Bipedal Robot Sets Guinness World Record For Robotic 100-Meter Sprint (newatlas.com) 32
A droid named Cassie has set a Guinness World Record for the 100-meter dash by a bipedal robot, "an impressive demonstration of robotics and engineering," reports New Atlas. From the report: Cassie is the brainchild of Agility Robotics, a spin-off company from Oregon State University, and was introduced in 2017 as a type of developmental platform for robotics research. And Cassie has continued to come along in leaps and bounds since then, in 2021 demonstrating some impressive progress by completing a 5-km (3.1-mile) jog in just over 53 minutes. This achievement involved the use of machine learning algorithms to equip the robot with an ability to run, overcoming its unique biomechanics and knees that bend like an ostrich to remain upright. With this capability, Cassie joined a group of running bipedal robots that include the Atlas humanoid robot from Boston Dynamics and Mabel, billed as the world's fastest knee-equipped bipedal robot. But in optimizing Cassie for the 100-meter sprint, the researchers had to head back to the drawing board.
The team spent a week fast-tracking Cassie through a year's worth of simulated training designed to determine the most effective gait. But it wasn't simply a matter of speed. For the Guinness World Record to stand, Cassie had to start in a standing pose, and then return to that pose after crossing the finish line rather than simply tumble over. This meant Cassie had to use two neural networks, one for running fast and one for standing still, and gracefully transition between the two. Ultimately, Cassie completed the 100-meter sprint in 24.73 seconds, establishing a Guinness World Record for a bipedal robot. This is a great deal slower than the sub-10-second times run by the world's best sprinters, but the researchers believe progress will only accelerate from here. You can watch Cassie's record-setting dash here.
The team spent a week fast-tracking Cassie through a year's worth of simulated training designed to determine the most effective gait. But it wasn't simply a matter of speed. For the Guinness World Record to stand, Cassie had to start in a standing pose, and then return to that pose after crossing the finish line rather than simply tumble over. This meant Cassie had to use two neural networks, one for running fast and one for standing still, and gracefully transition between the two. Ultimately, Cassie completed the 100-meter sprint in 24.73 seconds, establishing a Guinness World Record for a bipedal robot. This is a great deal slower than the sub-10-second times run by the world's best sprinters, but the researchers believe progress will only accelerate from here. You can watch Cassie's record-setting dash here.
Gait (Score:1)
Still waay slower and technologically inferior to humans human. First, the average fat slashdotter level human can run 100 meters a full 10 seconds faster than that robot. Second, that robot's gait is all F'd up .. it's gait cheating. Make one proper bipedal like humans, and with toes. Let's see a robot run on its toes.
Re: (Score:3)
First, the average fat slashdotter level human can run 100 meters a full 10 seconds faster than that robot.
So, in 14.73 seconds? An average slashdotter, let alone a fat one? Doubt.
Either way, give them some time. I suspect they'll be beating Usain Bolt sooner rather than later -- at least in raw speed under ideal conditions, anyways.
Re: (Score:2)
Re: (Score:3)
The average for humans is about 14 seconds.
Where did you get this number? I can see 14 as an average for somewhat trained runners who put in an all-out killer effort one time. Average for all humans? No. I can't see that. Back when I ran fast, I could put out 75 to 80 second quarters. Not great, not too shabby. Around 6:00-ish minutes per mile, which is probably faster than the average person can do it. And it was a good day if I saw a 14 second 100m.
Re: (Score:2)
sadly.
by crossing the lines.
the time is ignored.
and the robot disqualified
Re: (Score:2)
https://trackspikes.co.uk/aver... [trackspikes.co.uk]
For your link the average speed at age 23 is 10.89.
If you tried to figure out average 10k times by surveying people who hit the Boston Marathon Qualifying standard you'd actually have a less biased sample.
Re: (Score:2)
Your link isn't really clear, but it definitely seems to give average times for runners who actually enter 100m racing events, not average fat slashdotter level humans, who probably for the most part don't ever run at all unless there's an emergency.
Re: (Score:2)
Your link isn't really clear, but it definitely seems to give average times for runners who actually enter 100m racing events, not average fat slashdotter level humans, who probably for the most part don't ever run at all unless there's an emergency.
The link was from the AC, and seems to be a possible source of the 14 second figure people were using.
I'd also note that especially in the context of signing up for a race the 14 second average across all age grounds they have is a completely meaningless figure (if you ran 14s at a 20 year old at that event you'd be embarrassed, if you did it as a 60 year old you'd be a marvel).
Re: (Score:2)
The source for the "14 second figure people were using" is the top-level comment of this thread :
First, the average fat slashdotter level human can run 100 meters a full 10 seconds faster than that robot
(and the robot's time was 24.73 seconds.)
Re: (Score:2)
Stick a bag of cheetohs and some porn at the finish line. You'll see under 14 seconds for even the fattest of the fat. Hook up a WOW session beside the porn and see under ten seconds.
Re: (Score:2)
The average for humans is about 14 seconds. having said that I would be shocked if this community wasn't well north of that average and no way in hell any of the fat ones are doing it in 14 seconds.
Maybe an average active 20-30 male.
Even someone who's sub-13 is considered fast for an average recreational sports athlete.
Re: (Score:1)
Re: (Score:2)
Re: (Score:2)
Wait until they discover the wheel! (Score:1)
And if nobody can think of any compelling use cases for bipedal over wheels, that's pretty discouraging as to their real-world utility.
Re: (Score:2)
Re: (Score:3)
Seriously? Rocky terrain, mountainous regions, muddy regions, collapsed buildings.. etc. and combinations thereof.
Re: (Score:2)
Re: (Score:2)
Very true, but bipedal beats a wheel in those situations. Unless we are talking wheeled bipedal like Boston Dynamics Handle.
Re: (Score:2)
Re: (Score:2)
But seriously I wonder what kind of track and field event would favor bipedal robots over rolling ones. Hurdles, maybe?
High jump? Long jump?
Not a very smart robot (Score:2)
Instant DQ! (Score:2, Informative)
Re: (Score:2)
Surely the robot has to stay within its lane to have a valid time? Otherwise it's an instant disqualification, and that thing was all over the place! I'd forgive it if it had a fricken laser beam attached to its fricken head, but alas it doesn't have a head, or even a torso for that matter (looks ridiculous).
I actually found that interesting, how "organic" the clumsiness felt. I suspect that a result of the ML used to drive the motions.
In reality the fact it was clumsy impressed me more than a very stable straight line sprint since it showed the robot was able to maintain its balance and momentum while off-balance.
Olympics++ (Score:2)
Haha. (Score:2)
More PR BS. Even the fat guy filming it hardly had to work to keep up.