What University of Arizona researchers have been working on is an autonomous artificial intelligence framework that uses real-time data to work out how best to deploy various high-tech resource — ranging from ground vehicles and drones to smart sensors, and other technologies — to surveil the 1,900-mile-long border with Mexico. From a computer science perspective, such an undertaking is enormously complex: Working out when to send a person out on foot, a truck, or an unmanned aerial vehicle depends on factors ranging from weather and terrain to the likelihood that a target might be armed.
To build the system, the university received a three-year, $750,000 grant from the Air Force Office of Scientific Research. The project began in March and will run through 2020.
“The goal of our project is to firstly evaluate various combinations of existing border surveillance technologies, as well as newly available technologies, and secondly devise an optimal solution to coordinate them considering tradeoffs of multi-objectives that often conflict among one another,” Young Jun Son, professor and head of the UA Department of Systems and Industrial Engineering and principal investigator of the project, told Digital Trends. “Those technologies that are considered in our work include UAVs with various sensors and intelligent onboard algorithms, other airborne vehicles and intelligent onboard algorithms, stationary ground sensors, ground patrol agents, and newly developed technologies which may not exist as of today.”
At present, the work is still being carried out through computer modeling and simulation to help the U.S. Department of Homeland Security’s Customs and Border Protection unit gain a deeper understanding of how it will lead to swifter, better-coordinated border strategies. Having started with a relatively simple model, the researchers are now scaling up their simulations models to involve hundreds of drones and thousands of people.
“In the future, we will test them in a real-world environment,” Son said.
