Catherine The Creative

1) It's my understanding the CtC I basically just uses random movements to apply the paint to the paper. Is this right, or is there some other kind of 'decision' process that goes into applying the paint?

It was not entirely random. CtC uses a Lynxmotion 5-axis robotic arm. Using a test program, I located the servo positions (5 servo values) of the paint bucket area and about 10 different positions on the easel. I then wrote a program that would essentially: a) move the arm into the paint bucket b) pick a random number of points between 2-4 c) move the arm to each of the 2-4 points in turn d) go back to step a

2) Does the robot decide when the painting is finished and stop working or do you turn it off or remove the paper to end the work?

The artist I worked with (Anna Nedeau) used the robot as kind of a tool. She would squirt different colors of paints into the paint bucket area and would occasionally reorient the paper. It was fascinating to watch them work together. They got into a kind of rhythm that would go until Anna decided it was finished.

3) How does CtC II work? What does it draw and how is that determined?

CtC II is a much more involved project. It started as an investigation into Inverse Kinematics. Inverse Kinematics involves the mathematics of computing all of the angles for the robot joints to make the end point of the arm arrive at a particular x, y, z coordinate. It turns out to be a lot harder than it sounds. Even after seeking help from Robotics Engineering Textbooks and several friends ranging from a University of Texas mathematician to a PhD MIT physicist, we never got good results. Eventually, I punted and used the technique of Forward Kinematics which is a lot easier. This turns the problem around by taking all of the robot joint angles and computing where in 3-d space the end point ends up at. This allowed me to write a program that would try thousands of random angle combinations for each point on the drawing easel and picking the best choice. The program ran for about 3 days. At the end I had pretty good guesses for a 128 x 128 point wide drawing area.

CtC II loads the precomputed table for the 128 x 128 drawing area into memory. It them loads a simple graphics image that I drew with a drawing program. From there, it tries to arrange the points such that the dots are arranged in smooth continuous lines to make it easier for the pen to make smooth lines. CtC II could draw simple recognizable shapes such as a box, a circle, and an X. The interesting thing that came from this was that because each point was separately approximated, the drawings had a kind of "squiggle-vision" look to them.

4) How did CtC come about?