Sunbot Swarm Test 3


The third test of the Sunbot Swarm occurred this past weekend, March 9th, 2019. A few important changes have occurred in the design:

  1. Switched to an acrylic chassis for greater stability
  2. Added plastic domes to enclose the bots. With that, I had to consider how to attach the domes in a way that made them stable, but also easy to open and modify. I used long threaded rods and nuts. Working on a new tutorial to reflect these changes, with pictures.

I like the domes a lot. Visually, they begin to take the bots into a more playful place. They reference spacesuits and cyborgs. Practically, they allow the robots to be outside in cooler weather, and in theory, to survive in a greater range of ecosystems.

Things that are working well:

They are much more stable and exhibited some interesting behavior in terms of sometimes flocking together (perhaps amplified by infrared communication, the bots start emitting if they have enough light), and sometimes widely dispersing. They became much more active later in the afternoon. One of the ivy bots was satisfied with lower light levels, I need to check the sensors on that one, but it is a bit of an anomaly in the video, always hanging out in shade, though perhaps appropriate to an ivy. The rest performed pretty well. They generated a massive response from bystanders, lots of good conversations were had.


-A man asked if they were controlled by turtles. And then a child suggested they were robotic. We had a good chat altogether based on this.

-A man told me about his most beloved plant, the “people” plant, which attracts visitors to his home, sometimes too many (in-laws). Lots of discussion of our love for plants overall.

-Many people asked whether the plant itself was “controlling” the movement or acting as a sensor, which is a great question.

Things that need more iteration:


One thing I was uncertain of, was how much air-flow the plants would need. A few of the plants did steam up which seemed to indicate they needed more air holes. I wedged some sticks into the seam of the plastic dome to allow more air flow for those. For now, I am just leaving them in the enclosures for tests and documentation, and I take them out afterward (inside). I need to research more into terrarium/greenhouse design, to determine the right amount of air holes, that won’t let too much cool air in.

The other good thing about air holes is that it would allow rain water to get in, but I do have the issue of needing to drain the water to the sides, to avoid draining directly onto the electronics. That issue was better resolved in my original design for the Sunbots, the structure of which was meant to be vacuum-formed, and was prompted by an invited proposal for a installation in a public square in New York:

But this design became very expensive, when considering making multiples. So the DIY explorations reflected throughout this site have been taking a more piece-meal approach––solving one problem at a time. But the benefit of making and testing them this way, is that they are low-cost and easy to replicate, making it a more potentially open project, and also allows things to evolve in a more a natural way, such as with the addition of the dome.

Solar Panels

I designed the lower half of the Sunbot with an overhang, in part to make space for a solar panel to sit there. But I need to modify the dimensions somewhat to get a better fit for a bigger panel.

Sustainable Material Exploration

I’d love to incorporate materials with an even lower environmental footprint into the design, such as mycelium and bio plastic. Explorations ongoing.

Plants as Sensors

A great comment from bystanders and something I’ve been meaning to explore in any case, reading the electrical impulses in plants and trying to determine whether those can be used to map movement, in lieu of or in combination with light sensors.




Sunbot Swarm Test 2


Laser cut bodies made for more stability, but also a but more of a boring performance than the crazy garbage bots (version 1).


-In good light, they are pretty still. Some minor self-adjustments.
-Some anomalies, if one sensor loses light the bots begin to turn, turning may cause the other sensor to lose light as well and send it driving in wrong direction all together.
-Cardboard chassis are not quite stable enough, wheels on some are bowing. Need to replace with acrylic
-Unclear how or if IR communication is effecting behavior.
-Need to test in a more complicated lighting condition.

Sunbot Swarm Test 1


Sunbot swarm uses a slightly more complex circuit than the first Sunbot. A motor driver IC (L293D) as opposed to a transistor circuit allows the wheels of the bot to move backwards, which allows for greater navigational abilities and turning. An ultrasonic sensor is designed to help it avoid obstacles, though as the video shows, when the bots are different sizes and the sensors are in different places, they don’t “see” each other well. The bots also have IR emitters and receivers, in order to help them communicate if they’ve found an adequate amount of sun. However, in the conditions of this test, where the sunny area was limited, they in a sense competed for that small area. Next test will be to see if standardized bodies and sensor heights + changes to the code, can help prevent them from mauling each other.

The robots in this video were assembled by participants in a workshop at Open Source Gallery in Brooklyn, using recycled materials for the bodies.

Some photos from the workshop: