My research interests are developing solutions that sit at the intersection of Computer Science and Health. My recent projects have involved creating Deep Learning models to manipulate or extract information from audio, such as emotion or new bio-signals.
Prior to beginning my PhD, I worked as an Embedded Systems Engineer at a smart LED company, Stack Lighting, a subsidiary of Philips Lighting. In this role, I developed production-ready microcontroller C code for smart LEDs. I also designed signal processing and machine learning techniques for the RF doppler motion sensor embedded within each bulb.
In my spare time, I love staying active, including running, biking, and playing soccer, and traveling. I also believe in using technology to improve the lives of those around the world and am an active member in Engineers Without Borders.
Because each Stack light bulb contains an RF doppler motion sensor, I created an approach using the doppler data from 3 light bulbs to determine the location of a moving object in a room. The technique converts the frequency of the 3 sensors into angle of arrival, then uses an adapted triangulation method to find the object's location.
The goal of this project was to use light bulbs to detect falls in nursing homes. Since falls produce a distinct signal in the frequency domain when detected by the doppler motion sensor within each light bulb, I devised and coded a machine learning algorithm to extract Short Term Fourier Transform features and determine whether a fall had occurred.
To deal with shower temperature fluctuations at my apartment, I designed, coded, and 3D-printed a shower automation device. Using an arduino, water temperature data was collected via a temperature sensor water-proofed in thermally-conductive epoxy. Next, by processing the water temperature data through a machine learning algorithm and PID controller, the stepper motor was able to adjust the shower handle to continuously maintain the perfect temperature.
Wanting to validate the data on my sleep tracking smartphone app, I created a sleep tracker using an acceloremetor/magnetometer connected to a microcontroller. Using signal processing, such as moving average and differencing filters, to analyze the data, I was able to compare the motion plot from the smartphone app with my own data to confirm its accuracy.