Live data from Hacker News

An electric new era for Atlas

bostondynamics.com

281–290 of 443 posts

Re: An electric new era for Atlas

#281

Earlier quoted context omitted.

Their press release actually says electric atlas is more powerful. Though I wonder if that's higher peak torque, and not so much explosive power required for jumps. A commercial robot doesn't need to do parkour.

In that case the question is why did they use hydraulics in the first place.

Static holds. Once you pressurize the cylinder to make it move to a certain position, it can hold that position without using more energy.

This makes sense for quasi-static systems but obviously is a limiting factor for dynamic robots.

Re: An electric new era for Atlas

#282

Earlier quoted context omitted.

Their previous humanoid robot, Atlas, was using hydraulics. But Spot (the dog like one) is not.

Right, and I believe only Atlas was capable of leaps and flips.

https://www.youtube.com/watch?v=xNeZWP5Mx9s

different configuration, but electric motors are fine if you get momentum on your side. Humans use their entire range of motion get build up velocity to jump; this is motion control thing.

Re: An electric new era for Atlas

#283

As we enter an era of wide scale robotic deployment, we need to think long and hard about what the maintenance bottleneck will look like. We need to advocate now for reliable and open upgrades, replacement parts, service documentation, and diagnostics. Right to repair will be even more important for this technology than autos or general computing.

Never mind right to repair, of all the advancements, maintaining the new machines has always been the obvious new job that gets created. We created the loom and fired everybody? Well now there's a loom engineer job waiting for (some) of you. What happens to society when, instead of having a robot-fixing job, the robots can fix themselves? AGI is a distraction; much like the Turing test turned out to be the wrong test. It's not the problem of how can I fix the one robot I've taken out a second mortgage to buy that I'm worried about, it's when can I buy two robots and they can fix each other that I'm worried about. Because then there is no new job being created.

Re: An electric new era for Atlas

#287

Does anyone else think the joints seem stiffer than the hydraulic version? The head and torso are receiving a lot of shock forces with each step. That seems like a downgrade from the previous one. It clearly has a much larger range of motion and if it is also stronger as claimed then I can't wait for the acrobatics videos that are surely coming. But I think the most exciting thing is that it has hands from the start.…

Electric motors dont have a lot of "give", like hydraulics do. But yes force-torque controllers can be tuned to be squishier. Someday I think electric motors will be the muscles and we'll have some kind of elastic tendons. For energy efficiency, it seems obvious to harness impact energy in a mechanical spring system, as nature does. Or just use wheels / a wheel. This whole humanoid thing strikes me as an addiction to…

It doesn't work that way.

Hydraulic systems have very little "give", unless you put a hydraulic accumulator (an air tank with a fluid/air barrier) in the system. Electric motors have plenty of "give". Forcing a motor to turn backwards won't hurt it. The gear train is usually the weak point. As motors and controllers have improved, robot gear reduction ratios have decreased, which reduces the load on the gear train and lets the motor absorb shock loads. Direct drive robots eliminate the gear train entirely. Here's a nice one.[1] "You cannot strip the teeth of a magnetic field" - General Electric electric locomotive rep, around 1900.

With modern motors, you can get huge torque with light weight, and cooling becomes the limitation. Schaft used water-cooled motors in their direct-drive robot. Google bought Schaft, ran them into the ground and killed them.

[1] https://shop.directdrive.com/products/diablo-world-s-first-d...

Re: An electric new era for Atlas

#288
post #11

If all Boston Dynamics did was make Youtube videos, they'd have a pretty good business.

Is that their business? They've been around for 30+ years and I don't think they've ever successfully commercialized a product. So far as I can tell, they just hop from DARPA grant to DARPA grant and make cool videos of the results. I don't have any particular problem with that, but its a little weird? I figured they were a more traditional industrial robotics company that just did the humanoid robots as a side line…

Boston Dynamics needs a sugar daddy to subsidize them. First it was DARPA. Then Google. Now Hyundai. Their real achievement is that their management has been able to keep the money flowing for three decades.

Re: An electric new era for Atlas

#289
post #54

Earlier quoted context omitted.

The problem with that “1960s vision” is thinking to literally about having robots do exactly what humans do now. Likely there is a creative way to solve the need for humans to do those tasks via automation, but it’s not likely to look like a humanoid robot folding laundry.

Agreed. On the other hand, it'd be highly amusing if the future did involve humanoid robots out mowing the grass with a push mower, or getting into their car to drive to the grocery store. I'm reminded of this Adam Savage video of BD's Spot pulling a rickshaw (23:00). https://www.youtube.com/watch?v=zyaocKS3sfg Maybe in the future the passenger will be a humanoid robot being taken to its laundry folding job?

Humanoid robots making buggy whips to control ornery LLM-based robotic horses. With private/pair key encryption in the whips themselves- they can send a digitally signed wireless "threat" before actual contact is needed.

Re: An electric new era for Atlas

#290

A hill I'm willing to die on: bipedal robots are an evolutionary path that machines don't need to go down, we have lovely bearings and wheels that work perfectly with electric motors. Yes obviously there are limitations i.e. stairs and uneven terrain but there are wheeled/tracked solutions for those too Most of these robots will be used in factories that have very nice flat concrete floors

If we ever see a world where robots need to be useful outside of a factory with perfectly flat concrete floors, then yes - there needs to be continued evolution in traversal over uneven ground and around unanticipated objects. Bipedal locomotion is useful for this (although not the only solution). Right now the hardest jobs to replace will be those of plumbers, electricians, carpenters, etc where they need to operate…

i think search and rescue is a great application of humanoid robotics, you need something very versatile and a human body is not a bad model for a universal terrain form factor.
Post reply on HN