Introduction: Robotic and Human Collaborative Art

By

This project was created as a course assignment at the California State University, Long Beach; taught by Dr. Behnaz Farahi: DESN 490: Generative AI and Robotics.

There is so much untouched possibility in the world of human / robotic collaboration, and one of the many ways this can be applied is through art—For this project, we were prompted to collaborate with robotic design to see where this creativity could take us.

Supplies

Software / Technology:

  • Rhino
  • Grasshopper
  • Photoshop
  • MidJourney
  • RoboDK

Physical Materials:

  • UR10e Robot
  • Black Scratcher Boards
  • White Canvas (18x24)
  • Heavy White Paper
  • Tombow Markers
  • Vinyl Sheets (1'x5')
  • Black Acrylic Paint
  • Laundry Detergent
  • Dish Soap
  • Knitting Needles
  • Rubber Bands
  • 9mm Metal Spring

Step 1: Project Ideations

One of the main focuses for our design process were topographic forms. We had a lot of inspiration—Oceanic forms, flowers, animals, nature, shells, oil spills, loose abstractions of organic forms, and much more. Using our own research as well as MidJourney AI, we compiled many images that we would transform in Rhino and Grasshopper to turn those images into topographic versions.

Step 2: Model Testing in Rhino / Grasshopper

With our images in mind, I moved over to Rhino and Grasshopper to make topographic versions of our reference. For grasshopper, we imputed out images and coded them so that when they show up in Rhino, their structure and color values would be translated into topographic lines. We then used those lines as paths for the robot to travel along to create our physical image. Then once we were done with adjusting our images, we exported them to the robot.

Step 3: Model Testing on Robot

For our final design, we wanted to explore the concept of scratcher art, but at a much larger and impressive scale. Our execution of this concept split divided into two experimental phases, robotic and material. For the robotic side of things, our biggest issue was getting the height of our tool correct. If it was too low and it could break the robot or stab through our table, and if it was too high it wouldn't be low enough to scratch the surface to create our project.


In addition to height, we also went through a lot of tool prototyping and modifications to get the pressure, bounce, give, height, and structural integrity correct. We also conducted these experiments with a variety of failed and successful models to aid in our exploration. Then once we managed to perfect every aspect of this process, (tool, height, and movement), we were ready to start building our final design.

Step 4: Material Experiments

For the material side of things, creating a "scratchable" surface that could reveal another surface below was an extensive process. Aside from coming up with the scratcher part, we also needed to find a material that wouldn't appear damaged when the top layer was scratched away.


For the underlying surface, we found that vinyl worked the best. While the scratchable material could have been applied to any smooth plastic/like surface, we felt that vinyl gave us a lot of options in terms of color and size of our canvas. As for the scratchable material, we spent a lot of time figuring out what materials would work best together to get a clean result. Eventually we discovered that liquid laundry detergent and acrylic paint gave the best results for line smoothness and "scratchability". The only downside to it was that with detergent doesn't really "dry", it just stays sort of wet and can be smeared easily even when mixed with the paint base. Then to combat that issue, we used an iridescent acrylic sealant to help keep everything in place. Lastly, with our material set (the scratcher layer and the vinyl layer), we were able to move into the final designs.

Step 5: Final Design Process

For our final design, we wanted to create something abstract through using natural textures; with that in mind, our research consisted of exploring what kind of textures gave us the most interesting results. Using MidJourney, we found that the best generations came from more ocean themed designs, which helped us settle on our final design—shell textures. Once we found one that worked, we followed the same process from before (MidJourney > Rhino/Grasshopper > RoboDK) to create our final work.

Step 6: Final Design Completed / Collection of Work