Carnegie Mellon Researchers Develop Novel Tools To Make Mobile Devices More Energy Efficient

Assistant Engineering Professors Awarded National Science Foundation Grant for Cutting-Edge Work

Mobile devices are now so ubiquitous that they outnumber the human race, but most remain expensive to operate, consume gobs of scarce energy resources and compete for sparse spectral access on the airwaves. 

To help ease spectral scarcity and curb these energy hogs, Carnegie Mellon University's Jeyanandh Paramesh and Pulkit Grover are building novel tools and systems designed to stem the energy use of wireless data traffic.

With a four-year, $800,000 award from the National Science Foundation (NSF), the assistant professors of electrical and computer engineering are working to establish new information theory, concepts and technological solutions to keep pace with the explosion in wireless traffic, which is projected to outpace "Moore's Law," the trend that predicts that transistor capacity will double every 18 months.

"We plan to design and build prototype systems that are able to respond and adapt dynamically to the environment in order to operate with minimal energy use in changing environments," Paramesh said.

These systems operate in the so-called "millimeter-wave" frequency bands (e.g., 60 GHz) where large bandwidth is available. But the starkly different nature of radio wave propagation (compared to the sub 5 gigahertz bands where current radios operate) forces a conceptual and technological re-think of how millimeter-wave links are designed, the researchers said.

The professors and their research team report that energy consumption of mobile devices has become a bottleneck, and the energy overhead of any new spectrum access remains unclear. And because today's archaic systems are designed with rigid, fixed parameters, there is no room for energy efficiency.

But Paramesh and Grover aim to design a system that will be adaptable and efficient enough to rein in energy use by 50 percent, and perhaps much more.

"Part of our novelty lies in designing systems that reduce the energy it takes to send a message together with the energy it takes to decipher that same message," Grover said. "These two problems have traditionally been addressed in an isolated manner, leading to huge inefficiencies in energy usage by existing systems." 

Grover also is developing new energy-efficient protocols and circuits that can reduce energy consumption by including the right pattern of "redundancy" for correcting errors in communication links of big data centers.

The researchers said the broader focus of this NSF-funded research includes the training of new scientists with experience in cross-disciplinary research, the development of new courses and the inclusion of more women and underrepresented minorities in the research effort.