CAREER: Toward Dependable Intelligent Computing on Batteryless Intermittent Devices

Iowa State University, Ames IA

Investigators

Abstract

This project aims to develop software and hardware techniques required to distribute intelligent computations across a network of batteryless, intermittent sensor nodes. Rather than relying on traditional ultra-low-power chips that assume they have ready access to bulk energy (e.g., wired power or a battery), the project investigates approaches where nodes operate using power supplied solely by energy harvested from their operating environment and, thus, spend a significant fraction of time without sufficient power to operate at all. Specifically, the project focuses on investigating the following novel directions: (1) adaptive hardware design techniques to support a responsive, communication- and energy-harvesting-driven paradigm; and (2) application-specific computation scheduling approaches to efficiently offload computation across a network of batteryless, intermittent sensor nodes. Liberating batteryless, intermittent sensor nodes from needing to be closely-connected to a computationally-capable device (e.g., edge server or data-center) fundamentally extends their potential for impact in fields such as industrial automation, infrastructure monitoring, agriculture, environmental sciences, healthcare, defense, and space exploration. Indeed, the results of the project can benefit any field that needs pervasive long-duration intelligence at the edge that remains robust to poor or limited connection to cloud infrastructure. An inherent aspect of this project is a co-curricular experience that promotes students to follow the entire design process of a chip, thus replicating the real-world complexity of computer chip design as well as facilitating student leadership, ownership, and interpersonal skills in preparation for internships and the workforce. The resulting software and hardware systems, experimental data, and educational prototypes from this project will be made publicly available for scientists, engineers, educators, and industry to use. 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.