CAREER: Real-Time Scheduling of Intelligent Applications

University Of California-Riverside, Riverside CA

Investigators

Abstract

Emerging machine learning and data-driven applications are bringing smart technology, new functionality, and automated decision-making to devices and systems ranging from consumer products to major public infrastructure. These new capabilities require real-time scheduling where each intelligent software and their interaction must be performed in a timely manner for safe and successful operation. Doing so effectively calls for innovative approaches to the scheduling methodologies. This project will produce three important advances: (1) improving the real-time performance of data-intensive applications on heterogeneous embedded hardware architectures, (2) exploiting application-level requirements for data-oriented scheduling and efficient system design, and (3) enabling intelligent applications to deal with software infrastructure failures and dynamic environmental conditions. This project will enable greater capabilities across the Internet of Things (IoT) and cyber-physical systems like smart cars, smart buildings, factory automation, and mobile health. It will encourage more robust design of these systems so that they are easier to build, more resource-efficient to operate, and more trustworthy and reliable. This will make it possible to adopt intelligent real-time systems in applications where safety and timeliness are essential. The tools developed in this project will be made publicly available, serving as a design tool for industry and an educational tool for colleges and universities. Further, the research will serve as the foundation for education and training programs. 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.