PECASE: Algorithms for Tractable Computation of Decentralized Control Systems

Stanford University, Stanford CA

Investigators

Abstract

ECCS-0642839 PI: S. Lall Abstract Intellectual Merit The intellectual merit of this proposal lies in the development of a systematic mathematical and computational approach for the analysis and synthesis of decentralized networked control systems. Systems which require networked control include large formation space telescopes, distributed sensor and actuator arrays, the power grid, and the mechanisms for congestion control and routing in the Internet, all of which consist of multiple elements which interact and communicate to achieve system-wide performance. The research will focus on developing mathematical and computational tools that allow the control designer to explore the trade-offs between computational cost, communication constraints, and performance inherent to decentralized control. Broader Impact The broader impact of this research arises through potential impact across a wide spectrum of fields, as these areas are of concern in economics, engineering, and computer science. Immediate applications include formation flight of the space telescopes for the Terrestrial Planet Finder Interferometer, and the ground-based Thirty Meter Telescope. Broader impact will also be achieved through outreach. Undergraduates will be integrated into the computational aspects of the proposed research, in collaboration with the Undergraduate Research Opportunities (URO) program at Stanford. The participation of women and minorities in control engineering research will be actively encouraged by the principal investigator. The research and teaching knowledge developed under this proposal will be actively disseminated to both academia and industry via new curriculum development at Stanford, as well as workshops and short courses at national conferences for which materials will be freely available online.