As the demand for smaller and more energy efficient electrical devices continues to grow, electrical engineers are faced with the challenge of figuring out how to best fit all the electrical components into the devices.
Gayatri Mehta, assistant professor of electrical engineering, and her team of student researchers are taking an innovative approach to this challenge, turning the problem of efficiently mapping electrical components into a web-based computer game.
The game features various series of blocks inlaid on a graph. Players are asked to arrange the blocks more efficiently while adhering to certain constraints. Mehta and her team have been working to develop the game for about a year, and a beta version.
By visually and mathematically analyzing the graphs of the top scoring players, the team hopes to harvest human intuition and develop new algorithms, or mathematical equations, that will help engineers develop the next generation of cell phones, medical devices and other electronics. The interdisciplinary project is funded by a $499,924 grant from the National Science Foundation.
“Humans are very good at observing patterns. Algorithms pick nodes and randomly swap them, but humans can see clusters that can be moved together,” says Mehta. “What we are learning is how humans observe patterns and what strategies they use, and then, we plan to use that information to develop new, smarter and faster algorithms based on these strategies.”
This summer Mehta has 10 students working on the project, including students from computer science, computer engineering, electrical engineering and graphic design. Two of the students, Marc Reisner of John Hopkins University and Natalie Parde of UNT, are participating in UNT’s Summer Undergraduate Program in Engineering Research Program (SUPER). Two other students, Candace Calhoun of Johnson C. Smith University and Akeem Edwards of Old Dominion University, are participating in the the Computing Research Association’s Distributed Research Experience for Undergraduates program.
Anil Sistla, a UNT graduate student in electrical engineering, has been working with Mehta since May and says he enjoys the non-traditional nature of the project.
“This was a unique idea, so we got to start from zero,” says Sistla. “It has been interesting to understand how people think. We are already seeing that the human perspective is better than traditional algorithms in some ways.”
-Alyssa Yancey, News Promotions
Posted on: Tue 24 July 2012