Message received: Graduate student combats information bottlenecks in technology networks
By Rosie Haney
April 10, 2012
When you make a cell phone call, a signal is transmitted from your phone to a tower, and then to the phone of the person whom you're calling. But what if that tower was destroyed? Natural disasters, war and other factors quickly could disable our telecommunications systems.
A solution is in sight: ad-hoc networks. Researchers are looking to the nascent technology to help mitigate communication vulnerabilities, says Qian Zhang, a doctoral student in electrical engineering and computer science.
Zhang is the recipient of the 2011-12 Graduate College Fellowship, which provides an award of $14,487 plus a full tuition scholarship. She is one of five graduate students on the Ohio University campus who received a Named Fellowship this year for their innovative research, scholarship and creative activity.
Ad-hoc networks are decentralized, which means that signals are transmitted directly, device to device. Examples of devices not only include cell phones, but radios, laptops, PDAs that soldiers carry in military applications and sensors in wireless sensor networks, Zhang explains. Traditional networks for such devices are channeled through a center, such as a command station or radio tower.
The benefits of these networks are abundant. When enough devices are distributed in an area, by eliminating the centralized "base station," the signal has the potential for increased range, speed and larger amounts of data. The networks not only could reduce infrastructure costs, but also would eliminate the threat of a center being destroyed and ceasing all communication, making the system useful in times of disaster or war.
The advantage is also a disadvantage, though. With no center, the landscape of the network is always changing. A center provides relative consistency and also helps sort out the messages. Now scientists are challenged to bring all these capabilities to devices that previously only sent and received messages.
Most of the research prior to Zhang's work was aimed at new ways to transmit a message and how to divide and distinguish the data. Developing new devices and methods of operation has been important for this early-stage technology, but Zhang examined how to allocate time and frequency resources most effectively. Do these devices work better when they accommodate one domain or when both time and frequency variables are considered simultaneously?
Zhang found that splitting time and frequency resources was the best way for these units to communicate. This "two dimensional" network allows the devices to talk without relaying the signal through other devices, which prevents information bottlenecks.
Zhang used algorithms and mathematical models to simulate her experimental networks. Her advisor, David Matolak, a professor of electrical engineering and computer science, says that the algorithms were made to simulate extreme conditions such as physical obstructions and sending a message while all units are in arbitrary arrangements or even regular arrangements such as a line.
This industry is still budding, Matolak notes, and research such as Zhang's helps to make unbiased progress. It's tempting for companies to use their own metrics to shed more favorable light on their devices.
"We're working on the ability to efficiently and effectively build networks to solve key human problems," Matolak says.
Zhang will complete her doctoral degree in September and hopes to continue research and development in the private sector.