MEERKATs was my senior thesis at Columbia College Chicago, a project to build an ensemble of musical robots. It stands for Musical Ensemble of Electronic Kinetic Art Things. It is very much inspired by the LEMUR bots of Eric Singer and his crew.

Final Update

Here is a performance of the MEERKATs show piece, FibboNachos in the Project Room at Columbia College Chicago. The piece uses Fibonacci numbers in an number of ways to inform the rhythms. The piece ran on a constant loop for about 8 hours. I chose to use Fibonacci numbers because the piece I had been writing in anticipation of the installation had to be rewritten once the robots were completed, because they performed so differently than I had anticipated.

MEERKATs Performance from IAMColumbia on Vimeo.

Update 3

Below is a video showing the process of constructing the robots, which involved a laser cutting machine, and lots and lots of hot glue. This is a time lapse of a roughly 2.5 hour process. Unfortunately, this robot repeatedly fell apart during the final installation phase and could not be included in the final exhibition.

Update 2

This is the completed Diddly Bot, the centerpiece of the Meerkats. I knew that I wanted to build an instrument that wouldn't sound like anything else on earth, and I also knew that I didn't feel like building a resonating body with a hard surface, nor did I want the instrument to be purely electric, so I settled on a banjo like body with a hammer dulcimer mechanism to make the string sound. The plexiglass body ends up giving the instrument a dark tone with long sustain. The steady sliding became a the signature sound for the Diddly Bod. Here is the code on the Arduino. I use Loopbe1 as my virtual midi port, and Spikenzie Labs serial to midi converter.

Update 1

This rig was built to test and demonstrate the interaction of arduino with MIDI signals. The music you hear was generated by soft synths, but the midi signals are also routed to a virtual midi port, which talks to a midi to serial converter, which then sends the data to the arduino, and then some conditionals turn the various components on and off. This test became the basis of the software side of the installation.

Original Post

This is the first autonomous instrument I have ever built. This was really just a test of the mechanism, and a fail safe to prove that the arduino could send musical signals on it's own. . 

The piece the robot plays is just a test piece to demonstrate the speeds the mechanism is capable of. The drum will eventually become the body of a banjo-like instrument I'm calling the "diddly-bot". The rhythm is hard coded into the Arduino. I wrote a Lisp program to help me prepare rhythms more quickly, so that instead of typing out all of the

               digitalWrite(7, HIGH);

               delay(218);

etc. Instead, I can just type:

              (tempo 120)

              (beatval 4)

              (SoidArduinoRhythm `(4 4 4 ,(dotted 8) 16))

And then the lisp program calculates the durations based on the tempo and time signature I gave it, and then formats the Arduino code for me. 

I'm still building the banjo-bot. I'm preparing finalizing the software for these things. Here's my test of Arduino with MIDI. Using MIDI signals means I won't have to hard code the songs, which will shorten the testing cycle and ultimately make the instruments more versitile and easier to write for.