Student-Professor Team Researches Wireless Networking
Roanoke students are the only undergrads to present at a conference in Italy
For the past couple of years, Dr. Anil Shende, professor of computer science, studied problems with frequency assignments in wireless computing. One of the goals of his work is to maximize the utility of a wireless network while minimizing the number of channels or frequencies used by the network. In the summer of 2005, he had two problems for Roanoke computer science majors Mason Vines '06 and Scotty Smith '07 to work on and later present in Italy.
The problem with frequency assignment for wireless networks begins with the Federal Communications Commission (FCC). Cell phone and Internet industries are assigned a specific range of wireless frequencies by the FCC. With this constraint, wireless companies search for ways to use the least number of channels that will allow the greatest number of users to communicate without interference.
"Very soon," Shende says, "everybody will be carrying wireless devices. Being able to have all these devices communicate with each other will require optimized use of the limited bandwidth of available frequencies."
Shende worked with the students to solve this problem in two ways. First, he showed Vines his scheme to assign frequencies to the devices in a wireless network arranged as a cellular grid, so that the difference in frequencies assigned to neighboring devices is large enough to prevent interference. "Interference is caused when two transmitters are operating on closely related frequencies," Vines explains. Vines' task was to test Shende's conjecture that the difference exhibited by this assignment scheme is the largest possible.
When multiple transmitters are used in a network, it increases the complexity of assigning frequencies. Working with a cellular grid or flat surface would be like working within a one-story coffee shop having everyone on the same connection. However, working in a 3-D square grid, such as a two-story building, will result in more surface area, more frequencies and higher chances of interference. Therefore assigning frequencies to each transmitter in the network so that interference is minimized becomes a complicated process.
Smith's goal was to find such "assignment schemes," or patterns, for a wireless network arranged as a 3-D square grid. Smith was able to perform this research through "exhaustive search" by working with a cluster computer, which essentially is a collection of 64 computers that can act as one. This large computer system allowed Smith to sort through millions of numbers and patterns in a week rather than a month's work for one computer.
Vines and Smith were partially successful in their research by the end of the summer. Vines was able to prove Shende's conjecture under certain conditions, and Smith found one family of assignment schemes for the 3-D square grid networks. They continued working through this process last fall.
In October of 2005, Shende, Vines and Smith were the only United States citizens to attend the Workshop on Frequency Assignment Problems held in conjunction with the Italian Conference on Theoretical Computer Science in Siena, Italy. The two students presented their research and proved part of their thesis to work.
"One thing that impressed most people at the conference was that they [Vines and Smith] were the only undergraduates there," Shende says. Professors, engineers and Ph.D. students from Italy, Spain and the Czech Republic also presented similar work.
"Being able to present in front of an audience that was interested in the subject was a nice change," Vines says.
Shende says the students did a great job presenting their work clearly, which is not easy to do. Presenting their work at a conference and observing other presenters prepares students for graduate school and to be a part of a research community, he adds. "These presentation skills are invaluable as graduate students, whether they are presenting at a conference or defending their thesis," Shende says.
"This is what you do as a graduate student. You are given a problem that no one else has solved. You have to think of other mechanisms to solve the problem, and this gives the students the experience that is not part of their regular coursework," Shende says.
Roanoke's students are not only given the opportunity for a hands-on experience, but are doing the problem solving, researching and presenting that general graduate students have to face.
"Since we were the only undergraduates presenting, I was a little nervous; but now, I see that it was a very good opportunity to know what to expect in graduate school," Smith says.
Shende finished his bachelor's degree from Birla Institute of Technology and Science in India and received his Ph.D. from the University at Buffalo. During his tenure at Roanoke College, he has received a grant from the National Science Foundation and has used the money to support his student researchers as well as buy the cluster computer that helps students perform their research.