Buildings often contain more material than their structures require, but making and installing customized components can be costly. MIT researcher Miana Smith is exploring a different approach: small lattice blocks designed to fit together accurately even when the robots placing them are comparatively imprecise.

Smith, a recent MIT graduate and visiting researcher with the Center for Bits and Atoms team in the Program in Media Arts and Sciences, has focused on automating the production and assembly of the blocks, which her lab calls voxels. The work aims to make it possible to build larger, load-bearing structures from repeatable units.

The construction industry is often cited as responsible for more than one-third of global greenhouse gas emissions, according to MIT News. In steel-framed buildings, an estimated 30–40 percent of the steel could be removed without weakening the structure by tailoring beam cross sections. But custom parts can require expensive manufacturing and skilled installation, making oversized, uniform components cheaper to use at scale.

Putting precision into the blocks

Automated fabrication can accommodate complex designs, but conventional approaches face their own precision demands. MIT News reports that placing framing members or extruding concrete across a 10-meter span can require accuracy within a few millimeters. Machines capable of that precision are difficult and expensive to manufacture.

Smith’s approach shifts some of that burden from the machine to the material. The blocks snap together in a way that helps prevent errors from accumulating as a structure grows. She compares the principle to Lego bricks, whose mechanical design allows even an imprecise assembler to produce a precise result.

The robots used in her work have a margin of error of about 10 millimeters, compared with roughly 100-micron tolerances for most industrial robots. Because they do not need industrial-grade precision, the robots cost a few hundred dollars each rather than tens of thousands, according to MIT News. That lower cost allows multiple machines to work in parallel.

Getting the system to build at a larger scale required changes to the blocks. Earlier units needed as many as 24 connectors apiece, with each joint creating another potential point of weakness or assembly failure. Smith’s solution was to make the units larger and pre-attach as many internal joints as possible before robotic assembly. The team also adopted a more internally connected lattice to improve its strength or stiffness relative to weight.

Housing demonstrations and concrete plans

The team is collaborating with Druk Holding and Investments on Gelephu Mindfulness City, a housing development project in Bhutan. Bhutan officials wanted construction methods using local materials that could be reconfigured over time, an objective Smith sees as compatible with the voxel system.

Using plywood blocks, the researchers have built room-scale, load-bearing demonstrations, including a small house. They plan to test the system outdoors at a similar scale; MIT News does not give a date for that test.

Smith also wants to apply the reconfigurable-block approach to formwork for concrete. Complex concrete molds are expensive to make, and she expects the flexible system to reduce that cost significantly. Alongside her MIT work, she is pursuing a startup focused on more sustainable concrete construction through an Activate Fellowship.

Her route to the project included undergraduate research with architecture professor Larry Sass and a class taught by Neil Gershenfeld, her PhD adviser and director of the Center for Bits and Atoms. Before concentrating on robotics, she spent two years working with living systems, including bacteria in woven fibers and honeybees sent into microgravity. Smith describes living systems as distributed assembly systems, an idea she says helped inspire her move toward robots.

Further research is planned at the lab: Smith says a new group of students is exploring a wider range of robotic behaviors, including different robot shapes working together on larger structures.