Using AI to monitor borders and deploy responders
Border walls have an obvious limitation: People can go under, over or around them. Also, as critics of President Donald Trump’s plan to build a wall between the United States and Mexico point out, they are also expensive to build -- especially when the border is 1,900 miles long.
Researchers at the University of Arizona are developing technologies that offer a smarter way to control the border. Young-Jun Son, head of the university’s Department of Systems and Industrial Engineering, has received a three-year, $750,000 grant from the U.S. Air Force to build an autonomous monitoring and response system for Border Patrol land and air vehicles.
The idea, Son told the UA News, is to “devise a system to most effectively, efficiently and safely deploy Border Patrol resources." He and his team plan to do this by building a framework that ties together ground and air vehicles equipped with cameras to patrol the border, along with data from fixed ground sensors.
The real magic of the system, Son said, lies in its artificial intelligence algorithms designed to determine the most effective response to incursions. "Once we have detected, located and identified our targets of interest, we must decide which vehicles to deploy, and how many of each, to best meet objectives while considering trade-offs of performance, cost and safety," Son said.
Factors weighed by the software include not just the locations of targets and Border Patrol personnel, but also weather conditions, geography and even vehicles' fuel consumption at different altitudes.
"For example,” Son said, “to track a group of people moving in mountainous areas under clear blue skies, the optimal solution might be to deploy six UAVs and two trucks driven by Border Patrol agents; whereas for monitoring a group of the same size traveling in an urban area on a cloudy day, two UAVs and six ground patrol vehicles might be more effective."
He told GCN that “the AI heuristics find an optimal or near-optimal solution quickly.”
While the framework is in testing, deployments of personnel and vehicles are made automatically, but when the system is used in the field, it will advise human operators about ideal deployments, he said. In addition to integrating data from multiple sensor platforms, the intelligence in the system tracks many factors that might escape the notice of human operators.
“UAVs can monitor a broader region in a faster manner, but their monitoring capability wouldn’t work on a cloudy day, and they can’t see the objects under the trees,” Son said. The framework automatically takes such factors into account when determining the ideal deployment of resources.
Son, an expert in computer modeling and simulation, is teaming up with Jian Liu, an associate professor of systems and industrial engineering whose specialization is statistics and data analysis. The two professors have already created algorithms that simulate and predict how groups of people are likely to move in different border areas and under different weather and light conditions.
"We believe that by integrating multiple surveillance technologies, we can far surpass their individual capabilities," Son said. "In our integrated system, the sum is bigger than its parts." What’s more, he said the framework the team is building can be applied to many uses beyond patrolling the border, such as managing manufacturing supply chains and electricity grids.
Posted by Patrick Marshall on Jul 27, 2017 at 12:40 PM