Dissertation Title: Scalable Robotic Voxel Construction

Abstract:

This work advances distributed robotic assembly through the development of scalable, structural robotic material assembly systems based around modular, lattice-based building blocks, or voxels. These voxels combine structural, electrical, and robotic functionality to enable the automated assembly of large-scale, load-bearing structures. Current methods for the digital fabrication of large-scale structures typically use complex, costly, and inflexible machine or robotic systems, limiting the design freedom, material efficiency, and mechanical performance of the resulting structures. This is because accurately and reliably producing large structures using analog processes inherently competes precision, speed, adaptability, and build size against each other. In these kinds of systems, such as large-scale 3-D printing, small errors can accumulate into significant failures without non-destructive repair methods. Over the lifetime of a structure, modification and repair often present significant challenges with limited end-of-life options.

Distributed robotic systems working with digital material systems offer an alternative approach, where multiple small robots cooperatively construct large structures by traversing and building directly on the workpiece using an error-correcting material system. When paired with material systems made from renewable materials in structurally efficient geometries, this type of assembly system has the potential to enable the assembly of sustainable, reconfigurable structures that can be adapted for new uses over their lifetime. However, prior systems have struggled to achieve sufficient load-bearing capacity, assembly throughput, and geometric flexibility to scale to real-world building scenarios. Similarly, while earlier voxel-based material systems have demonstrated diverse and designable mechanical performance, their efficient assembly into full-scale, load-bearing structures remains underexplored and a critical barrier to their use in practical applications.

This thesis develops new methods for the scalable robotic assembly of load-bearing, electronic, and self-replicating structures. First, I introduce new hardware for electronics-integrated voxel systems, which serve as the foundation for modular robotic assembly and the direct-write fabrication of electromechanical structures. These systems are then used to demonstrate robotic self-assembly toward recursive, self-scaling swarms, as well as alternative robot geometries for application adaptability. Next, to improve scalability and efficiency, I introduce new hierarchical interlocking voxel architectures, supporting faster, more cost-effective robotic assembly at large scales. This system is evaluated in the context of architectural applications, where this approach offers a sustainable and fully reconfigurable route to structural on-site building construction. This is demonstrated in a comparative study evaluating the proposed robotic voxel construction system against other existing automated and manual approaches to structural construction methods. The material-robot system is then extended to the architectural scale and evaluated through the construction of a set of load-bearing functional structures: a prototype house, a demo car, and reconfigurable furniture, all made from reusable parts and renewable materials. Together, these results demonstrate a path toward self-scaling, distributed robotic construction systems capable of building functional architectural structures with reduced material use, faster assembly, and minimal infrastructure.

Committee members:

Neil Gershenfeld
Director, Center for Bits and Atoms
Center for Bits and Atoms

Caitlin Mueller
Associate Professor
MIT Architecture and Civil and Environmental Engineering

Maria Yablonina
Assistant Professor
University of Toronto - John H. Daniels Faculty of Architecture, Landscape, and Design