Bookmark and Share

Low power/energy design of real-time embedded systems

Principal Investigator(s): 

Abstract: 

As technology scales, transistors become smaller and smaller. It enables more complicated functions to be integrated on a single chip to build all kinds of embedded systems such as cell phones, PDAs, digital cameras, and sensors. Many such embedded systems are portable devices powered by battery, and power efficiency has emerged as one of the most important and challenging design criteria. Power efficient designs not only have longer battery (and thus system) lifetime, but also have better reliability and less cooling cost.

At the Embedded Systems Research Lab in the University of Maryland, Prof. Qu and his colleagues' research covers many aspects of low power system design. The research team is attempting to identify the most power/energy efficient hardware implementation of application software. The researchers study scheduling problems for system that performs multiple real time applications. They investigate circuit level techniques for dynamic and leakage power reduction. They validate our techniques by both computer simulation and system prototying in field programmable gate arrays (FPGA). They also develop low power computer aided design (CAD) algorithms and tools.

Prof. Qu was featured in MIT Technology Review for his research aimed at saving processing power in handheld multimedia devices. In an issue of the Association for Computing Machinery's Transactions on Embedded Computing Systems, Qu and fellow researchers at the University of Maryland describe a simple way for multimedia handheld devices to save power.

The premise of the technique is to alter the playback performance in the multimedia device by removing approximately six frames of digital video per second, a slight reduction that is unnoticeable for human viewers. Lowering the decoding standards for the digital video in this way requires less work for the device's processor, and results in lower power consumption, providing a 54 percent energy reduction, according to Qu and his fellow researchers.

Qu's power-saving scheme could be implemented in either hardware or software.

Prof. Qu and his team's research is funded primarily by the National Science Foundation.