Tuesday, January 7, 2020

URSP Student Daniel Sovine Tests an Alarm that Would Recognize the Unique Sound of a Gun and Set It Off Using FFT

When I began my first semester as an electrical engineering student, I took Calculus I which was being taught by my future mentor. There was a day in class when he had a small tangent and mentioned a project that might interest any student in ECE or engineering students. It was an alarm that would recognize the unique sound of a gun and set it off using FFT. I wasn’t too interested at the time but I kept it in the back of my head. After the months of terrible news coverage of gun violence, it was brought to my attention again by my future partner in the project. We met with my professor and he introduced us to the URSP program. 

Before this project, I was not involved in any large research project. As I’ve progressed through our project, I became more familiar with the research process as a whole. Although all research projects are different and an engineering project is drastically different from a biology project or a sociology project, I feel like I would enjoy this environment for a while. I really went into this curious to see how I would feel about research and if I could pursue it further for the rest of my college career. I hope as I continue on to other projects, I remember what I did for my Shooting Detection Alarm System this semester.

This semester was a very stressful one for my course load. I had to learn how to balance my URSP project with my projects from classes and work. It taught me a lot about time management and what I am capable of. I normally contact my partner, whos in Pennsylvania, through Facetime and we discuss the research that we did between the time of the last talk. I would have all of the materials for the hardware base system and he would take care of most of the software. We had to find a better way of working together, so we used a VNC server to connect the system across networks. If we weren’t waiting for the parts or took a break for midterms, we’d be working on the project each week. We spent most of September waiting on parts to come in the mail. We did most of the fast fourier transform Matlab work in October. In November, we hit a dead end with the audio neural network. However, now we are working with the image of the frequency domain in the neural networks and it’s settling much better than the audio neural networks.

If there was one thing that this term taught me, it would be that plans are never concrete. Things don’t always go as you expect them to. This is especially true in any research project. You may not get approval for your project in time, or you might not get your parts in time. The ability to mold to the circumstances that are out of your hands is a skill that I have gained from this project. That is not to say that you should change your plan too much, it just means that you have to work with how things are.