Stateful self-assembly

Abstract

Nature shows us many organised structures that form through interactions between their components with little external guidance. These self-assembling systems range from simple crystals to considerably more complex biological structures and organisms. Inspired by these systems, the development of programmable self assembling systems could lead to mass manufacturing processes that produce individually unique items. Current artificial self-assembling systems involve small numbers of centimetre-scale components, and have not resulted in structures anywhere near the complexity seen in natural systems. This thesis argues that to advance artificial self-assembling systems towards this complexity, the statistics of the interactions within self-assembling systems need to be empirically examined and understood. However, the pursuit of this involves the resolution of a variety of technical challenges. These are approached in this work through the development of a self-assembly toolkit that allows the collection of these statistics from a physical system with larger numbers of components than in previous works. A novel capacitive communication interface is developed for the components of this toolkit, which allows messaging between neighbouring components that are constrained to the surface of a plane. As self-assembling components reduce in size towards the microscale, the penalty for incorrect activation of a component’s binding mechanism is likely to increase. With this in mind, this capacitive communication interface is optimised to provide spatial alignment sensing, with the aim of allowing informed binding mechanism activation. The toolkit developed in this work uses components that are constrained to two degrees of freedom of motion. In pursuit of the development of programmable self-assembling components for 3D structures, a new design of alignment sensor for use in 3D is created. Simulation of this sensor, which is developed using an evolutionary algorithm, indicates that it is suited for detecting the alignment of components with three degrees of freedom. Approaches using computer vision are developed for the spatial tracking of the components of the toolkit, allowing the collection of empirical data regarding the interaction of components. The technical advances described within this work will allow the progression of data-driven self-assembly process design.

More information

Identifiers

Catalogue record

Export record