The previous strongest argument for “why make a humanoid robot?” was that the world is already designed for humans to navigate and manipulate. But there are many robot designs that could navigate the world equally well, especially if we only care about a subset of tasks and environments. The new strongest argument is that if the robot is humanoid then you can pre-train its policy on human actions.
For navigation, I’d assume that wheels are easier, even for humans to control, because you don’t need to worry about balancing the robot, and humans are certainly good enough at using a joystick. Plausibly humanoid hands could be more intuitive though.
Before continuing, I should say that overall I think humanoid robots is not a good idea.
Wheels work for flat ground but not for staircases. Afaict wheel on leg hybrids handle staircases well but at that point it looks like it is more complex than having 2+ legs. You can also have wheel extensions for humanoid robots, im pretty sure ive seen it in Unitree’s videos.
Controlling any sort of robot to do basic navigation, is something that I imagine will be close to fully automated that it doesn’t matter in the discussion of humanoid vs other robot. Though, do you know humanoid robots handle the balancing themselves, and that people do use controllers to control without needing to wrry about balance?
The fact that humanoid robots can handle a larger set of tasks and environments is also its appeal exactly because mass production can reduce cost significantly.
I really sound like a humanoid robot proponent but I just don’t think these are that good of arguments against it. I can write a longer list of issues I have with humanoid robots when the time comes.
The previous strongest argument for “why make a humanoid robot?” was that the world is already designed for humans to navigate and manipulate. But there are many robot designs that could navigate the world equally well, especially if we only care about a subset of tasks and environments. The new strongest argument is that if the robot is humanoid then you can pre-train its policy on human actions.
or that it is easier for a human control a humanoid robot I guess.
Humanoid robots controlled by surgeons did world-first operation on live pigs
For navigation, I’d assume that wheels are easier, even for humans to control, because you don’t need to worry about balancing the robot, and humans are certainly good enough at using a joystick. Plausibly humanoid hands could be more intuitive though.
Before continuing, I should say that overall I think humanoid robots is not a good idea.
Wheels work for flat ground but not for staircases. Afaict wheel on leg hybrids handle staircases well but at that point it looks like it is more complex than having 2+ legs. You can also have wheel extensions for humanoid robots, im pretty sure ive seen it in Unitree’s videos.
Controlling any sort of robot to do basic navigation, is something that I imagine will be close to fully automated that it doesn’t matter in the discussion of humanoid vs other robot. Though, do you know humanoid robots handle the balancing themselves, and that people do use controllers to control without needing to wrry about balance?
The fact that humanoid robots can handle a larger set of tasks and environments is also its appeal exactly because mass production can reduce cost significantly.
I really sound like a humanoid robot proponent but I just don’t think these are that good of arguments against it. I can write a longer list of issues I have with humanoid robots when the time comes.