Image Image Image Image Image Image Image Image Image Image

Bartlett School of Architecture, UCL

Scroll to top


No Comments

Chaos Theory for Interaction

  • On March 7, 2018

The optical illusion by Feliks Konczakowski that has been making the rounds on social media lately.

Speculating Chaos Theory

The initial idea of the “sand plotter” came out of my fascination with magnets, and I thought a steel ball drawing patterns in sand would make an impactful visual effect. It is also what informed the design decision later to add the lightbox, which accentuated said effect. I have made references to a variety of drawing machines because that was the only type of interface that offered the precision and control I was aiming for. However, as the project developed, I have found that the most interesting results happened when things fall “out of control”.

I have always wanted to explore chaos theory and the recent exposure to “Edge of Chaos” by the IALab has made me reflect on the theory a lot more. This has inspired a more concentrated design direction that I’d like to explore in the next two weeks, and eventually build a final prototype for. Since the XY plotter now is only operating in two dimensions, I thought I could build the Z axis up and gradually introduce the element of chaos. I am not exactly sure how it could be achieved yet, but it seems that the unpredictability of material can serve this goal quite well.

This would be an interesting concept to further investigate. For a long time, robotic design has been about precision and control; but with chaos theory, we realize that things will never be precise because they’re not functioning in linear systems. We can magnify this theory through a series of chain reactions and give it a visual presentation. For instance, I could employ the most precise machine with a simple task, like drawing a line, and yet each time the machine repeats, influenced by unforeseen “chaotic” factors, the outcome becomes completely altered.

I tested out some of the digital ways of adding chaos, calling functions such as “random” or “noise” along with the gcode locations. Most of the results look a bit like what’s shown in the video below. I imagine if I added a puppet-like figure to be controlled, the overall effect might become even more interesting.



Interaction Experiments

Before this week, I have always relied on a software called “Universal Gcode Sender” to “grbl-communicate” with the XY plotter. While such method was extremely effective, it takes away the possibility of real-time responses, which is easily the most exciting part for interactive art.

Through scraping online and getting help from tutors, this week, I have finally found success incorporating Processing with my previous design. The following videos are to show a few compelling methods I have managed to achieve for controlling the plotter. Going from here, it is not difficult to imagine a variety of interactions that could be attained. I will have to narrow down and pick the best one in the week to come.



Quick, Draw! with Google

In the process of designing for interaction, I was inspired to look into the Quick Draw Dataset produced by an online experiment with Google Creative Lab. Following a recent tutorial by Daniel Shiffman and making a few adjustments, I created the below sketch where all drawings of “apple” are randomly displayed one per second.

Originally I thought it would be nice if I could have the XY plotter draw these doodles in sand, and have another mechanism display the location and time stamp of the original creation. I soon realised it was not possible using the bitmap files I have downloaded, because they are raster images, whereas an XY plotter uses vector. I am not sure at this point if it would be a worthy exploration to alter ndjson files, which is one of the only other formats of data Google provides from this experiment.


Submit a Comment