PDA

View Full Version : [Question(s)] Robotic multi-axial joint



alonso
11-20-2012, 11:32 PM
A question for you anatomy buffs. Is the way that the human shoulder can can get such a large range of motion because the socket part of the ball and socket covers a very minimal surface area, and it's being held in place by the elasticity of the muscles?

I know that typically people break it down into a pivot per axis of actuation, but has anyone heard of or built a robot using a multi-axial actuation in a single joint?
4319
-Alonso

Th232
11-21-2012, 02:01 AM
Yep, that's pretty much how the shoulder works. The "ball and socket" for the shoulder is more like a golf ball and tee than anything else.

I've seen a few cases similar to what you're looking for, look up a Stewart platform. Not identical in that the base and top plate aren't directly connected, but the concept is there. Bear in mind that you'll run into a rather interesting set of problems when you start working with parallel manipulators. This may be reduced or eliminated if you do make a physical connection and are then able to reduce the number of actuators.

alonso
11-21-2012, 12:41 PM
Hey TH232,

I had never seen a Stewart platform, Very cool! Thanks for sharing. That set-up not only gives you multi-axial rotation, but translation as well much like the way hexapods can move.

In an earlier project I did something similar to this Dual-axis joint, but I ran into some problems with the angular limitations of using a U-Joint.
http://youtu.be/9tv6RwcdmHk?t=2m10s

I was thinking of trying to use machined springs (http://heli-cal.com/cm/Products/U-Joints/Home.html) as a type of joint. Anyone have any experience with these? it seems like they might give you a higher range of motion than the traditional U-Joint.

jwatte
11-21-2012, 01:15 PM
I think the human shoulder also has secondary motion from the clavicle and other shoulder bones. I e, the "socket" itself actually moves around when you move your arms.

tician
11-21-2012, 01:40 PM
I'm pretty sure I've linked this thing (http://spectrum.ieee.org/automaton/robotics/industrial-robots/superfast-robotic-camera-mimics-human-eye) before in a previous thread. Uses three linear actuators to create a joint with three rotational degrees of freedom affixed to a ball end for support.

One of those projects I would love to build, but will never get around to doing is to make a 3~4 meter tall pirahna plant with a couple dozen of these joints (pneumatically actuated not piezoelectric). The design I had thought of using was essentially a bunch of identical cylindrical segments with a pneumatically actuated joint on one end (a single air supply tapped by all segments with nitinol wire air valves in each segment to control each of the three pistons in each segment).

alonso
11-21-2012, 05:22 PM
Very cool link, thanks! I had never heard of ultrasonic piezo actuators. So tiny!!!!! I wonder what type of torque the produce?
http://youtu.be/oTrzNwjWCBo

And check out this spherical ultrasonic piezo actuator!
http://youtu.be/_O68scHOwnY

I also got curious about 3 DOF force feedback joysticks and found this website (http://electronics.howstuffworks.com/joystick8.htm).

tician
11-21-2012, 05:54 PM
Very cool link, thanks! I had never heard of ultrasonic piezo actuators. So tiny!!!!! I wonder what type of torque the produce?IIRC, it usually is in the milliNewton range. Very high speed, but low force.