Summer Undergraduate Research 2008
The first two projects were sponsored The John S. Rogers Science Research Program. This popular Lewis & Clark College program prepares outstanding students for careers in the sciences by supporting collaborative scientific research between students and faculty. The program aims to attract and retain outstanding students and faculty in the mathematical and natural sciences. Sensor Network Security: Elliptic Curve Cryptography on SunSPOTs
An emerging class of small computers, outfitted with sensors and linked by radio transceivers, can form perceptive sensor networks that will more closely connect the Internet to the real world of birds and trees, ships and bridges. While several security mechanisms and protocols have been developed for networks of regular PCs, many new challenges arise due to the characteristics of power-constrained wireless sensor networks. In our work, we focus on the Java-programmable SunSPOT device and elliptic curve cryptography (ECC). We developed our own implementation, and we investigated run times and energy consumption. Our point multiplication in prime fields is faster than that of Bouncy Castle, an open source ECC implementation. ECC takes between 3.25 and 24.89 seconds and lasts for far more than 600 digital signatures on one battery charge. We explored the impact of various parameters including key size, non-adjacent form, and prime versus binary fields. We believe such investigations to be an important step towards addressing the challenges of energy efficient security for Java-based sensor networks. Samuel W. Bock ‘11, James Elwell ‘10, Dennis P. Gosnell ‘09, Travis S. Mandel ‘11*, Jonathan Perry-Houts ’11, Jens Mache, Department of Mathematical Sciences, Lewis & Clark College (* Carnegie Mellon University). Sensor Network Security _____________________________________________________________
Using Accelerometer Input for 3D Gesture Recognition: Lessons Learned
In recent years, handheld devices have become increasingly common, and more and more have started to contain accelerometers, thermometers, and other environmental sensors. Our goal was to use those sensors not only to sense the surrounding world but also to sense and interpret a user's gestures. To this purpose we made use of the accelerometer included in Sun SPOT wireless devices and experimented with different strategies to analyze acceleration readings and extract meaningful gesture data. We had mixed results but managed to identify a number of different strategies that do not work, and several that show potential for future development. Samuel W. Bock '11, James Elwell '10, Dennis Gosnell '09, Travis S. Mandel '11*, Jonathan Perry-Houts '11, Jens Mache, Department of Mathematical Sciences, Lewis & Clark College (*Carnegie Mellon University). Using Accelerometer Input for 3D Gesture Recognition _____________________________________________________________
The next project was sponsored by The Willamette Valley Consortium for Mathematics Research. This REU program is supported by the National Science Foundation (NSF), Willamette University, Lewis and Clark College, Linfield College, and the University of Portland.
The Game of Go: Statistical Approaches to Artificial Intelligence
Writing programs to play the classical Asian game of Go remains an open problem in the field of artificial intelligence. While programs outperform humans at Checkers, Backgammon, Othello, and Chess, the best Go programs are still easily defeated by strong human players. Recent advances involve Monte Carlo techniques, which choose moves based on the results of many simulated games played more or less randomly to the end. This summer's research involves improved statistical models, attempts to break the board into regions (as human players do), better allocation of time (spending more time on "difficult" moves), and learning via simulated evolution (genetic algorithms). Kevin Imber '09, Seth Pellegrino '10, REU students Andrew Hubbard (University of Maryland) and John Stogin (Princeton), RET teacher Jason Galbraith (Sunset High School), and Lewis & Clark faculty Yung-Pin Chen (statistics) and Peter Drake (computer science). Orego: UCT search Techniques Orego: Strategies for Allocating Thinking Time
|
