with conventional traffic laws.
Cars in the contest must also prove context-dependent speed control to ensure safe operation, including adherence to posted speed limits. They must also exhibit safe-following behavior when approaching other vehicles from behind. No tailgating, thank you.
Robot cars must, additionally, exhibit safe check-and-go behavior when pulling around a stopped vehicle, pulling out of parking spot, moving through intersections, and in situations where collision is possible. They have to stay on the road in a legal and appropriate travel lane en route, yet make sharp turns, and move smartly through intersections (a definition file specifies GPS coordinates of stop signs).
DARPA Challenge vehicles must also be able to navigate in areas where GPS signals are partially or entirely blocked, and they have to follow paved and unpaved roads, and stay in lane when encountering sparse or low-accuracy GPS waypoints. They can change lanes when legal and appropriate, such as when passing a vehicle or entering an opposing traffic lane to pass a stopped vehicle, but may not pass other vehicles queued up at intersections.
Vehicles must also merge safely with traffic moving in one or more lanes after stopping at an intersection, and be able to pull across one lane of moving traffic to merge with moving traffic in an opposing lane. They must also stop safely within one meter of a stop line at a stop sign, and proceed in less than ten seconds, abiding by precedence rules.
DARPA Grand Challenge vehicles must be able to navigate toward a destination in a large, open area where minimal, or no, GPS points are provided, such as loading dock areas and parking lots. They even have to be able to pull into, and back out of, parking spaces. Finally, competing cars also have to demonstrate the ability to execute one or more three-point turns, and make a U-turn. All of this while using sensors to recognize pedestrians, traffic lights, and stop signs.
Back in the real world of Manhattan, New York City cab drivers maintain GPS trackers are an invasion of their privacy. They also argue credit card readers make them vulnerable to payments being rejected, especially after a customer has left a cab. Of course, the drivers know they'll now have to declare every cent of fares they get, including tips.
What these cab drivers don't know is that one of these days sensor-equipped autonomous vehicles may supplant human taxicab drivers. The DARPA Grand Challenge proves it's possible. Which driver would you prefer?
