The ElectroVoxel: MIT Scientists 3D Print Self-Reconfiguring Robotic Cubes

The term “robot” still tends to immediately conjure up imagery of blocky, metal men who speak in mechanical voices; or perhaps, otherwise, Westworld-esque, super sophisticated cyborgs, indistinguishable from naturally-occurring Homo sapiens in every way save for method of creation. The more advanced that robotics becomes as a field, however, the clearer it is (even to admitted robophobes like myself) that robots will make their greatest real-world impact in all forms not directly inspired by the human anatomy.

Indeed, some of the most exciting developments lately in the world of robotics, rather than increasing the structural complexity of the objects in question, seem to be headed in the opposite direction: radical simplicity. However, this doesn’t mean the jobs these robots are being designed to perform are getting less complex. Again, the opposite is the case; and 3D simulation and 3D printing are, in large part, facilitating this dynamic. For instance, last month we covered squishy robots that an Austrian research team made from renewable materials. Without 3D printing, it’s difficult to envision how the scientists behind the project could’ve both created the material facilitating the experiment, and figured out how to operate the robotic system they created, all in about two years.

Along similar lines, a research team at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) recently published a paper, “ElectroVoxel: Electromagnetically Actuated Pivoting for Scalable Modular Self-Reconfigurable Robots”, which it will present at the upcoming (May 23-27) 2022 International Conference on Robotics and Automation (ICRA) in Philadelphia. The ElectroVoxel (it could use a better name, imo) is the team’s version of modular robots, which, as the paper’s subtitle conveys, are small robotic blocks which self-reconfigure to create large, shifting assemblages of robotic structures.

In the paper, the team’s lead author, MIT PhD student Martin Nisser, writes, “When building a large, complex structure, you don’t want to be constrained by the availability and expertise of people assembling it, the size of your transportation vehicle, or the adverse conditions of the assembly site. While these axioms hold true on Earth, they compound severely for building things in space. …By applying this technology to solve real near-term problems in space, we can hopefully incubate the technology for future use on Earth too.”

The ElectroVoxels are small cubes — their side lengths are about 2.4 inches — made from ferrite core magnets, wrapped with copper wire. Inside each cube is a circuit board, the scaffold for which was printed on an Ultimaker 3; the cube’s magnetic sides are attached to each other with connectors printed on a Form 2 from Formlabs. Each cube costs just around seventy dollars to produce, and an hour-and-a-half to assemble. Again, this clearly wouldn’t have been possible without 3D printing.

Moreover, as the team explains in the paper (p. 3), “Manually planning pivoting maneuvers and their associated electromagnet assignments becomes intractable for more than a few cubes. To let users visualize and plan reconfiguration maneuvers, we developed a simulation (Fig. 3) that computes all electromagnet assignments based on desired reconfiguration maneuvers specified by the user.” So, the most unique aspect of this particular project — the creation and utilization of wireless hinges, a fairly mind-blowing concept — is facilitated more or less entirely by 3D imaging software.

As Nisser points out, the ElectroVoxel’s most natural applications, to start, would be in space, not just because of “the favorable dynamics provided by microgravity”, but equally, owing to the simple practical matter of optimizing storage space. Additionally, as with the renewable-based squishy robots, the inexpensiveness of the robotic cubes makes them particularly well-suited for any job that they’d likely be destroyed by. This creates the potential for versatility and accessibility that could never be achieved with large, heavy, capital-intensive machines. The integration between 3D simulation and manufacturing reality is not only giving individual researchers in the robotics field more control over their experiments, but, in doing so, is opening up entirely new understandings of how robots could ultimately function.

Images courtesy of MIT