I’d go further to argue that it’s very necessary for robots to have even more mobility such as wheels to take advantage of momentum and increase efficiency. There’s a reason the wheel was invented, things like bicycles, wheelbarrows, etc.
The ANYmal wheeled legged robot is an example of that.
I’d argue that the wheel was invented not because “walking” was inefficient, but because flesh is weak and gets tired.
A robot doesn’t have that weakness. It thinks nothing of running five hours at high speed if necessary. It has no need of wheels if it can just Gump it cross country with cargo on his back a la Death Stranding.
Robots have battery capacity limitations, they get “tired” in a different way. Your claim is true if you invent a battery that never runs out of power.
I wonder if someday in the future we might use reinforcement learning to iterate over different mechanical designs to explore even more exotic combinations of wheels, springs, hydraulic pistons, steel wires, legs and joints (optimizing for metrics like mobility etc). I even wonder if flexible joints made out of hard rubber could offer any advantages on bipedal motion
I’d go further to argue that it’s very necessary for robots to have even more mobility such as wheels to take advantage of momentum and increase efficiency. There’s a reason the wheel was invented, things like bicycles, wheelbarrows, etc.
The ANYmal wheeled legged robot is an example of that.
https://newatlas.com/robotics/anymal-swiss-mile-quadruped-wheeled-standing-robot/
I’d argue that the wheel was invented not because “walking” was inefficient, but because flesh is weak and gets tired.
A robot doesn’t have that weakness. It thinks nothing of running five hours at high speed if necessary. It has no need of wheels if it can just Gump it cross country with cargo on his back a la Death Stranding.
Robots have battery capacity limitations, they get “tired” in a different way. Your claim is true if you invent a battery that never runs out of power.
But does walking necessarily use more energy than rolling?
Good point well made. I hadn’t considered that.
Yes, very good point!
I wonder if someday in the future we might use reinforcement learning to iterate over different mechanical designs to explore even more exotic combinations of wheels, springs, hydraulic pistons, steel wires, legs and joints (optimizing for metrics like mobility etc). I even wonder if flexible joints made out of hard rubber could offer any advantages on bipedal motion
They might be able to ride bikes at some point.
I think the benefits to making them humanoid are underrated in this comment section.