Embedded systems are computers within devices that are not generally recognized as computers, e.g., computers within automobiles, medical devices, and personal communication devices. The design of embedded systems is complicated by their numerous conflicting objectives, e.g., short design time, low price, low power consumption, high performance, and reliability. Manually designing embedded systems that meet all these goals is an arduous task. This project targets the automatic analysis, design, and control of embedded systems.
This project is developing optimization algorithms and software that automatically designs low-power, low-price, high-performance embedded systems and integrated circuits. A fast and accurate profiling and memory access tracing based infrastructure for detailed power and performance analysis of embedded applications and operating systems will be implemented. The theory for power-aware operating system scheduling and device control will be further developed and the results embodied and made public in a new version of a widely used embedded operating system.
Embedded systems (special-purpose computers) make it easier to communicate, improve the speed and safety of travel, manage information, and assist people with hundreds of other tasks. This project is making it possible to build better and less expensive embedded systems. In addition, software developed in this project is publicly released as a reference implementation to improve the battery lifespan of embedded systems that run open-source operating systems. This project is integrated with Northwestern University courses on embedded system design and real-time systems. Through it, undergraduate students are benefiting from, and actively participating in, academic research.
