Not so long ago, it was groundbreaking for wireless companies to use drones at all. Among the first news-making developments for unmanned aircraft in the industry were deployments to inspect cell phone towers.
 
But recent research projects highlight far more fundamental ways that wireless leaders are using UAS — to help plan the standards and configurations for next-generation networks, and maybe even to provide telecommunications infrastructure that can fly in where ground-based networks are unavailable.
 
A few examples:

• Qualcomm Technologies in May announced results from months spent testing how well existing 4G LTE networks can keep aircraft systems connected at up to 400 feet above ground level. The upshot: Current networks work pretty darn well, but improvements will be important to accommodate the expected surge in drone traffic and other connected devices.

• Researchers in Finland have used drone photography and photogrammetry to build highly detailed computer models for
predicting problems with signal propagation. The technique captures smaller elements of buildings and the landscape than traditional modeling – rising to a challenge presented by the millimeter wavelengths that are increasingly important for large-data transfer.

• Facebook is developing drones that could one day act as networks unto themselves in areas that lack infrastructure or where infrastructure has been damaged.
 
A foundation of data
 
“I don’t think it’s necessarily news, anymore — using drones for agriculture, surveillance, and inspections and things like that,” says Paul Guckian, vice president of engineering at Qualcomm. “Those are very good applications and very powerful, but we can’t stop there.”
 
The push now at Qualcomm, he says, is toward ensuring that the industry is equipped to handle such activities, and more, on a larger and larger scale and in complex environments.
 
To that end, Qualcomm sought this year to lay a foundation of data on the strengths and weaknesses of current 4G LTE networks with an extensive trial — more than 1,000 flights over several months, plus simulations for more aggressive tests — at the Qualcomm UAS Flight Center in San Diego.
 
Qualcomm is sharing and discussing its research as a participant in 3GPP, which in March accepted a study item to enhance LTE support for drones; and in the FAA’s Drone Advisory Committee.
 
Qualcomm’s flight center is in Class B airspace near the Marine Corps Air Station Miramar, as well as being surrounded by diverse landscapes and dense residential areas. That means the trial required extensive Federal Aviation Administration permissions and was grounded in real- world challenges and relevance.
 
“I think that we were aware of testing that had been done, a little bit of testing here and there,” says Harris Teague, Qualcomm’s lead engineer on the trial. “But what we wanted to do was complement that with kind of a full trial effort that would allow us to get enough data in a controlled way so that we could make some statements that we think would generalize to overall network performance.”
 
For its extensive 4G LTE trial, Qualcomm used drones manufactured and maintained in the company’s robotics group. The workhorse of the outdoor exercises was the 39QC. This custom-designed quadrotor aircraft has a takeoff weight of about 1 kilogram, and it can fly 17-18 minutes at a time on battery power. It can be controlled via Wi-Fi and/ or LTE in flight. The on-board flight computer, called Snapdragon Flight, is a custom board also designed at Qualcomm that uses the Snapdragon 801 processor.

Part of Qualcomm's UAS connectivity testing. Photo: Qualcomm
 
Better than expected
 
What was the biggest surprise in the results? Seamless transitions as drones moved from one base-station connection to another.
 
“So that is called the handover,” Teague says, “and the performance in handover was better than I think anyone expected. In fact, we had essentially 100 percent success in maintaining those connections over handover events during flights.”
 
A related finding: Signals at higher altitudes were strong, in general. In some cases, drones were connecting to base stations as far as 14 miles away. Teague attributed signal reliability in flight to the absence of ground-level clutter.
 
“When you’re on the ground and you’re moving, there’s all kinds of clutter and all kinds of obstacles and multi-path sources for signals,” he says. “When we got up at altitude, yes, there was more interference from a larger number of base stations that can be detected. But it turns out that those signals are not changing as quickly, and therefore the algorithms that are trying to maintain connectivity have a little bit easier time of performing.”
 
Teague emphasizes, however, that the expected surges in drone traffic and other connected devices mean there is still room for improvement as the industry plans for 5G.
 
If a drone in the air has multiple bases from which to choose to receive a signal, that can translate to stronger connectivity for that drone, Teague says. But if the drone is communicating with a lot of base stations in selecting a signal, that’s noise for those stations that could impact other users in a crowded network.
 
