CAREER: Scheduling Driving Sensing and Control Nodes in Nonlinear Networks with Applications to Fuel-Free Energy Systems

Power and water systems, the Internet, and other infrastructure are assembled into intertwined networks. These networks evolve nonlinearly in time: small changes in a system’s inputs result in disproportionate changes in the sensed outputs (e.g., a minor car accident can disrupt traffic for hours). These systems form nonlinear networks that rely on ubiquitous sensors and controllers.

Examples include water flow meters and pumps in water systems, ramp meters and cameras on highways, and solar panels and smart meters in energy networks. As a result, the scheduling of sensors and controllers embodies a major step in the reliable operation of various systems and urban infrastructure. When and where sensors and controllers are placed still present challenging questions for experts in this field.

These are not merely engineering questions---they are, by and large, socio-economically imperative. To that end, this Faculty Early Career Development Program (CAREER) award supports fundamental research to understand this scheduling. This project presents an opportunity for scientific advancements by creating innovative algorithms and educational tools for the sensors and controllers scheduling problem in nonlinear systems.

This is a shift from existing literature that focuses instead on linear approximations of nonlinear dynamics, which result in underwhelming performance for various applications. Furthermore, motivated by the urgent environmental need to decarbonize energy systems, applications to the placement of renewable energy resources are explored. The project will also (a) create crowdsourced educational material through open-source computing and (b) build on curriculum development through engaging students from under-represented groups.

The project offers a novel framework for the exploration of the sensors and controllers scheduling problem in nonlinear networks, offering theoretical breakthroughs in dynamic network sciences. In particular, the project will investigate new approaches to efficiently quantify observability and controllability, while exploring parameterization of nonlinear dynamics, first principles in control and network science, and efficient binary search methods.

The foundations will lead to generating subsets of driving control/sensing nodes and have major implications in identifying vulnerabilities and network failures. The research will also create scalable optimization methods that are endowed with theoretical properties that make them suitable for offline/online placement or scheduling of sensing and control nodes, for networks of varying sizes, for systems modeled via differential algebraic or ordinary differential equations, and under the presence or absence of benign or malicious uncertainty.

The project will apply the theoretical foundations in fuel-free power networks that are characterized by the intermittent nature of renewables and consumer behavior.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.