Experience Design & Physical Computing, 2019
Draw Words Drawing Robot
An interactive created to give the user the ability to visualize their favorite text. Perhaps a poem, email, song, or text msg.
How it Works:
Final Result:
Design Process:
Ideation: We were initially inspired to build a meeting visualizing robot, drawing different markings for speaking frequency or content contribution. After some research, we found many of these robots existed. We pivoted seeking to create a more novel robot that was both artpiece and text analysis tool.
Prototyping: Our first prototype was on wheels and drove across a paper surface. We decided the driving was unnecessary and clunky. This lead us to a mounted robotic arm solution which offered us more control over the final drawing. We prototyped the arm using two servo motors, cardboard, a marker, and an Arduino.
Software Design: To make the robot translate emotions found in text, we use two programs.
Sentiment Analysis Program: Evaluates the emotions expressed in the text
Visualizer Program: Creates a virtual image of the line drawing the robot will create
For the Sentiment Analysis Program we use VADER. This tool gave positive, neutral, or negative values for each line of a piece of text. We then developed a visualizer program. The program manipulates line segments that we drafted in photoshop. These line segments had plotted points along them that later translated into polar coordinates. Each line segment had a set of 10 coordinates, which were then modified corresponding to the values produced by the sentiment analysis program.
Digital image from the visualizer program
The visualizer program was able to draft a mockup of all the line segments for a drawing. With these images we were able to prototype digitally what the final drawings would look like.
Our final design featured six different line segments (see below). The size each segment was drawn at correlated with the value from the sentiment analysis. For example: a positive value of 10 would create a larger loop than a positive value of 2.
Final Line Segments
Labeled with the sentiments they represent
Hardware and Electronics: We used a Raspberry Pi to run our Python script. The Raspberry Pi takes a picture, runs text recognition software on it, then (using formulas we derived) it generates the angles to turn the servos to. We were unable to get smooth servo motion on the Pi, so we send the angles over a serial line to an Arduino, which drives the servos and creates the drawing. When the phone displaying the desired text is in place, a simple button is used to activate this process.
Physical Design: We chose to build the majority of the device out of black spray-painted MDF to give the project a clean, modern feel. The pieces were designed in CAD, laser-cut, and glued together. For easy setup, we designed a stand to hold the user’s phone. This stand sits parallel to the Raspberry Pi camera & mount.
CAD Drawing of Base
Base Layout
Signage Design: We created labels and instructions to enable the user to easily participate in the experience. This signage also clearly communicates to the user the meaning of each shape being drawn by the machine.
Additional Images:
Takeaways & Conclusions:
Our final design was successful in that it…
was an effective elegant art piece that composed a line drawing from text
generated an image unique to each text based on sentiments from VADER calculations.
is clean, understandable, and branded - accomplished by specific material choices and signage
This design project allowed me to..
experiment further with Arduino and Raspberry Pi hardware
practice working with graphic design and UX communication
With more time and resources, I would like to improve the marker load an unload mechanism. As a whole, the project offered a unique opportunity for me to pair my hardware and fabrication skills with my visual design skills.
(Created in collaboration with Eliot Laidlaw and Talie Massachi)