Research & Work

Associative Processing using Coupled Oscillators

In the spring and summer of 2013, I did some analog design research at the University of Pittsburgh. The purpose of the research is to investigate the usage of coupled CMOS ring oscilltors as a means of pattern matching. Each oscillator is configured such that its oscillation frequency is a function of an input voltage. This input voltage is proportional to the difference between two patterns. These oscillators are coupled resistively such that similar input voltages produce larger oscillations than different inputs. This is because similar inputs produce similar frequencies that lock together in phase. The coupled signal is rectified to a DC value and can be used to guage the similarity of the inputs. The works was presented at the 50th Design Automation Conference.
DAC Poster (9.1 Mb)

Lunar and Planetary Science Acadamy at NASA Goddard

I was fortunate enough to be a part of the 2012 Lunar and Planetary Science Academy at NASA's Goddard Space Flight Center. During my appointment, I worked in the magnetometer lab designing an anti-aliasing filter for the Solar Probe Plus mission. It was nice to finally get to play with some circuits on the job. I also worked on a project for the Juno mission that automatically interprets sequences of commands and generates administrative approval forms, reformatted output commands, and a history of all past commands. The output files were formerly all entered by hand, so this program will make things much simpler in the future. It gave me a chance to brush up on my Python programming. Overall, the academy was an excellent experience, I met some really great people and learned a lot of cool science.

The Federation of Laptop Orchestras

I was involved in The Federation of Laptop Orchestras as a part of Roger Dannenberg's Computer Music Systems and Information Processing course. It was pretty cool to be working on such a large project inside of a class. I was on the sound design team, so it was my job to make the instrument samples that everyone else would be using. The project was pretty intense. The class designed the software that not only turned our laptops into musical instruments, but networked with other groups around the world. We performed live with six other orchestras at Stanford, Texas A&M, Lousiana State University, the University of Colorado, the University of Huddersfield in England, and Queen's University in Ireland. It's a pretty cool concept. Due to latency over the network, the concert will sound a little different at each venue. Below are a couple press releases and other links from around the web. You can see me in the picture, I'm the guy who looks like he didn't know a picture was being taken.

Predictive Indoor Navigation on Commercial Smartphones

The Indoor Navigation app features onboard localization and predictive path planning and was designed for the Nexus S smart-phone. My contribution to the application was the design of the hierarchical path planning algorithm, map representation scheme, and user interface. The path planner recursively establishes a route between the users position and the specified or predicted destination. A fine-grain path is calculated using an adapted version of the D* planning algorithm and clearly displayed for the user. D* is an incremental search algorithm that replans a user's route in a dynamic environment. The interface was designed to be as intuitive and simple as possible. Our team successfully overcame the challenges of performing computationally intensive tasks on a handheld device that is fairly limited in memory and processing power.

Modeling Self-Healing Materials

The design of materials that can spontaneously heal any damage imparted on them is of great interest to the fields of biomedicine and material science. These simulations were designed to test the integrity of polymer materials cross-linked with strong carbon bonds as well as lower energy bonds, disulfide or hydrogen bonds for example. As the material experiences tensile deformation, the lower energy labile bonds can reconnect with reactive groups on adjacent nanogel particles. My role in this research was to introduce the Hierarchical Bell Model to a new domain, one in which the bonds could not only rupture, but also reform. I designed the probabilistic model governing the behavior of parallel labile bond systems and implemented it into the original model. Additionally, I ran numerous tests and varied parameters to optimize the material's properties.