Students from the Department of Electrical and Computer Engineering developed technology that uses sound to detect the location of gunshots. The project earned special recognition at the school's Undergraduate Research Celebration.
In recent years, extremist groups and lone wolf actors have increased the frequency and severity of active shooter incidents.* To combat this trend, a team of Mason students developed a novel solution for community leaders and governments who demand safety and security for their citizens.
Ben McCall, a senior majoring in Electrical Engineering, and his classmates Aryan Toughiry, Puja Patel, Joel Williams, and Rohini Shah developed the Active Shooter Tactical Response Assistant, or ASTRA a system that can discern gunshots from other sounds, and also localize the shot.
The high-level concept for the system is a nodal network consisting of three independent acoustic phased arrays. A phased array is a fixed arrangement of sensors (in this case microphones) that measures the time difference between an event (in this case a gunshot) that determines the direction from which the event (sound) originated.
Each array could each detect a gunshot using matched filter techniques and create a vector pointing toward the origin of the shot using cross-correlation. The information is then reported back to a central receiver over wireless links where it displays the nodes’ vectors on Google Earth and places the location of the shooter at the intersection of the vectors.
“One of the advantages ASTRA is that the nodes were made of off-the-shelf components and open-source software,” said Patel. “This means that anyone anywhere in the world could continue our research and/or implement the system on their own.”
One of the students’ biggest challenges was finding a place to test a gunshot detection system.
“Our number one priority was safety, but we also needed the freedom to operate the system on our schedule and have ‘on demand’ gunshots when we wanted them,” said McCall.
The team found an outdoor shooting range that allowed them to validate a single node implementation and confirm its successful operation and also test the complete multi-node system. They used multiple firearm types, recording and collecting test data for further processing as the system grew. Once they completed the single node version, they began the multi-node stage of ASTRA.
This next stage involved building two additional nodes (for a total of three) and creating wireless network architecture so that each node could report its detection and angle of shooter to the central data receiver. Additionally, the students wrote a program that displayed all of the data in real time to simulate what law enforcement or first responders would be monitoring in a real-world deployment.
“Since our second testing at the outdoor range, we have checked our results for accuracy, implemented an adaptive threshold to account for environmental noise, and created a user-friendly graphical representation of the results on the Google Earth terminal,” said Williams.
A poster presentation about ASTRA was accepted for the school’s Undergraduate Research Celebration on April 10. Learn more about the system by attending the event.
*An FBI study found that between 2000 and 2006, an average of 6.4 incidents occurred annually, but between 2007 and 2013 that average increased to 16.4 incidents annually.