Focus on new faculty: Xinyu Zhang, mixing disciplines to improve wireless communication
Wireless networks don’t always do the best job of communicating with each other, and neither do the disciplines behind those networks. That’s why Xinyu Zhang brings an interdisciplinary approach to sorting out the obstacles that keep us from realizing the full potential of wireless networks and mobile computing.
Zhang says there’s a gap between communications systems thinkers and computer systems researchers when it comes to wireless networks. Zhang combines the two fields: He holds a bachelor’s degree in communications and electronic engineering from Harbin Institute of Technology in his native China, an MS in computer engineering from the University of Toronto, and a PhD in computer science and engineering from the University of Michigan.
Not only is each half of the equation important, but Zhang says people from either side tend to neglect the nuances of the other.
“Traditionally, computer systems people abstract the underlying communications layer as just hardware that provides you with a certain bit-rate, but it’s much more complicated than that,” Zhang says. “You need to understand how the communications system actually works.”
It’s fortuitous that Zhang understands wireless networks from both angles, because wireless networks present a broad range of problems.
Even before joining the Department of Electrical and Computer Engineering in 2012 as an assistant professor, Zhang contributed to research advances that caught the eye of the computer and wireless industries. As a PhD student at the University of Michigan, Zhang helped professor Kang Shin invent GapSense, which PC World described as a way for wireless networks to say “excuse me” to one another. By enabling Wi-Fi, Bluetooth, and ZigBee networks to send energy pulses to tell each other when to hold off transmissions, GapSense can reduce interference among the networks.
Before that, Zhang and Shin tackled all the battery life that smartphones waste while they’re in idle mode, but still searching for wireless signals. “About 80 percent of the time, energy is wasted in idle mode,” Zhang says.
The solution, known as E-MiLi (short for Energy-Minimizing Idle Listening), slows down a phone’s Wi-Fi card’s clock until the phone recognizes an incoming message. The research won Shin and Zhang the ACM MobiCom Best Paper Award in 2011.
Again, the research enabled Zhang to bring his computer-systems and communications-systems training together. “Computer-systems people don’t know how the underlying communications system works, so we try to bridge this gap by redesigning Wi-Fi protocols so devices will consume less power during idle moments,” he says.
The technology has since been licensed to a major smartphone manufacturer and might make it into that company’s next generation of smartphones.
Perhaps not surprisingly, the journey from China to Canada to the United States has left Zhang with an unusual combination of interests. As an undergraduate, he learned tai chi. While in Toronto, he developed a taste for foosball, and notes that he once beat two of his labmates using a single hand. While at Michigan, he played bamboo Chinese flute in the Ann Arbor Chinese Traditional Music Ensemble. He also plays tremolo harmonica.
Shortly after completing his PhD in Ann Arbor in July 2012, Zhang joined the UW-Madison faculty as an assistant professor of electrical and computer engineering. Here, he has continued marrying his computer systems and communications systems expertise in pursuit of better wireless-network performance.
In his time at UW-Madison, Zhang has taught courses in mobile computing and turned his research focus to another big obstacle wireless networks face, which he dubs the “capacity gap.”
“The communication capacity of wireless networks is very high, but actual network performance is several orders of magnitude slower than that,” he says. “That’s mostly because of interference between different access points or different smartphone devices.”
One common and frustrating example, he says, is when hundreds of people at a conference are trying to access a wireless network at once.
“In that kind of situation, many people can’t even get access because of the interference between different devices,” Zhang says. “The overhead becomes so great that the network can’t handle it.”
But by developing protocols to police that interference, more of the users at said conference will be able to get better Internet service.
“I am trying to bridge the gap by enabling signal-level cooperation between devices,” Zhang says. “Such cooperation is transparent to users, and can substantially improve network quality of service.”