Games Gadgets n Technology
Sunday, November 04, 2007
No driver, no worry
Autonomous robotic cars race through simulated life-like road conditions, throwing up the possibility of unmanned vehicles for military as well as civilian purposes
The race began with mechanics moving eleven cars to the starting line. The monsters roared their engine at the starting line. But at the fateful moment, the drivers abandoned their vehicles, and the cars began moving on their own, eliciting cheers from thousands of spectators!
Robotic cars built by prominent universities in the US competed Saturday in a high-stakes race organised by the US military’s Defense Advanced Research Projects Agency (DARPA), which hopes to use driverless vehicles as weapons on urban battlefields by 2015.
The competition was staged at an unused military base, and promises a first-place prize of $2 million.
 |
| A member of the MIT team tunes their entry’s computer before the start of the race |
The cars picked up pace as they found their way along streets of the closed-down base without any help.
The challenge sent them along neighbourhood roads, through traffic and around jams created by humans. About 50 humans piloted other cars to simulate real road conditions.
Over and over, cars with a warning honk and roof-rack of space-age gear came to a perfect halt at the stop sign of a deserted intersection, then pulled through smoothly.
In fact, the biggest error was a minor ‘fender bender’, which did not stop or significantly damage either vehicle.
“There’s more computing power in the back (of the cars) than most companies have,” said Paul Barrett, team leader of the Massachusetts Institute of Technology’s (MIT) entry, while describing the technology that went into their modified Land Rover.
Only seven contenders remained by the end of the race, and Stanford University’s ‘Junior’, crossed the finish line first.
Carnegie Mellon University’s entry finished second, while Virginia Tech took third place.
Finishing first does not guarantee victory because speed is just one of several criteria used to determine who will be the champion and get the prize money.
Competitors are also rated for precision and performance while navigating the course.
Of the five finalists who bogged down early, one car ended up in a driveway with its sensors continuing to swivel, while an 11-tonne self-controlled green truck called the TerraMax halted inches away from mowing down a column.
In qualifying events, some robot cars simply stopped, lost in thought, climbed over curbs and sideswiped parked vehicles.
However, this scenario is still a lot better than the first DARPA race held in 2004, which had no finishers.
While the teams get fame and large cash rewards, it’s the US military that hopes to be the real winner, by gaining technology that would enable the military to make a third of its vehicles robotic by 2015.
Manufacturer Oshkosh trucks, which built the TerraMax, supplies the US military in Iraq, and a driverless version is exactly what is needed to cut the number of soldiers’ lives at risk in battle.
Universities also see interesting artificial intelligence problems to solve, while corporations see the building blocks of an automobile of the future.
However, there’s enough interest for the common man as well, as hundreds of spectators turned out for the event.
Along with the military uses, DARPA sees the race as ensuring better public safety as well.
While the cars may not be ready to hit the vehicle market just yet, DARPA said the contenders should throw up a few options in the forseeable future.
Inside the Winner
Here’s a look at what goes inside Stanford University’s ‘Junior’, which finished first.
The body
Model: 2006 Volkswagen Passat wagon
Engine: 4-cylinder turbo diesel injection
Transmission: 6-speed direct-shift gearbox
Fuel Consumption: City: 10.8 km/l;
Highway: 18.1 km/l; Combined: 14.7 km/l
Top speed: 203km/h
Acceleration: 0-100km/h in 10.1sec
The senses
Cutting-edge sensors and custom AI software enable Junior to determine its position and perceive its surroundings, day or night, even in adverse GPS conditions. To measure its localisation and perception, the car uses an array of Lidars – a measuring system that detects and locates objects on the same principle as radar, but uses light from a laser. It can estimate positions with an accuracy of within 5cm.
The brainsHardware: High-tech servers crunch sensor data up to 20 times a second, and run Junior’s artificial intelligence software.
Software: Junior’s intelligence comes from a suite of integrated, custom-coded programs that include four major components: a planner to make decisions and choose paths; a mapper to transform sensor data into environment models; a localiser to refine GPS position and road map structure from lane markings; and a controller to turn decisions into driving.
posted by strangebeauty at 8:53 PM
0 Comments:
Post a Comment
<< Home