Collision Avoidance Technology for Unmanned Aircraft Years Away

By Yasmin Tadjdeh
Experts have said that the integration of small unmanned aerial systems into the national airspace will open up a slew of new applications and pump millions of dollars into the economy.
However, a key enabler of unlocking that promise has so far remained elusive for federal regulators and industry.

Sense-and-avoid technology is a conundrum for unmanned aerial vehicle engineers and researchers. Ideally, such a system — which could be airborne, ground-based or a combination of the two — could autonomously alert a drone that it is about to collide into another aircraft, manned or unmanned, a building or even a bird.

Much research has been devoted to finding a viable solution, but experts say at best the technology is still years away from implementation.

“We’re going to be looking at this issue for a long time,” said retired Air Force Maj. Gen. James O. Poss, who leads the Alliance for System Safety of UAS through the Research Excellence (ASSURE) coalition.

ASSURE comprises 21 research universities and more than 100 industry and government partners.

UAS integration into the national airspace will have to be done in baby steps, Poss said. It is the “first wave” of autonomy to really affect the public, he noted.

One major milestone occurred earlier this year when the Federal Aviation Administration — the agency tasked with regulating drones in the national airspace — released a proposed rule for small UAS. The document, which was up for public comment through April, stipulated that operators must fly during the daytime and within a user’s line of sight.

While the FAA is moving cautiously, those rules — particularly requiring line-of-sight operations — could severely hamper the commercial UAS industry, Poss said. “Most industry partners require beyond visual line-of-sight authority to make money,” he said.

For instance, while precision agriculture has been touted as one application for drones, only select farmers will be able to take advantage of the technology as the rule stands now, he noted.
“If you’re doing walnut farming or … [growing grapes for] wine in California you can maybe get away with an unmanned vehicle that can only fly five to seven miles from its operators because they’re relatively compact crops,” he said. But for many farmers whose land spans hundreds of acres, a drone will have to go beyond line of sight to be effective.

However, even if the rule were revised, “it’s very difficult to go beyond visual line of sight without some type of sense-and-avoid [system],” Poss said. “And if you don’t get beyond line of sight it’s just difficult to have a vibrant commercial UAS industry.”

Current sense-and-avoid systems — ground-based or airborne — have their share of issues, he noted.

“The problem with airborne sense and avoid is it requires size, weight and power on the air vehicle … and those are all in very short supply on UAS,” he said. Attaching any type of transponder to an aircraft could quickly eat up precious power.

A number of airborne technologies would likely use the automatic dependent surveillance–broadcast system, a GPS-enabled tool that transmits information about aircraft altitude, speed and location to track other systems, Poss said. The FAA is already requiring that most aircraft in controlled space use it by 2020. But the problem with that is non-cooperative targets, he said.

“If everybody is going around flying with … those ADS-B transponders — which they will in 2020 — you don’t need a sophisticated airborne or ground sense and avoid because the two systems will pick each other up and then all you have to do is write some avoidance algorithms inside the UAS,” he said. “The problem you run into is non-compliant targets — balloonists, hang gliders, flocks of birds, you name it.”

Ground-based systems could pick up on a number of items in the sky including non-cooperative aircraft. However, the stations too have their share of complications, he noted. They require the installation of potentially expensive radars and they only cover a fixed area. Further, if an aircraft loses its communication link to the ground station, it will be left in the blind, Poss said.

“You’ve always got to be prepared for the aircraft to lose [its] link in one way, shape or form, so you’ve got to have some rudimentary detect and avoid [on board],” he said.

Both NASA and the Army have promising ground-based systems, he said. In late 2014, the Army installed its first ground-based sense-and-avoid radar for unmanned aircraft at Fort Hood, Texas.

North Dakota is also leading efforts to develop the technology, said Bob Becklund, director of the Northern Plains UAS test site. Northern Plains is one of six FAA-designated drone test sites across the country that were established in 2014.

One intriguing program called the Limited Deployment – Cooperative Airspace Project (LD-CAP) is emerging from the University of North Dakota, Becklund said.

