Low cost Robot Hand (with only one motor)

Hello,

Here there's a video of a robotic hand I build! :slight_smile:

I know (at least in my experience) that the end effector can be a very complex part to make: because I'm very satisfied of this result I've obtained using only one motor, I'd like to share it with you, hoping to be useful to someone looking for ideas!

Let me know,
A.

That's awesome, great job!!!!

That's very good... thanks for sharing that.

I was going to do this but couldn't find nice thick straws and in anycase they're very flimsy. The garden hose is a great idea...

the music sucks

Your assertion that you need N motors for N movements is wrong.

A small three-wheeled robot with 2 independent motors has 4 basis movements,
forwards, spin right, spin left, or backwards.

It is also quite clear in your video, that the forearm is being manipulated off-camera
by a human, and not by any kind of robot arm mechanism.

Your assertion that 90% of hand movements involve grasping objects or picking up objects
is also very dubious.

I type ( as in keyboard ), 10's of thousands of movements a day and I pick up maybe a dozen objects a day.

michinyon:
the music sucks

Well, I'm not sure this is a constructive feedback :slight_smile:

michinyon:
Your assertion that 90% of hand movements involve grasping objects or picking up objects
is also very dubious.

It's known that the major part of the use of our hand is just grasping. You can grasp different objects in different ways but is one movement.
90% maybe is exaggerated, but I was trying to summarize in only one frame the whole preface.

michinyon:
Your assertion that you need N motors for N movements is wrong.
A small three-wheeled robot with 2 independent motors has 4 basis movements,
forwards, spin right, spin left, or backwards.

You're talking about another context. I don't have weels, because I tried to build a hand, not a car.
I just said taht moving N fingers independently requires N motors.

michinyon:
It is also quite clear in your video, that the forearm is being manipulated off-camera
by a human, and not by any kind of robot arm mechanism.

Well sorry but here I understand that you didn't realize what the purpose of the video was.
I just wanted to realize the end effector. All the tests, are tests on the functionality on the end effector.
I never said I created a robot capable of playing chess, did I?

What size of garden hose did you use in the low-cost robot hand?

jeffmorris:
What size of garden hose did you use in the low-cost robot hand?

The diameter of the hose is 1.5 cm.

The thickness is approximately 3mm for thumb, index, middle, and 2mm for ring and pinkie (if you see the video if can recognize two different kind of hoses). This is because a thicker hose is harder to pull, but is more effective to create the opposable thumb mechanism.

A.

jeffmorris:
What size of garden hose did you use in the low-cost robot hand?

I was wondering that too... looks about 10-12 mm to me, 3/8"-1/2"?

Edit... ah, I see. The different wall thickness is good thinking!

I just bought some of that hose... it's quite curled from being rolled up in the shop. I'm wondering how to straighten it? Or once I cut those notches at the knuckles maybe it won't matter....

JimboZA:
I just bought some of that hose... it's quite curled from being rolled up in the shop. I'm wondering how to straighten it? Or once I cut those notches at the knuckles maybe it won't matter....

I'm glad you liked the idea :slight_smile:

Well, when I bought a long piece in the shop it was too quite curled, but when I cut the five fingers (in my case 12cm each, more or less) the shorter hose pieces were much more straight!

A.

I'm also wondering how many bend / straighten cycles the tube can handle before it loses its springiness to loosen its grasp 8)

But right now that doesn't matter...

Well, so far I can guarantee it at least for some hundreds of cycles :smiley:

Nice project, I'll probably try to do something similar in the future :slight_smile:

It's known that the major part of the use of our hand is just grasping.

Known by who ? You just claiming it, does not make it true.

As I said, I type thousands of characters a day, and I pick up maybe a dozen things.

I quite like your hand, it's good ! It is just the dodgy claims in your captions that I don't like.

I just wanted to realize the end effector. All the tests, are tests on the functionality on the end effector.
I never said I created a robot capable of playing chess, did I?

No, but you did claim that your hand uses only one motor/servo. To have a usable hand for picking anything up,
you would need some kind of arm which will require more than one motor or servo.

michinyon:

It's known that the major part of the use of our hand is just grasping.

Known by who ? You just claiming it, does not make it true.
As I said, I type thousands of characters a day, and I pick up maybe a dozen things.
I quite like your hand, it's good ! It is just the dodgy claims in your captions that I don't like.

My source is a TED Lecture i saw about the "Intelligence of hand", by Antonio Bicchi, who is a Professor of Robotics at University of Pisa. You can find it here:

but unfortunately I found it only in Italian. He explains that our human hand is incredibly complex, and has more than 20 degrees of freedom. Nevertheless, we can have a quite useful robot hand with only ONE degree of freedom, and in his lecture he explains the possible important implications of this discovery.
Above all, he says this can be useful to make simple but light and cheap prosthetic hand.

My youtube video is not a TED talk, and I'm not a professor of Robotics and I've not 20 minutes to explain my ideas. I just tried to summarize all the important facts in just one frame, and English is not even my mother language. If you find my effort a "dodgy claim", I don't know what to say :roll_eyes:

michinyon:
you did claim that your hand uses only one motor/servo. To have a usable hand for picking anything up,
you would need some kind of arm which will require more than one motor or servo.

Sorry but I think you confuse the end effector with the rest of the robot (whatever it is). The end effector I realized is complete, and requires only one motor to do his job (grasp something). You can just google "end effector"...
Then, if you want to realize a whole humanoid robot, probably you will need other 40 motors, but this is another issue. :slight_smile:

No, but you did claim that your hand uses only one motor/servo. To have a usable hand for picking anything up,
you would need some kind of arm which will require more than one motor or servo.

Spin control! I don't think any claims to arms were made. The discussion title is "Low cost Robot Hand (with only one motor)". No claims to hands that require "arms".

Attached is a pic of my proof-of-concept finger based on ndrplz's video.

I was surprised at how much force it takes to curl the finger: servo sounded like it was straining and was drawing about half an amp.

But all good fun, and I thought, what the heck, might as well mount in in a surplus box to hold the servo steady while it hauled on the string. And lucky I had those two old electric conduit clamps in the cellar.

Note to self: get new blade in junior hacksaw, that fibrous hose is a bugger to cut otherwise.