Page 44 - fall2017

which vehicles to deploy, and how many of each,
to best meet objectives while considering trade-
‘Our goal is to devise a system to offs of performance, cost and safety,” Son says.
“For example, to track a group of people
most efficiently and safely deploy moving in mountainous areas under clear blue
skies, the optimal solution might be to deploy
border patrol resources.’ 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.”
Ground-based vehicles have their own In their simulations, Son’s team also will be
advantages. Their sensors better detect objects on adding aerostats, lighter-than-air craft such as
cloudy days or beneath trees, and they produce unpowered balloons that are increasingly used
higher-quality images for identifying individual to track drug traffickers’ low-flying drones and
objects or people. intercept traffickers.
The challenge for the UA researchers is to Using NASA geographical data from the
choose the right combination of aerial and border, the UA researchers have written hundreds
ground vehicles, given different terrain and of algorithms to simulate and predict how groups
weather conditions, and activate them at just the of people may move when traveling on flat desert
right time. and mountains, uninhabited areas and cities, in
“A major task of unmanned vehicles in patrol dry and dusty conditions or during monsoons.
missions is to detect and find their targets’ While the UA researchers are not doing
locations in real time,” says research collaborator field tests at the U.S.-Mexico border, they are
Sara Minaeian, a UA doctoral candidate in conducting experiments outside the lab. They
systems and industrial engineering. “This can be have two quadcopter drones, one purchased
challenging for many reasons: For example, the and the other built with off-the-shelf parts, and
surveillance vehicles and targets are all moving, a ground vehicle resembling a toy car. All are
and the landscape’s uneven nature may alter how remote-controlled and carry a variety of sensors.
targets appear.” In experiments this spring, the researchers
In a paper with Liu and Son published in the used an aerial drone outside on the UA Mall
July 2016 IEEE Transactions on Systems, Man and and inside the Student Union Memorial Center
Cybernetics: Systems, Minaeian describes their to track 10 student volunteers walking in a
novel motion-detection and geo-localization group before randomly dispersing. They also
algorithms for enabling aerial and ground deployed their unmanned ground vehicle to
vehicles to work in teams to precisely locate identify individual people and serve as a moving
targets and decide how to respond. landmark to prevent the UAV from losing sight of
The researchers also have been analyzing and its subjects.
testing different wireless network technologies The researchers are using their experimental
for drones to communicate and cooperate over data to better understand various crowd
varied distances. behaviors, such as gathering and splitting, and
to refine their algorithms to more accurately
predict and track the crowd’s movements. From
Delicate Balancing Act
experiments with a few drones and students,
Establishing when and where to send UAVs the researchers are scaling up their simulation
versus personnel on foot or in trucks is a delicate models to involve hundreds of drones and
balancing act. Factors to consider include fuel thousands of people.
consumption at different altitudes, accessibility, “We believe that by integrating multiple
weather conditions and whether subjects may be surveillance technologies, we can far surpass
armed. their individual capabilities,” Son says. “In our
“Once we have detected, located and integrated system, the sum is bigger than its
identified our targets of interest, we must decide parts.”

42 ARIZONA ALUMNI MAGAZINE