Designing Self-Assembling Systems Via Physically Encoded Information

Throughout nature, complex self-assembled entities emerge from decentralized components governed by simple rules. Natural self-assembly is dictated by components, their environment, and the interactions among them -- physical information -- forming a system, describable by a set of rules. The process of self-assembly is equivalent to performing a physical computation, using the interaction and transformation of physically encoded information in a system to build physical information structures - entities as the output. However, designing artificial self-assembling systems remains an elusive goal. Understanding the interplay between information and the generation of a process is required for engineering emergence in this form of natural computing. To investigate this interplay, a self-assembly design methodology was developed, the three-level approach, comprising of: (1) specifying a set of rules, (2) modelling these rules to determine the outcome of a system in software, and (3) translating to a physical system by mapping the set of rules using physically encoded information. The contributions of this thesis stem from using the three-level approach to demonstrate how physical information can be used to: enable the self-assembly of desired entities while reducing errors during the process, address the algorithmic constraints of the problem of designing self-assembling systems, and divide the self-assembly process into time-intervals to create more complex desired entities not otherwise possible. These contributions are substantiated through a set of proof-of-concept experiments. Mechanical components in two and three spatial dimensions (in terms of component movement), are confined to a surface or suspended in a fluid. Vibrational energy from the environment facilitates component mobility. Component shape and magnetic-bit patterns are used to encode interactions. Components and environments are fabricated using rapid prototyping. The successful results demonstrate the feasibility of designing self-assembling systems via physically encoded information.