“We found that for a limited number of drones in the network, (the noise) is really just not a big deal,” Teague says. “However, when we go into simulations and we look at scaling to larger and larger numbers of drones, this interference grows. And so some of the efforts that we’re looking at for improving and optimizing next- generation networks are to deal with these problems.”
 
Overcoming obstacles
 
As Teague notes, maintaining clear wireless signals on the ground can be quite challenging.
 
But researchers from Aalto University and Tampere University of Technology, both
in Finland, have developed a technique that could help network planners to avoid or overcome problems with signal propagation in obstacle-laden urban landscapes.
 
The idea for the 2016 research came from Vasilii Semkin, who was a doctoral candidate studying radio-wave propagation at Aalto. Semkin, who has a hobby of drone photography, recalled recently that he discovered the science of photogrammetry after taking pictures of some castles, thinking to himself that it would be nice to turn the photos into 3-D models.
 
As Semkin learned about the level of detail possible with photogrammetry, he began to wonder: Could more detailed models improve propagation prediction?
 
Conventional models leave out small elements of landscape and architecture, such as plumbing or air conditioners — elements that didn’t matter much to wireless communication in the past, Semkin says. But those small elements are big enough to disrupt “millimeter wavelengths,” which are increasingly important in next-generation networks because they can transmit more data at a higher speed amid heavier traffic than larger wavelengths.
 
To test the photogrammetry technique with propagation modeling, Semkin and his research partners prepared both a conventional geometric model and a photogrammetric model of an area of Aalto University that spans about 15,500 meters.
 
The photogrammetric model required many more photos from more angles to enable the model to have 28,000 faces, as opposed to just 800. And with one drone — specifically, a DJI Phantom 3 Advanced quadcopter — the team was able to gather the necessary photos in only about 30 minutes of programmed flight time, Semkin says.
 
He was quick to point out that working with a photogrammetric model was still time-consuming, requiring more calculations and more computer memory than working with the simpler model. But the results showed the photogrammetric model was more conducive to accurate propagation predictions.
 
“So I wouldn’t say that [the new technique] is very fast. But it depends on what you need as an output,” Semkin says. “If you need very accurate results and you are working with millimeter wavelengths, then it would be necessary to use this.”

Researchers in Finland used a drone for cell signal propagation testing. Photo: Vasilii Semkin
 
Unmanned infrastructure in the sky
 
While some recent research involving drones has focused on planning and improving ground-based communication networks, Facebook aims to develop aircraft that essentially would serve as networks in the sky.
 
One such UAS is the Aquila, a solar-powered plane that Facebook founder Mark Zuckerburg hopes to reproduce as fleets that would provide internet access in underdeveloped areas. The plane, which is designed to fly at high altitudes for up to 90 days at a time, crashed in a December test flight. But if completed a second test flight at Arizona’s Yuma Proving Ground in May that the internet giant hailed as a success.
 
Another UAS project under development at Facebook is the Tether-Tenna, an unmanned helicopter about the size of an automobile that designers hope can provide emergency internet access after natural disasters. Tethered to a fiber line and power source on the ground, the Tether-Tenna woiuld fly a few hundred feet above the ground to act as what the head of Facebook’s Connectivity Lab, Yael Maguire, has called “insta- infrastructure.”
 
‘Crawl, walk, run’
 
Brent Terwilliger has followed the intertwining paths of UAS and wireless communication with particular interest. He is chairman of the master of science degree program in Unmanned Systems at Worldwide Embry-Riddle Aeronautical University in Orlando, Florida.
 
Terwilliger applauded the forward-thinking collaboration of industry players and academics who are investing time and money in work that will enable future business and humanitarian opportunities, but will not necessarily pay off quickly.
 
“It’s important to look strategically at the areas of growth and opportunity here,” Terwilliger says.
 
“Don’t expect to make a quick buck. Invest in technology that’s going to be the underpinning in a strong foundation … Because, you know, how we use this [UAS] technology, where we use it, what industries are supported, what fields contribute to it, it’s really limitless. I like to tell people: This is our generation’s space age.”