September 2, 2004
ECE senior Paul Li is on the move conducting wireless research on security measures for transit systems for the Transportation Research Board (TRB). The busy Integrated M.S./B.S. Degree (IMB) student assists Yang Cai, a systems scientist in the School of Computer Science's Human-Computer Interaction Institute. Their project team includes ECE Professor and Center for Wireless and Broadband Networking Director Dan Stancil (Li took 18-758, Wireless Communications with him) and Mel Siegel from the Robotics Institute.
Li became involved in the study as part of a course assignment in Cai's class. They worked with Richard Shnyder from the Port Authority field testing prototypes and proposing new technical concepts. Let's connect with Li:
Because I heard that the ECE program in CMU is one of the best in the states and I am very interested in programming and computer hardware.
At the beginning of the project, I helped Professor Cai to find the suitable equipment, e.g. cameras and wireless cards, for the project. After that, we captured real-time images from the Port Authority Transit (PAT) buses and I started the image processing.
I also wrote Java programs to transfer images between two laptops in an ad-hoc wireless network. They were used to analyze the distance and transfer rate when using different wireless antennas and wireless protocols.
The most challenging part was that I didn't have any experience with image processing. I used MATLAB as the tool and wrote a MATLAB script to detect the moving, human, part of the pictures. There are some techniques (e.g. optical flow) that I learned and used in the script to detect the moving parts. But since there are a lot of variations to light intensity, the result wasn't very accurate. Also, I didn't have a lot of experience with network programming in Java. Now, I am very comfortable doing all of these tasks.
The best thing about being involved in this project is that I can apply what I learned to a real application. Other than that, I also had a chance to try some of the latest network cameras, which are high resolution and wireless ready.
Average 10-15 hours per week.
18-791, Digital Signal Processing I. There is a lot of math in this course, but it is really amazing. You can do a Fourier Transform on a signal input (e.g. voice), process the transformed signal (sample up/down, change the frequencies), convert it back, and obtain a meaningful signal.
18-396, Signals and Systems, is the pre-req of 18-791 and I enjoyed it as much as 18-791.
For GE Energy... My job was to write software to control steam turbines.
I want to be a software engineer or an electrical engineer.
Being involved in a research project is a very good opportunity because it is a lot more fun to learn the application, rather than just the theory behind it.