Typhoon Storm 2142, on 30 July 2013 - 05:41 AM, said:
I would like someone explain to me how a 'Mech is able to turn on the spot, but not be able to make a sidestep. The same mechanics are used in both movements. EVEN THOUGH they supposedly have no ball-joints or such.
It could be done via a lot of rotation about the ankle joints - though, that places a lot of stress on a relatively small joint and would be generally slower and less efficient than using properly-actuated hips.
For example, if a
Mad Cat (which weighs 75 metric tons, or ~165,347 lbs) with no lateral or vertical hip actuation (reference the image linked in
my previous post) wanted to turn in place to the right:
- lift the right foot, keeping the left foot planted
- rotate clockwise about the left ankle
- place the right foot on the ground, then left the left foot
- rotate clockwise about the right ankle
- place the left foot on the ground
- repeat steps 1-5, swinging the lifted leg forward or back as appropriate to keep the feet laterally aligned
- repeat step 6 as necessary to turn through the desired arc
Note, however, that being limited to such a turning motion would mean that the
Mad Cat (in addition to placing
a lot of strain on an ankle joint where the balancing plate is ~2 meters wide and the ankle shaft is maybe one-third of that) would not be terribly maneuverable - it would not be able to turn very far per step before losing its balance, which means lots of very tiny rotations and a corresponding number of steps, which translates into a lot of movement for very little, and ultimately very slow, progress.
Volthorne, on 30 July 2013 - 08:55 AM, said:
Believe me, I had considered that particular instance. However, the topic is specifically about side stepping, which implicitly requires the use of ball-and-socket joints on hips, and largely excludes feet (if you can muster enough momentum to throw 20+ tons of metal sideways a few meters, not having a large tilt on your feet isn't going to matter much). Therefore, that was largely a moot point when I was attempting to convey my logic (which also doesn't account for compression of the ground beneath you, provided there it can be compressed. A hill made of dirt or sand or ash would be much more accommodating to a 'Mech walking laterally across it than say, one made from granite, thus reducing the stress on ankle joints).
Actually, side-stepping does not "implicitly require the use of ball-and-socket joints at the hips".
See the video of Aldebaran Robotics' NAO linked in
my (other) previous post; it accomplishes a similar degree of maneuverability without apparent use of such a ball joint.
Here are a couple of close-up images of the NAO, allowing for a rather close look at the hip assembly:
The lower (gray) part of the visible joint rotates about the lateral axis (allows for the forward and back motion needed for walking/kicking/etc) as well as about the longitudinal axis (moves side-to-side; allows side-stepping) within the upper (white) part of the joint, which then connects with a second joint inside the pelvis that rotates about the vertical axis (swings the hips forward and back, as seen in the standing sequence in the video).
If anything, it's more like a
universal joint.
Ball joints are not the only means of achieving the required flexibility, or of doing so with an acceptable (with regard to maintaining a humanoid form) degree of compactness.