The research objective of this award is to develop a methodology for examining system-theoretic aspects of semi-autonomous networked system. These systems consist of a collection of dynamic units that evolve according to a predetermined interaction protocol over an information-exchange network, and a set of units that act as interfaces between the networked systems and its exogenous operational environment. The system-theoretic attributes of particular interest in this project include controllability, observability, and feedback performance of semi-autonomous networks, as well characterization of most influential interfaces- which in this project- will be examined with respect to graph-theoretic and algebraic structures of the underlying network. In this venue, the network features that are interest include its symmetry, degree sequence, diameter, centrality of its nodes, connectivity, and incidence relationship between its spanning trees and cycles. This framework will be developed in the context of a canonical dynamic model, namely, diffusion-like models of interactions and their extensions. Deliverables include a documentation of research results in terms of journal and conference publications, modeling and analysis tools for semi-autonomous networks, engineering student education, and curriculum development.
If successful, the results of this research will provide a framework for augmenting a variety of networked systems for improved performance, including engineering realizations of formations, distributed sensor networks, as well as synthetic lattice-like structures in nano-systems. This award will lead to developing a new course with the aim of bridging a gap between networked systems analysis and control and their scientific and engineering realizations in science and engineering. This goal will further be facilitated by organizing a workshop on networked systems during the final year of the project and the development of a new curriculum for graduate program in systems and control.
The goals of this project consist of three facets: fundamental research in the area of networked dynamic systems, educational activities in terms of curriculum development, including a textbook, in the area of networks viewed from the vista of control and system theory, and broader impacts in terms of influencing research in non-engineered networked dynamic systems as realized in biological and social systems. In a nutshell, a networked system is a collection of dynamic units that evolve according to an unforced or controlled protocol over a signal exchange network. These systems are ubiquitous in diverse areas of science and engineering. Examples include biochemical reaction networks, cellular organisms, neural networks, and intricate synthetic lattice-like structures in nano-systems, as well as engineered robotic systems. In this project, we have developed a design paradigm where system-theoretic attributes of networked systems, such as those pertaining to their controllability, observability, and feedback structure, are examined with respect to the graph-theoretic and algebraic structures of the underlying network. In this venue, the network features that have been of particular interest include the symmetry structure of the network, its degree sequence, diameter, and connectivity, as well as centrality of its nodes, and incidence relationship between its spanning trees and cycles. This framework has been developed in the context of a canonical dynamic model, namely, the controlled Laplacian/diffusion dynamics and its nonlinear extensions. Along the way, we have examined a network-centric controllability and observability criteria, as well as the proposing adaptation mechanism for network evolution toward improved collective system theoretic properties. In the educational domain, we have developed a course entitled Networked Dynamic Systems at the University of Washington as a multidisciplinary course in the systems and control curriculum. The outreach component of this project has led to extensive collaborations with researchers in non-engineering disciplines, including our collaborations with researchers in the area of cellular differentiation who are interested to adopt a network control point of view for a deeper understanding of various intricate biological and cellular processes. Moreover, the PI has collaborated with a colleague at Georgia Institute of Technology (Prof. Magnus Egerstedt) in completing a textbook on the topics of networked dynamic systems.
