Morphological Plasticity for the Design, Control and Deployment of Complex Engineering Systems
Abstract
The objective of the proposed work is to develop a framework for understanding the properties of artificial, thermodynamically open systems (ATOS) with morphological and/or neurological plasticity, as well as how to enable such systems to perform useful work. To approach the objective, the PI will examine three fundamental questions: ( 1) What is the adaptive benefit of morphological plasticity for an ATOS, compared to an equivalent system that adapts to its environment only through neurological plasticity? (2) How can an optimized ATOS capable of morphological plasticity be automatically and gradually reduced to a system of equivalent capability, but is thermodynamically closed? (3) Are ATOS s more scalable than systems that lack morphological plasticity? To approach the first question, the PI will perform a series of randomized computational tests in which two sets of ATOS s are optimized: those that have access only to neurological plasticity (the control), and the other which has access to both morphological and neurological plasticity (the treatment). Leveraging.prior work, the PI will approach the second question by establishing conditions under which appropriate geometric, material, and control structures will not change during the operation of a single ATOS. This will involve an investigation of objective functions that favor the rapid genetic assimilation of initially environmentally-triggered traits. For the third question, the PI will investigate scalability of the morphological plasticity argument with the following hypothesis: If an ATOS is optimized to perform some useful work, and two parts of the system should grow differently in response to two different stimuli, work in theoretical evolutionary biology dictates that these two systems will gradually become independent of one another, thus yielding a modular system. Since modular systems are known to be more scalable than non-modular systems, it is here hypothesized that systems capable of morphological plasticity can more easily be scaled up to greater complexity and competency. The proposed project will be deemed a success if it provides proof, through optimization and analysis, that this relationship between optimization, thermodynamically open systems, and modularity will enable the automated design of ATOS s beyond the complexity and competency of those ATOS s that have been manually designed to date.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jan 12, 2017
- Source ID
- W911NF1610304
Entities
People
- Josh Bongard
Organizations
- Army Contracting Command
- United States Army
- University of Vermont