Designing a Digital Traffic System for Commercial Drones

commercial drones
by Ken Kaplan
Executive Editor
, iQ by Intel

NASA scientists and technology industry leaders are creating an automated drone traffic system aimed at balancing flight efficiency and safety.

By 2020, lower altitude skies could become wickedly dangerous. That’s when an estimated 2.6 million commercial drones will be in the wild, not to mention several million more flying robots owned and operated by consumers.

“When I say every home will have a drone, many people don’t believe it, but I do,” said Parimal Kopardekar, principal investigator at NASA for unmanned aerial systems traffic management.

PK, as Kopardekar is known to friends and colleagues, is dead serious. Serious enough to create an unmanned aircraft systems (UAS) traffic management (UTM) system to help avoid catastrophic collisions caused by rampant drone traffic, especially if regulations allow drones to fly beyond the line of sight (watch his Talks at Google on how to make skies safer).

“I call what we have today ‘eyeball traffic management,’” he said. “Once regulations allow drones to fly beyond visual line of sight, drones need to be electronically tracked.”

drone launch
A participant launches a fixed-wing drone at a TCL2 demonstration at Reno-Stead Airport. Credits: NASA Ames/Dominic Hart.

The potential for commercial drones is growing as companies find ways to use them for deliveries, inspection of infrastructure, science, and search and rescue missions. Kopardekar wants to get ahead of the curve by creating an air traffic system that will help the fledgling drone economy flourish.

“Drone technology is maturing rapidly,” said Kopardekar. “We can’t wait until 2.6 million commercial drones are flying then build air traffic management.”

He remembers it wasn’t until after two aircraft collided over the Grand Canyon that the government and industry introduced an air traffic management system in 1958.

“We need to have this in place ahead of time so that we can balance efficiency and safety, and allow the economic value to be realized for drone operations in low altitude air space,” he said.

Kopardekar and his NASA teammates began building a drone traffic system in 2015. Powered by cloud computing, artificial intelligence and software automation, it uses application programming interfaces (APIs) so that other services can plug into the traffic system.

“We need a cost-effective, reliable system that ensures safety without burdening operators like today’s main aviation system does, where everybody has to keep talking to the air traffic controller,” said Kopardekar.

For example, Delta Airlines sends the flight plans to a control tower, where a human mediator identifies and separates any conflicts, and then dictates a final flight path.

“We want to reduce that bottleneck by giving operators access to flight constraints, locations of other aviators operating nearby and let operators decide where and when they fly,” he said.

A flexible and scalable traffic system will allow millions of drone operators to use it simultaneously in a safe and reliable manner.

Building on the Past

Kopardekar finds parallels with life in the 1930s, when multiple modes of transportation illuminated the need for balancing efficiency and public safety.

“At intersections there were horse and buggies, trams, cars, and people walking, and it created chaos,” he said. “That is when the lanes, stop signs and all those things emerged to maintain efficiency while insuring safety of every mode of transportation.”

Ames Research Center
NASA’s Parimal Kopardekar (far right) discusses air traffic control systems at NASA’s Ames Research Center in Silicon Valley, California. Credits: NASA/Dominic Hart.

Kopardekar has led air traffic management related research since 1993. According to the Aircraft Owners and Pilots Association, no single scientist has a more central role shaping the integrated airspace of the future.

If his UMT research comes to fruition, commercial drone operators will simply connect to a networked communication system that determines when and how a drone will fly safely and efficiently to its destination. It would be fully automated, similar to how the U.S. Department of Transportation is approaching self-driving cars.

“If you want to operate in the airspace, you grade your trajectory or any other flight operation and send it into the system to check that someone else is not using it,” he said.

Traffic and Other Services

Working with the Federal Aviation Administration (FAA) and a variety of industry leaders, Kopardekar’s NASA team will continue their research and development through 2019, then the FAA will work with drone technology innovators to complete and implement a traffic system by 2025.

The data exchange system will share only necessary information wirelessly between operators and the FAA. Technologies and services that provide planning, scheduling, tracking and sensing could come from a wide variety of industry leaders.

“Suppliers could emerge to help commercial drone operators plan flights and secure access to the air space in a safe manner,” said Kopardekar. “Tracking can be provided by a variety of technology or service providers like Verizon or Intel. Multiple operators and suppliers can interact with each other on one system.”

NASA’s work is inspiring collaboration across industries and universities.

Intel Aero Ready to Fly Drone simulates automated search and rescue operation at Virginia Tech as part of NASA UTM Trials. Photo courtesy of Virginia Tech Mid-Atlantic Aviation Partnership (MAAP).

In early June, Alphabet’s Project Wing, Intel and Virginia Tech performed a variety of exercises to test a drone delivery traffic system being created by the Project Wing team.

