The proposed research is to develop a universal framework for the analysis and design of Network Embedded Feedback Systems (NEFS). Any such theoretical development requires a theory of distributed control where sensors and actuators are linked to distributed controllers through communication links, such as a wireless network. Two key ingredients of such a theory are a unified view of communication control and a systematic view of distributed algorithms for problems of optimization and control under uncertainty. Over the last few decades, there are extraordinary advances in microelectronics, in software (as reflected by modern object-oriented systems, design patterns and middleware) and in communication networks (including protocols, wired and wire-less networking). Today, everything is getting sensed: vehicles, roads, buildings, airspaces, and the environment. Sensors are getting smaller, cheaper, more pervasive and powerful. The benefits of pervasive networked sensing are compelling. Applications include monitoring wildlife habitats for conservation programs, waterways to understand long-term environmental effects, urban areas for intruder detection, and various components of automobiles for integrated safety and warning systems.
Intellectual Merit:
The rapid convergence of sensing, computing and communication technologies on cost effective, low power, miniature devices will enable a revolution in NEFS. These are distributed control/estimation systems with large numbers of sensors, actuators, and computing nodes that communicate over wire-line or ad hoc wireless networks. Enabling NEFS across this extremely diverse application set requires a unified analysis and design methodology. The feedback nature of NEFS demands extra-ordinary reliability. For this, they need to adaptively respond to unpredictable events such as node failures, communication link failures, and network conditions such as latencies and packet losses. These are the core problems that are addressed in the proposal.
Broader Impacts:
The scientific impact of the proposed research will stem from efforts at developing fundamental system-theoretic analysis and modeling methods for NEFS. This fundamental theoretical development is vital to enable subsequent design and analysis research for NEFS. The technological impact of the proposed research is that it will provide insight into the systematic, cost-effective design of NEFS in a variety of applications. This research is linked to a layered architecture approach which permits design tradeoffs such as computational resource allocation, communication protocol choice, and power/reliability compromises to be readily navigated. The aim is to link this research to a doctoral program by developing a graduate course on Distributed Control over Networks.
