Grey

Nearly everyone today has a cell phone. They allow us to communicate and operate at a level hardly imaginable just 10 years ago. With their rapid growth, which shows no signs of stopping, cell phones are quickly becoming the first truly ubiquitous computing environment. People have long imagined such a pervasive computing platform, and those dreams are quickly becoming a reality.

Grey hopes to expand on the current power of the cellular telephone to allow you, not only the standard convenience of anywhere communication, but the power to control objects around you. Open doors, log into computers, and yes, even go shopping, all while carrying nothing more than your cell phone. Grey promises to develop all of these, and to do it in a way that far exceeds the security to the user offered by any current method.

Through the careful application of security measures, Grey can offer expanded convenience to its users, while simultaneously providing them with better security and control. From capture-resilient keys, to one-time use credit cards, Grey can offer all this, and much more. Allow your co-worker access to your office when you're on the road, and be certain only he can enter. Let your secretary read your email, without giving her your password. With Grey, all of this will soon be a reality.

Claytronics

Claytronics is the next logical step in recording and playback. It started with audio and the telephone. Soon after came video and the television. The next generation is full 3D rendering of any object. We have called this new media "pario", and claytronics are the building blocks necessary to realize this dream.

Initially conceived by Seth Goldstein and Todd Mowry, claytronics, when finished, will revolutionize countless fields. Teleconferencing will now occur with everyone physically in the same room. You can watch a hockey game played out in 3D on your desk. Movies can come to life, really! The applications are endless.

Presently, however, we are a long way from that reality. Right now we can only simulate such a world and attempt to solve the software problems, so that when the hardware is ready, the software will be too. By addressing the problem now, before it is financially viable to build prototypes, we can have solutions to the problems of power routing, motion planning, stability, communication and many others by the time they are needed for deployment.

Steganography
Rensselaer Polytechnic Institute - Center for Image Processing

Working with Jeremiah Harmsen, I extended previous work in the area of steganalysis - detecting information hidden in images. Just like cryptography and cryptanalysis, one tries to hide information, while the other tries to retrieve that information. The only difference is, in steganography, that information is hidden in pictures, rather than in cipher-text. This has the advantage that quite often it is impossible to even know that information is being transmitted. By altering only the least significant bit of a color pixel, the image remains unaltered to the human eye, yet enormous amounts of information can be transmitted discretely.

Thankfully, however, for those trying to catch this information transfer, the image is noticably different to a computer. Specifically, what Jeremiah had found, was that by computing the "center of mass" of the color intensity histogram, it was possible to determine with extremely high accuracy whether a picture had been modified from its original, even without having the orignal. Esentially, all "real" pictures inherently have a random property which can be measured by the "center of mass".

Unfortunately, calculating this "center of mass" is extremely slow and was of little good in any practical implementation. Instead of calculating the center of mass on all 3 colors at the same time, it was possible instead, to calculate a modified center of mass on the three 2-color pairs. Even though the new algorithm required three computations, the time necessary for each decreased a couple orders of magnitude, with only a slight decrease in accuracy. To read the paper, see my publications

NIST SURF

As part of the Summer Undergraduate Research Fellowship (SURF) at the National Institute of Standards and Technology, I analyzed the leasing protocol in Jini service discovery protocol. Working with Dr. Kevin Mills we simulated the leasing proceedure in the Jini protocol using a event driven simulation written in SLX. In the process we developed two adaptive leasing procedures that increased systems responsiviness without degrading performance.