current research

My current area of focus is Urban Sensing or Participatory Sensing. Urban Sensing focuses on developing technological advances for using embedded and mobile sensing in order to enhacne civic life and invigorate public space.

We are looking into using cell phones as sensors (since they have an imager, microphone, location capability, and the ability to connect to other sensors in close range). Particularly, we are embracing the concept of participatory sensor networks where cell phones enable public and professional users to gather, analyze, and share local knowledge. We plan to design an architecture to enhance data crediblity, quality, privacy, and shareability in such networks. The application model follows the idea of having campaigns for such purposes as urban planning, public health, cultural identity, and natural resource management.



recent docs

• FIND Presentation
• DietSense
• Exploratory Sensor Search
• ImageScape

focused areas

• context-aware sensing

So the idea of context-aware sensing is essentially using defined "states" to determine which sensors to employ, at what rate, and for how long. The benefit/cost metric can be power, spatial/temporal coverage, or trying to characterize an event. I would like to employ these techniques for the types of data collection we are performing on the phone (diet/activity monitoring, etc..)


• selective sensing

In the case of selective sharing, I'm interested in determining ways to change the resolution of measurements (namely images, acoustic, and location) to protect privacy and preserve anonymity. Also, selective sharing can be thought of as a end-to-end problem where the idea of preserving privacy can be implemented on the source, a mediator, or on the back end. The ImageScape project that I'm working on right now focuses on ways to browse the thousands of images that are collected to make it easy to indicate which images should be public vs private.


• privacy preserving data mining

Finally, privacy preserving data mining deals with how we can ensure that the user's privacy and anonymity are enforced when collecting data and also when sharing the data. We can think of using techniques such as query restrictions on the data, k-diversity, aggregation etc...

past work

I have also done some work in photonics with Dr. Ali Adibi at Georgia Tech and worked on VLSI design automation with Dr. Sung Kyu Lim of Georgia Tech and Dr. Lei He of UCLA