5 ms·
This solution seems to have a lot of “gimbal lock” positions so I’m not sure 3 degrees is really accurate. The gears are so large that you lose 120° of rotation
by hinkley 4y ago
This solution seems to have a lot of “gimbal lock” positions so I’m not sure 3 degrees is really accurate. The gears are so large that you lose 120° of rotation along one axis. 2.6 DoF is probably more accurate.
- xpe 4y agoWhat are some ways the robotics field / industry quantifies degrees of freedom?
- hinkley 4y agoThe six degrees I’m aware of are translation through three dimensions and rotation along each axis. Assembling a car takes a lot of those for some fiddly bits for the interior and under the car. Some configurations of this device can rotate 180° on one and about 240° on another. One configuration, the one that gets the most screen time, is worse than that, which is what got me wondering in the first place.
- ajb 4y agoNot sure - the controllers have 4 degrees of freedom, so maybe they managed to arrange only to lose at most one at any given time, so that they always have three?
- hinkley 4y agoI’ve been noodling on this while doing shores today. I think the main thing is not degrees of freedom here, but degrees of freedom around a fixed point. There are four servos in this thing, and about nine gears. You can rotate a probe through a full three degrees of freedom with three servos and six gears, but you also need the X Y and Z coordinates if you want to spin an object without moving it. I think one more motor would give you full 3 degrees of freedom by allowing you to move the braces out of the path of movement. But I also suspect that one more of these motor assemblies would get you close to five degrees, and four of them total would get you six, with some wiring troubles to deal with. In the end you’d have an arm with just a “shoulder” and a wrist, which could be pretty interesting. I aLso wonder if you could invert the gear and the cog, with motors inside the sphere.