Search This Blog

7.4.12

陰謀論: 美國軍事部門DARPA尋求更逼真的人形機械人來應付即將到來的大挑戰

DARPA seeks more lifelike humanoid robots in upcoming Grand Challenge

Madison Ruppert, Contributor
Activist Post

The Defense Advanced Research Projects Agency (DARPA) is now looking for more lifelike humanoid robots to add to the military’s arsenal, which already has some quite creepy devices in the work including the SAFFiR robot, the jumping “sand flea” robot and an unbelievably fast legged robot.

DARPA’s next contest will task robotics specialists around the globe with the behemoth challenge of creating robots capable of navigating their surroundings and even handle tools with skills and dexterity near that of a human. One must expect that DARPA will have high standards this year given the recent leaps forward the military has made when it comes to robotics.


Indeed, recently we have seen DARPA fund the creation of the world’s fastest legged robot known as the Cheetah, the successor to Big Dog, known as Alpha Dog, which has the ability to take commands from humans and carry hundreds of pounds of gear through rough terrain, a robot capable of jumping onto buildings, and a robotic appendage that can answer phone calls and open doors autonomously.



DARPA – the Pentagon’s mad scientist research arm – is expected to announce their new contest in the next few weeks.

They are expected to ask engineers and designers to create a bipedal (two-legged) robot capable of a wide variety of tasks ranging from opening doors to driving cars, crossing rough terrain and conducting industrial repair activities.


The theme of DARPA’s upcoming Grand Challenge was leaked to the robotics-focused website Hizook, and later confirmed by Kent Massey.

“The goal of this Grand Challenge is to create a humanoid robot that can operate in an environment built for people and use tools made for people,” said Massey in an email to Hizook. “The specific challenge is built around an industrial disaster response.”

Massey is the director of advanced programs at HDT Robotics and he reportedly attended a recent speech during which the Grand Challenge was outlined.

The speech was held by Dr. Gill Pratt, DARPA’s program director. Dr. Pratt briefed the audience on the upcoming Grand Challenge and Massey’s account was reportedly confirmed by other attendees contacted by Danger Room.

The focus of industrial disaster response seems similar to other military robotics project like SAFFiR, which is designed to assist firefighting efforts on submarines and other naval vessels.

In recent years, DARPA has put a larger focus on these Grand Challenges, with the earlier initiatives focusing on self-driving cars.

These efforts helped push forward the development of autonomous vehicles considerably, changing the notion of self-driving vehicles from science fiction to reality.

Like many DARPA efforts, before they got involved, the area seemed wildly fantastical and wholly unrealistic. Obviously they changed that quite a bit.

Partially thanks to DARPA’s push, autonomous vehicles have improved quite a bit, even allowing a legally blind man to take a drive in one of the self-driving vehicles.

Back in 2004, DARPA began their first Grand Challenge, worth $1 million, which asked researchers to create autonomous vehicles able to travel off-road, although not a single car was able to travel over eight miles on the course.

In 2005, DARPA doubled the prize money to $2 million, and the challenge saw much greater success. This time around, five autonomously controlled vehicles were able to complete the course, a total of 132 miles, with a Sport Utility Vehicle (SUV) built by scientists from Stanford taking the prize money.

Then in 2007, DARPA gave $2 million in prize money to a collaborative team from Carnegie Mellon and Tartan Racing. Their entry was not only able to drive itself, but it was also capable of many other astounding feats like avoiding other autonomous vehicles and even obeying traffic signs.

It is hardly a leap to see why DARPA would shift their focus to humanoid robots after their previous efforts have proven to be highly successful.

It is worth pointing out that this newest challenge will open up greater opportunities to much smaller (perhaps even amateur) operations, whereas building a robot car would require resources only available to institutions like Stanford and Carnegie Mellon.

It might also give players with fewer resources the ability to begin to compete with giants like Boston Dynamics, which is behind some of the world’s most famous military robots like the four-legged Alpha Dog and many more.

DARPA will only provide funding to a dozen different teams, but that will not stop unpaid teams from entering the competition.

“If some of the unpaid teams perform better than the paid teams … they will begin receiving funding,” according to Massey.

Indeed, in 2005, it was quite shocking to see that one of the finalists was not a team made up of incredibly well-funded scientists from a major university but actually a group of individuals from an insurance company in Louisiana.

Massey says that this year’s contestants will likely have to create some astounding machines capable of “get[ting] into the driver’s seat [of a vehicle] and drive it to a specified location.” Such a robot might then exit the vehicle, stroll towards a locked door and then “unlock it, open the door, and go inside.”

That very well may not be the end of the robots travels, according to Massey.

Such a robot might be instructed to “traverse a 100 meter, rubble strewn hallway” after which it the human-like robot “locates a pipe that is leaking” and repairs it.

DARPA has yet to publicly release the full details of this year’s Grand Challenge, but what little information we have available to us indicates that this project will not be a walk in the park, even for the most advanced experts like the ones at Boston Dynamics.

Yet, considering the fact that having a robotic car drive a 132-mile course seemed near impossible just years ago, I believe we should steer clear of thinking of this year’s task as unattainable.

It will be fascinating to see what contestants come up with, and even more interesting (or potentially disturbing) to see how these humanoid robots are actually rolled out in the military arsenal.