Formal Foundations of Algorithmic Matter and Emergent Computation
Abstract
The overarching goal of this MURI is to develop and test experimental, simulation and theoretical frameworks within which to discover the fundamental principles to understand, enable, and synthesize emergent behavior in self-organizing systems. Self-organization and emergent behavior arise naturally across many fields, including physics, biology, computer science, and swarm robotics; however a unifying framework for the science that drives emergent behavior is still lacking. This MURI focuses on systems that define algorithmic (active) matter: Ensembles of particles that interact locally leveraging their physical characteristics and their interaction with the environment, using limited computational resources, bounded communication, and bounded memory to achieve complex tasks. Hence, this project takes a hybrid computational approach to algorithmic matter that combines traditional logic-based programming and non-traditional computational methods (such as task-embodiment and interactions with the environment), as well as predictions of basic properties of the emergent behavior by the collective (e.g., if the collective behaves like a gas, fluid or solid). In this context, emergent behavior can be viewed as emergent (collective) computation. This MURI explores the computational capabilities of emergent behavior through a continuous feedback loop, with theory and experiment informing each other within a multidisciplinary approach. The research plan uses a three-pronged approach to better understanding, enabling, and synthesizing emergent behavior. Specifically, this MURI addresses how to * predict physical and computational requirements for emergent computation, * determine what non-equilibrium characteristics cause these systems to evolve towards the desired emergent behavior, * design efficient collective computational systems to achieve specific task-oriented goals. The prediction and understanding of the evolution of emergent computation will allow one to determine the minimal computational requirements at the micro-level that can drive the desired task-oriented collective behavior, which in turn will facilitate the realization of engineered systems that can achieve the desired task-oriented goal. The MURI team and structure is experimentally novel in that it will compare and contrast two primary distinct, experimentally realizable testbeds of algorithmic matter: One consisting of macroscopic (~centimeters), robophysical entities (smarticles) with limited but programmable and reconfigurable function that move and translate in 2D, responding to sound and light; the other consisting of microscopic (~microns) semi-autonomous entities (SynCells) capable of harvesting their own energy, moving in solution in 3D, and collecting limited information about their environment. The two systems allow access to different timescales for translation, rotation, relaxation, and equilibrium, for example. The team is particularly well-suited to handle the research goals outlined above, bringing complementary expertise in physics, computer science, swarm robotics, mechanical engineering, and nano-fabrication. The questions being asked cannot be answered within any one of these disciplines alone and will require extensive collaboration across all of those, pushing well beyond current interdisciplinary approaches. This MURI has the potential to revolutionize the design of engineered collective systems across applications and at many physical scales, from swarm robotics to autonomous systems automated manufacturing, to medical and health applications. Moreover, the two experimental platforms occupy two distinct application spaces of future interest to robotics, medicine and remote monitoring. Therefore, domain specific principles uncovered in this MURI will be important to realizing future applications important to the DoD, including self-healing and reproduction, which are broad-reaching functions of interest to the DoD.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jun 25, 2019
- Source ID
- W911NF1910233
Entities
People
- Dana Randall
Organizations
- Army Contracting Command
- Georgia Tech Research Corporation
- United States Army