4 ms·

The mind-blowing thing is pricing - this thing starts at just $16K. We’re going to witness an explosion of affordable humanoid robots in every home & business,

The mind-blowing thing is pricing - this thing starts at just $16K. We’re going to witness an explosion of affordable humanoid robots in every home & business, and as they become increasingly capable the use cases will be nearly limitless. Think deliveries, cleaning, even childcare (Bicentennial Man, anyone?). I think what will be shocking is just how fast this happens, and they’ll likely outnumber humans at some point.

verdverm 2y ago
Stanford Aloha (open sw/hw) is around $32k with all 4 arms (2 for training) https://mobile-aloha.github.io/ https://mobile-aloha.github.io/ There was this video last week of a "westworld" factory in China too: https://newatlas.com/robotics/chinese-humanoid-robots-realistic/ https://newatlas.com/robotics/chinese-humanoid-robots-realis...
Animats 2y ago
Are they direct driving most of the joints? It looks like they have a direct-drive motor that can operate joints with no gears.[1] Is that right? If they use all direct drive motors, these robots are mechanically simple. No gears, no strings, no pulleys, maybe no separate bearings. A production line for a custom motor is expensive to set up, but once production is running, each motor is cheap. An encoder is cheap if designed into the motor. This is the route to volume production. Machine learning has made the control problem much easier. The classic approach was to work out the dynamics analytically. (I used to struggle with that in the 1990s). That's the old Boston Dynamics approach. Now that's obsolete. BD must be using some machine learning by now. The video of Atlas doing a flip showed the planning process briefly. It was very pre-planned. They must be beyond that by now. [1] https://shop.unitree.com/products/go1-motor https://shop.unitree.com/products/go1-motor
- margalabargala 2y ago
  The motor you linked, and what they are using, is not direct drive. That motor has a 1:6.33 integrated gearbox.
  - Animats 2y ago
    Too bad. It looks a lot like these real direct drive motors.[1] Those are in the same price range, and could get much cheaper if produced in quantity. The nice thing about direct drive motor robots is that you can't strip the teeth of a magnetic field. Impact forces are not a big problem. Robot gear trains are usually the weak point. [1] https://shop.directdrive.com/collections/m15-series https://shop.directdrive.com/collections/m15-series
    - Scouttman 2y ago
      The 8Nm/kg is impressive for the direct driver servo. It's still however a long way off from the 43.3Nm/kg gained by adding a gear box. Direct drive motors have other advantages such as better torque control and no backlash but out side of very novel motor design or making motor diameters impactical it's hard to beat the torque density of a gearbox. Maybe one day.
      - Animats 2y ago
        Watching those direct-drive robots jump and land smoothly is impressive. No problem with shock loads. They look like they're on springs, but there are no springs; it's all magnetic fields in the motor. With good temperature monitoring and over-speced MOSFETS, those direct drive motors can probably be way overdriven for a few hundred milliseconds. That's probably the future. You don't need that extra power all that often, but sometimes you do.
    - maxwell 2y ago
      Gear trains are the killer robots' weak point, eh? Good to know.
delfinom 2y ago
>We’re going to witness an explosion of affordable humanoid robots in every home & business, and as they become increasingly capable the use cases will be nearly limitless. For the first year, after that society will collapse as the largely now unemployed low income and middle class arm up and start pillaging for survival.

verdverm 2y ago

Stanford Aloha (open sw/hw) is around $32k with all 4 arms (2 for training) https://mobile-aloha.github.io/ https://mobile-aloha.github.io/ There was this video last week of a "westworld" factory in China too: https://newatlas.com/robotics/chinese-humanoid-robots-realistic/ https://newatlas.com/robotics/chinese-humanoid-robots-realis...

Animats 2y ago

Are they direct driving most of the joints? It looks like they have a direct-drive motor that can operate joints with no gears.[1] Is that right? If they use all direct drive motors, these robots are mechanically simple. No gears, no strings, no pulleys, maybe no separate bearings. A production line for a custom motor is expensive to set up, but once production is running, each motor is cheap. An encoder is cheap if designed into the motor. This is the route to volume production. Machine learning has made the control problem much easier. The classic approach was to work out the dynamics analytically. (I used to struggle with that in the 1990s). That's the old Boston Dynamics approach. Now that's obsolete. BD must be using some machine learning by now. The video of Atlas doing a flip showed the planning process briefly. It was very pre-planned. They must be beyond that by now. [1] https://shop.unitree.com/products/go1-motor https://shop.unitree.com/products/go1-motor

margalabargala 2y ago
The motor you linked, and what they are using, is not direct drive. That motor has a 1:6.33 integrated gearbox.
- Animats 2y ago
  Too bad. It looks a lot like these real direct drive motors.[1] Those are in the same price range, and could get much cheaper if produced in quantity. The nice thing about direct drive motor robots is that you can't strip the teeth of a magnetic field. Impact forces are not a big problem. Robot gear trains are usually the weak point. [1] https://shop.directdrive.com/collections/m15-series https://shop.directdrive.com/collections/m15-series
  - Scouttman 2y ago
    The 8Nm/kg is impressive for the direct driver servo. It's still however a long way off from the 43.3Nm/kg gained by adding a gear box. Direct drive motors have other advantages such as better torque control and no backlash but out side of very novel motor design or making motor diameters impactical it's hard to beat the torque density of a gearbox. Maybe one day.
    - Animats 2y ago
      Watching those direct-drive robots jump and land smoothly is impressive. No problem with shock loads. They look like they're on springs, but there are no springs; it's all magnetic fields in the motor. With good temperature monitoring and over-speced MOSFETS, those direct drive motors can probably be way overdriven for a few hundred milliseconds. That's probably the future. You don't need that extra power all that often, but sometimes you do.
  - maxwell 2y ago
    Gear trains are the killer robots' weak point, eh? Good to know.

delfinom 2y ago

>We’re going to witness an explosion of affordable humanoid robots in every home & business, and as they become increasingly capable the use cases will be nearly limitless. For the first year, after that society will collapse as the largely now unemployed low income and middle class arm up and start pillaging for survival.