Engineers from the University of Colorado’s Atlas Institute and the University of Tokyo have designed shape-changing swarming robots that can both individually and collectively reconfigure themselves for any number of applications, including displaying data, actuating objects, and act as tangible controllers.
Known as ShapeBots, the robots feature thin modular linear actuators that enable them to reconfigure themselves into various shapes.
“Swarm user interfaces support interaction through the collective behaviors of many movable robots. By combining such capability with individual shape change, we can enhance the expressiveness, interactions, and affordances of current swarm user interfaces. For example, self-transformable swarm robots can support representations that are not limited to moving points, but also lines, and other shapes on a 2D surface.”
The ShapeBots linear actuator was inspired by a retractable tape measure, which has a small footprint but extended range. The real-based linear actuators are equipped with small DC motors, which are used to feed and retract the thin polyester material that’s connected to an end-cap. By using the motors at different rotations, they can crease or warp the material on a side, allowing for contortion, or shape change. It’s also how the robots can push or grab objects for interaction.
An ESP8266 controls the robots, while a DRV8833 motor driver controls the linear actuators, which are powered by a 3.7V (110mAh) LiPo battery. Tracking and orienting the robots are done by using ArUco fiducial markers positioned on the bottom of their 3D printed housing, while a Logitech C930e camera placed on the underside of a transparent table uses computer vision to monitor the robots. The engineers developed a control algorithm that provides the robots application instructions utilizing a host computer for any given tasks at specific target locations, which enables the robots to swarm while avoiding collisions.
So far, the engineers have used ShapeBots for simple tasks, such as cleaning desks and visualizing state populations, which is an accomplishment considering they are only in the prototype stage. They also hope to reduce their size to create more massive swarms to create higher visualizations.