Highlights from the 2014 LA Auto Show

Cheetah robot outpaces Usain Bolt in setting new robotic land speed record

By

September 6, 2012

DARPA's Cheetah robot has set a new robotic land speed record of 28.3 mph (45.5 km/h)

DARPA's Cheetah robot has set a new robotic land speed record of 28.3 mph (45.5 km/h)

Usain Bolt might have taken home the 100 meters gold medal from the recent London 2012 Olympics, but things could have been different had DARPA’s Cheetah robot been allowed in the field. Living up to DARPA's original goal of developing a robot that could outrun any human, the quadruped robot has set a new robotic land speed record of 28.3 mph (45.5 km/h) for a 20-meter (65.6 ft) split, bettering Bolt’s human speed record by 0.52 mph (0.8 km/h).

The Cheetach robot’s 28.3 mph run on a treadmill broke its own previous record of 18 mph (28.9 km/h) set earlier this year and outpaced the fastest man on Earth, who, according to the International Association of Athletics Federations, set the world record for a human in 2009 when he reached a peak speed of 27.78 mph (44.7 km/h) for a 20-meter split. Although DARPA points out, the fact Cheetah ran on a treadmill equates to a 28.3 mph (45.5 km/h) tail wind advantage.

Being developed and tested by Boston Dynamics under DARPA’s Maximum Mobility and Manipulation (M3) program, Cheetah is being designed to perform in emergency response, humanitarian assistance and military missions. With such applications certain to involve rough terrain, DARPA intends to test a prototype on natural terrain next year.

In its current version, Cheetah is powered by an off-board hydraulic pump with a boom-like device keeping it centered on the treadmill. DARPA says improved control algorithms and a more powerful pump were responsible for the increase in speeds since the robot set its previous record.

Video of Cheetah’s record breaking run appears below.

Source: DARPA

About the Author
Darren Quick Darren's love of technology started in primary school with a Nintendo Game & Watch Donkey Kong (still functioning) and a Commodore VIC 20 computer (not still functioning). In high school he upgraded to a 286 PC, and he's been following Moore's law ever since. This love of technology continued through a number of university courses and crappy jobs until 2008, when his interests found a home at Gizmag.   All articles by Darren Quick
17 Comments

Mhhh, good work, they are getting there, but it won't be the real thing until the hydraulic pump comes on-board.

Riaanh
6th September, 2012 @ 04:00 am PDT

"In its current version, Cheetah is powered by an off-board hydraulic pump" - well, I can understand off-board power supply because batteries are not that good yet and THE REST of technology of robotics should continue developing because batteries WILL become better, but what's the point of developing something which can not function without most essential of its part being somewhere close connected with pipes?

Robert Silagadze
6th September, 2012 @ 04:18 am PDT

Interesting, but it looks like the black n yellow beam that is attached is giving it stability. If it could achieve this speed balancing itself, I would be properly impressed.

More so, if it could do it with it's own power supply, even if only for a short burst of time.

Brendon Walker
6th September, 2012 @ 04:25 am PDT

Interesting, but too unstable looking to be very practical in my opinion.

Quadruped is not at all ideal, but just what we are used to.

Any decent attempt with robotics should have at least 6 legs, and maybe 8.

Rigby5
6th September, 2012 @ 09:23 am PDT

Cool, but I'll be a whole lot MORE impressed when its propulsion system is entirely on-board. Granted, the tether does look slack, so it's probably not stabilizing the gizmo.

rocketride
6th September, 2012 @ 09:36 am PDT

Some part falls off at 0.13 / 0.14. Imagine what would happen in rough terrain.

Can it make corrections based on terrain and moving targets at that speed like animals and humans can?

Siddharth Mehta
6th September, 2012 @ 10:51 am PDT

Exit foot, to the left, at about 0:14.

The first thing I notice is that the left and right feet are operating in sync with each other. No animal actually runs this way. Springboks and similar animals use all four legs together in hops. But generally, animals use each leg separately. Well, not really seperately -- in sequence. The way I like to imagine it is to imagine a wheel. The wheel has four spokes, all on the same side of the wheel, equally spaced for the basic configuration. I think that should be the goal.

Also, add wrists and ankles. There need not be power to flex these joints. They can function in a springlike fashion. Use something like Oscar Pistorius used in the Olympics.

Victor Engel
6th September, 2012 @ 11:42 am PDT

The cheetah perhaps shouldn't be designed to run faster than the average human, since someone is going to need to run alongside it holding the tether, power supply, hydraulic pump equipment, while someone else will need to run alongside it to hold the controller arm. It's a nice experiment, but still has a long way to go.

Laura Ward
6th September, 2012 @ 12:18 pm PDT

Ah, Laura, then I guess you haven't heard about the CSV (Cheetah Support Vehicle) being developed for just that purpose. It's the size of a tank, contains a massive power generator, and supports the Cheetah with a large boom. Goes 30 mph, so no worries.

Arf
6th September, 2012 @ 01:11 pm PDT

Arachnid model would be more effective.

Dawar Saify
6th September, 2012 @ 02:01 pm PDT

Was it drug tested for being a cheater? Couldn't resist :)

Max Gallo
6th September, 2012 @ 06:22 pm PDT

meh, unimpressive. let me know when all the thethers and stuff isn't there, because that's just pure cheating.

SpaceBagels
6th September, 2012 @ 06:52 pm PDT

Very impressive.

You're looking at the Wright Brothers version of flight on a small budget.

Imagine what the SR-71 version will be in 30 years. No one will be out-running that.

Stagger_lee69
6th September, 2012 @ 09:40 pm PDT

Stagger, I think the Wright Brothers had a small budget. DARPA doesn't have small budgets, or they may be stingy with certain projects. I will be impressed when it can beat ursain bolt in the standing 100m.

Hang on doesn't someone have a donkey that can at least stand up, or that is DARPA again?

Next for a robotic racehorse which can win the Kentucky Derby.

MD
7th September, 2012 @ 01:11 am PDT

I dont understand why they dont simply map the movements of what ever animal they want, plotting how it moves and program it accordingly. I know game designers do this for advance game characters. Simply start with what is know, and work from there with an adaptive/intuitive system.

work on basic fluid movements, and speed will come later. Otherwise y our wasting time.

yinfu99
7th September, 2012 @ 09:10 am PDT

Max Gallo has the best post here. As for the others, gee whiz I am pretty sure they explored some of the options mentioned and as for the tether I suspect it is just part of natural progression.

I'm sure those who have to wear prosthetic limbs are glad that some inventors/manufacturers released their early imperfect models.

It's difficult to go from an idea to the final solution (if there is such a thing) in the first iteration. Speaking as a robotics engineer I think it's a great effort and I'm sure they are looking forward to perfecting issues. Wait a minute I'm not a robotics engineer? Oh well, same statement applies.

I'm glad Max Gallo didn't resist, great bit of subtle humour.

Dr. Veritas
8th September, 2012 @ 10:03 am PDT

Change the programming and run the belt the other way and see how it works.

It's built like a cat but they're running it in reverse.

Gregg Eshelman
12th September, 2012 @ 02:05 pm PDT
Post a Comment

Login with your gizmag account:

Or Login with Facebook:


Related Articles
Looking for something? Search our 29,509 articles