Researchers studied various algorithms and tested whether unmanned and manned systems could successfully avoid each other, he said. “What we would like to do is continue that research, expand on [it], and we would love to do that in coordination with FAA and broader industry.”

LD-CAP brought together a number of partners, including NASA-Langley Research Center, Draper Laboratory, MITRE and more, said Mark Askelson, a UND professor who works on the program.

During one test, researchers operated a drone in restricted airspace and repeatedly tricked it into thinking another aircraft was on an intercept course and about to collide with it. The system’s algorithms “worked extremely well,” Askelson said.

“We extended that then and created a whole bunch of intercepts so that we could work the algorithms harder,” he said. We “tested several different algorithms, one from UND, one from MITRE and one from Draper Laboratory.”

Defense contractor Exelis is also working on the issue.

In 2007, the FAA awarded the company a $1.8 billion contract to help build the next-generation air transportation control system, also known as NextGen. As part of that award, Exelis installed 654 radio stations across the country to collect airline data using ADS-B and other surveillance systems. The company integrates that information with data pulled from the FAA’s already existing 425 radars to give the agency unprecedented situational awareness into most of the nation’s controlled airspace, said Christian Ramsey, unmanned aerial system program manager at the company.

“Because of the relationship in the contract, we have joint data rights to the data we collect through the system, and we can build products on top of that,” he said.

Exelis is taking that line of products — known as Symphony — and applying it to unmanned aircraft, he said. Symphony RangeVue is a web-based application that integrates the data from Exelis’ radios and the FAA’s radars and displays them. Users can build geo-fences to ensure they stay within their airspace and can see other aircraft that are using surveillance systems, such as ADS-B, he said.

“It’s as good as manned aircraft get today. They get that same data,” Ramsey said. The company plans to offer Symphony RangeVue as a subscription service.

Exelis is not yet calling the product a sense-and-avoid tool because it doesn’t cover all of the national airspace. Rather, the company is pegging it as a way to gain situational awareness, he said. 

While the entire nation is covered by the ground stations, they can only see to a certain altitude, Ramsey said. Low-altitude aircraft are still tough to detect, he noted. The FAA has stated that commercial UAS must fly below 400 feet, which is below the system’s purview.

“If you think of the whole problem as a puzzle piece, we’ve got … 95 percent [of the] puzzle solved,” he said. Ramsey emphasized that Exelis wants to work with other companies to try and solve the issue but it could take years to implement.

Fabrice Kunzi, chief technology officer of Panoptes Systems Corp., a Cambridge, Massachusetts-based UAS startup company, said he was optimistic that a sense-and-avoid system could be found within the next three to five years, and implemented within five to 10 years.

Kunzi believes an airborne system is the safest and most reliable answer for industry.

“Unmanned aircraft have a tendency to lose [its] link and so if we lose link to the aircraft and we don’t have a collision avoidance system onboard, we can’t reliably prove the safety of the operation,” he said. “There has to be something on board the aircraft as a last layer of protection if we, in fact, want to enable safe operation and safe integration of these aircraft.”

Panoptes, which spun off from Aurora Flight Sciences in 2014, is the developer of the eBumper. The system, which includes four sensors and software, can autonomously command a drone to swerve away from a target in order to avoid collisions with other aircraft and buildings, Kunzi said.

It is designed to be platform agnostic, but has been retrofitted on a DJI Phantom, a popular recreational drone, with little effect on its endurance, he said. The eBumper adds about 80 grams to the UAV, but because of lightweight echolocation sensors it only reduces the drone’s endurance by two to three minutes depending on weather conditions, he noted.

The sensors can see out about 10 to 15 feet, he said. Panoptes is working on a second-generation version that would have a range of 150 to 200 feet. A prototype could be developed by the end of 2015, he said.

Topics: Robotics, Unmanned Air Vehicles

Comments (0)

Retype the CAPTCHA code from the image
Change the CAPTCHA codeSpeak the CAPTCHA code
Please enter the text displayed in the image.