MIT engineers design self-replicating robots capable of assembling giant structures
When it comes to the manufacturing of commercial aircraft, different parts are manufactured at various locations.
Before finally bringing them all together in a central plant and putting the finished aeroplane together, the tail components, the fuselage, and the wings are made at different plants.
Many other large structures, besides aircraft, are also built in sections. But what if instead of sections, the entire assembly could take place in one go using a team of small robots?
A doctorate student, Amira Abdel-Rahman, and Neil Gershenfeld, professor and director of MIT Center of Bits and Atoms (CBA), are envisioning precisely this in their doctoral thesis work which was recently published in Nature.
Self-building robots formed of superior voxels
The new research provides a step forward towards building robots that can basically make anything. It builds on years of research that employed small, identical lightweight components, including the assembly of a functioning race car and a malleable aeroplane wing.
The MIT team have successfully produced prototype robots capable of assembling small structures which connect to build whole vehicles and buildings- and even bigger robots.
They can transport data and power from one unit to another
Like earlier experiments, the team's approach uses voxels (the volumetric equivalent of a 2-D pixel), which are a collection of tiny identical subunits that create big, useable structures.
However, in this case, the voxels employed are more complex than earlier ones in that each of them can transport data and power from one unit to another. Previously, voxels were only mechanical structural components.
“When we’re building these structures, you have to build in intelligence,” Gershenfeld said in an MIT article.
Additionally, older versions involved bots that were connected through wire bundles and control systems. The new system is more efficient as voxels ensure there is a single structure — no bundles of wires or power sources.
The voxels join end-to-end to constitute the robots, and they can use their attachment points to grab another voxel and move it into its desired position. When the voxel is attached to the structure, it is released in that position.
When the structure size becomes large enough, the robots can make bigger robots in order to be efficient and reduce travel time across distances.
Algorithms handle the robots’ decision-making
One of the main focuses among researchers is to create algorithms that handle the robots’ decision-making as to when to build the structure, when to build more robots, and when to build larger ones.
According to Gershenfeld, an automatic robot assembly system that is capable of assembling massive structures and self-replicating will still take years to set up. However, this new work is a huge step towards achieving that goal.
Ultimately, this technology can be used to make large airplanes, factories, automobiles, etc., in a single go. Robotic workforces can quickly scale the materials and personnel needed to assemble the desired structures making the entire process that much more time and labor efficient.
You can compare this robotic system that assembles big structures like that to a kid making a large castle using LEGO blocks. The work has attracted significant interest from potential users like NASA, who the researchers are now collaborated with. Additionally, Airbus SE, a Europe-based aerospace firm, is sponsoring the study.
A simulated moonwalk in Arizona allowed engineers to test a wearable for future Artemis astronauts.