Friday (Saturday) Night Robotics – MANOIlangelo

Posted by Erin, the RobotGrrl on Saturday, May 30th, 2009

If the title wasn’t a big enough hint (MANOIlangelo), my next goal is to program MANOI to write the alphabet and numbers!

In case this somehow turns into a big project of mine, or the focus of my summer research, I’m going to write a very detailed version of my procedure (just to be safe). It will sound pretty un-blogtastic, but there will be pictures to look at, though!

After I finished creating all of the movements for MANOI’s arms possible (waving, hand shaking, worm/slide), I was wondering what else I could do with MANOI. After thinking about all the motions we do with our hands, I finally came across writing!

I brainstormed about the set-up a bit, and I decided to go with a simple set-up, instead of trying to make an eisel/board. The paper is about waste height, it’s held there using books.

Next part to figure out was the pencil. I played around with it a bit to see where it would be the best way to place it. The best way to place it is along the servo bracket so it doesn’t limit any of the motions.

Now I had to get the pencil to stick there! I actually didn’t use hot glue (not a good idea to use hot glue in room at night). Instead, I used a combination of tape and hair elastics. I would have use tie wraps instead of the elastics, but I didn’t have any.

After it was all attached and all, it was time to start trial and error’ing!

In order to see the << insert smart word here >> of the servos, I created two simple frames to move servo #2 back and forth. Through visual observation, I tuned it so that the pencil never left the paper. The result is a line! Due to chaos though, if the motions is repeated, it is actually an ellipse. This is because everything from the position of the paper (may be moving if it’s not level), the attachment of the pencil, the length of the pencil lead, the movement of the robot, and the possible-yet-minimal errors in the servos.

I repeated the previous step, but for servo #3.

I then combined the motions of servo #2 and servo #3 to make a rectangle!

Results:

1 iteration of the rectangle

Many iterations, started 2 times. Moved the paper between trials.

Several iterations, strong demonstration of chaos, most likely due to the length of the lead

A few iterations, little sign of chaos

MANOI drawing

It is pretty cool! Looking at the resulting rectangles though, it is obvious that there is a section where the pencil does not touch the paper. This is known as a singularity.

If one was to take the inside of a hemisphere, and draw two types of lines on it, you would get this:

At the bottom of the hemisphere, there is more concentration. However, as it gets closer and closer to the semi-circle line or equator, it becomes sparse.

One could apply this idea to the rectangle, where it is 1/4 of a sphere. Note, by sphere I actually meant hemisphere. EDIT: Scratch that, I did mean sphere. I’m having problems with fractions, evidently! LOL!

Why does this make sense? If you think about it, in 3D space the end-effector (pencil) will not touch the paper exactly the same anywhere. There will be more pressure in some spots than others. However, since we want MANOI to draw away from his body, and infront of itself, this leaves only 1/4 of a sphere.

The next step is to create a simple test program that will draw dots all over this area of the rectangle. This way we can gain a better understanding of how to make a robot draw letters.

The interesting part about this is that, in this scenario, it would actually be easier to create AI for the robot to learn how to write the letters by simple feedback, compared to programming the letters in. However, I believe that I should program them in first, and then possibly make AI for it. This way, there can be a goal node in the AI if it was to be made.

Posted in: MANOI, Projects, Robot.