“Within a few years, Wing and other companies are likely to have fleets with thousands of UAS in the air at any one time, so we’ll need systems that can dynamically route UAS not only around each other, but around manned aircraft, buildings, terrain, weather patterns and special events,” explained James Ryan Burgess, co-lead of Project Wing, in a blog post.

Ryan’s team tested their UTM platform at an FAA site run by the Virginia Tech Mid-Atlantic Aviation Partnership (MAAP). Three Wing aircraft, piloted by a single Wing operator, performed package pickup and delivery missions in the same area where Intel was piloting two Intel Aero Ready to Fly Drones and MAPP flew a DJI Inspire on an automated search and rescue mission.

“Operators have historically had to steer their aircraft away from obstacles manually; instead, we demonstrated that our UTM platform can automatically manage the flight paths of all these different types of UAS, planning new, clear routes for each aircraft if and when conflicts arise,” explained Ryan.

Capiton: Intel’s Yun Wei, Dale J. March, and Karim Tadros work on Intel Aero Ready to Fly Drones at Virginia Tech during NASA UTM Trials. Photo courtesy of Virginia Tech Mid-Atlantic Aviation Partnership (MAAP).

Nanduri’s team used Intel Aero Ready to Fly Drones, a commercial drone development system with built-in depth sensing and vision technologies. They performed scenarios that a UAS operator might encounter when it becomes common for commercial drones to fly beyond line of sight. They tested technologies that allow flight path conformance monitoring, dynamic adjustments to drone operation plans and contingency management, which requires coordinated communication between the UAS and the traffic management system both for flight planning and for real-time flight monitoring.

“The outcome of these UTM trials will pave the way for the standardization of how drones will communicate with each other and other unmanned aerial systems,” Nanduri wrote in an editorial after the June testing.

Commercial uses for drones are proliferating. Nanduri’s team recently demonstrated how autonomous drones can conduct bridge inspections using a new application dubbed Intel Mission Control running on an Intel Falcon 8+ System commercial drone. The software allows the drone to create an optimized flight path for capturing images of the bridge. It can simplify and automate a bridge inspection, which otherwise requires a skilled pilot to fly the drone to capture the structure’s many angles, and repeat it later to compare and identify any new fissures or weak spots on the bridge.

“Conventional methods of inspecting or surveying sites can result in hours of downtime, revenue loss, delayed work and sometimes even an increased risk of injury when inspectors or surveyors need to access unusual or hard-to-reach places,” said Anil Nanduri, vice president and general manager of unmanned aviation systems at Intel.

“Deploying a drone to capture aerial data for the same purpose can be safer, faster and more effective. Because of automation, drone technology is easy to use, reliable and can deliver high return on investment (ROI) for companies who use drones.”

Reliable Tracking Systems

A traffic system that that securely exchanges air traffic, weather and other information requires a sufficient amount of wireless communication bandwidth.

“Drones are flying data capturing, processing and transmission machines that will increasingly depend on robust, low latency wireless network connections,” Nanduri said.

Industry leaders are exploring how the advent of 5G wireless network technologies will benefit commercial drone operators, he said.

drone traffic
Drone traffic could increase significantly by 2020, creating the need for air traffic control systems in lower altitudes.

For now, Kopardekar is leveraging existing technologies that allow commercial drone pilots to schedule and plan areas of operation, and then broadcast coordinates so others can avoid or adjust their flight path.

“Tracking can be done through cell phone, radius, satellite-based or beacon-based systems,” he said. “Our goal is not to pick the technologies — our goal is to say this is the performance you need.”

Cell phones can reach approximately 95 percent of the world’s population but service only covers 55 percent of the land mass. Where cellular connectivity is not available, satellite-based communication is the way to go, Kopardekar said.

Drones Get Better Senses

Today, drones aren’t allowed to fly beyond their operators’ sight unless pilots get special permission from the FAA. Many believe regulations will soon allow commercial drones beyond the pilot’s line of sight. Some pilots may use goggles and the drone’s camera, but it’s more likely commercial drone use will rely on automation.

As more drone traffic fills the sky, drones will fly closer to one another. Computer vision technologies like Intel RealSense depth cameras and artificial intelligence allows drones to see, detect objects and quickly avoid collisions with electric wires, trees or other flying drones. Kopardekar said some delivery drones will need to be smart enough to detect if children or people are nearby and adjust the delivery location.

Kopardekar saw it as a giant leap forward in May when the International Civil Aviation Organization (ICAO) announced the UN agency would help the aviation industry bring such a global framework for managing drone traffic at low altitude airspace. Kopardekar said it will require governments and industries, even if some compete with one another, to collaborate.

“All of them have a common interest, which is safe access to airspace,” he said.

 

Editor’s note: Follow the airborne revolution and drone innovation news from Intel.

iQ newsletter signup

Share This Article

Read Full Story