Air traffic control for drones: Engineers introduce low-cost UAV detection technology

With the exponential rise in drone activity, safely managing low-flying airspace has become challenging — especially in highly populated areas. Just last month an unauthorized drone collided with a ‘Super Scooper’ aircraft above the Los Angeles wildfires, grounding the aircraft for several days and hampering the firefighting efforts.

Traditional radar systems are powerful but cannot effectively detect low-flying aircraft below 400 feet. While the Federal Aviation Administration (FAA) has some regulations to manage small, unmanned aircraft systems (UAS) or drones, tracking and safety can be problematic – especially in congested or restricted airspaces. BYU researchers may have the solution.

Using a network of small, low-cost radars, engineering professor Cammy Peterson and her colleagues have built an air traffic control system for drones that can effectively and accurately track anything in an identified low-altitude airspace.

“Radar has been around for a long time,” said Karl Warnick, co-author and BYU professor of electrical and computer engineering. “Instead of having a $10 million spinning dish like you’d see at an airport, we have a simple thing that could be built for a few hundred dollars. The small radars don’t have all the capabilities of a higher-end radar, but a network of small radars can work together effectively.”

Peterson explained how the drone air traffic control system works:

• Multiple ground station computers are connected to radar units, which are distributed around an area.
• These radar units are pointed toward the sky to detect any moving objects within their field of view.
• When a radar unit identifies an object, it records the position of that object in addition to the radar unit itself.
• This information is then converted to a global coordinate frame to be shared with other ground stations to create a comprehensive, time-varying picture of air traffic in the area.

This conversion allows all ground stations to accurately interpret the object's position in real space, Peterson said. To achieve a dynamic air traffic picture, each radar unit must be calibrated or provided with the necessary data to convert from the local frame to the global frame.

“Each radar has a field of view as it’s pointed up at the sky,” said fellow researcher Tim McLain, a BYU professor of mechanical engineering. “You want the radars to be calibrated so they all see an individual aircraft at the same place in the sky,”

Researchers said the small radars could potentially be installed on structures such as light posts or cell towers.

Peterson recently published a paper about tracking drones with their air traffic control system, explaining that their research, completed using funding provided by the National Science Foundation, provides more certainty about real-time drone location — important when considering how to prevent collisions between drones.

While the BYU researchers focused on three radars — each able to track a circular airspace about 500 feet across — the technology could be scaled to a broader network with many radars.

“One company (like Amazon or Walmart) can’t take the whole airspace for an hour, right?” Peterson said. “To be cost effective you need to allow multiple vehicles from different companies to travel through the same area during the same time window. If you want to be safe, you’ll want to know where the other drones are at.”

The effectiveness of the system could be compromised due to weather or an object that bumps into a physical radar unit, causing it to move and point in an unintended direction. But an online calibration allows the radar units to adjust for an inadvertent change in its position as it is collecting data, and to correct for any problems.”

“An exciting aspect of this air traffic control system is that in the course of 10 seconds, our radars can correct for a unit’s new position,” said graduate student and co-author Brady Anderson.

To come to the 10-second correction time, Anderson focused on a mathematical equation that performs the online calibration. The research team demonstrated that this dynamic calibration technique showed clear improvements over research with recorded or “batch” data.

Peterson said that with the algorithms driving the system, the radar units could be swapped out or more could be added, allowing for different capabilities depending on the needs.