Congress has ordered FAA to enable civil unmanned aircraft to routinely access U.S. airspace by 2015. FAA needs to lay down the rules of the road. But more is required to get there than one might think, including real-world operational data and work on sense-and-avoid, ground control station, and command and control (C2) technologies.
NASA is assisting FAA on two fronts. The agency plans a Centennial Challenge to enable operators of unmanned aircraft systems (UAS) to fly their vehicles on representative missions in fenced-off airspace. The exercise will test the maturity of UAS technologies, spur innovation and contribute data to the FAA’s rulemaking effort. The agency is also two years into its five-year UAS in the NAS (National Airspace System) program that is looking at requirements for sense-and-avoid systems, C2, and a common ground control station with a view to maintaining airspace safety.
NASA’s Airspace Operations Challenge, among other things, will test the view in the UAS community that their technology is mature and needs only a free rein to take off. Phase 1 of the two-part contest is expected to be held next spring.
“We will provide a venue in restricted airspace where they can fly their vehicles in a simulation of actual airspace,” said Garry Qualls, the project manager. NASA plans to award prizes of $500,000 and $1 million for the winners of Phase 1 and Phase 2, respectively. The size of the purses effectively limits the contest to individuals, universities and small to medium-sized aerospace companies, he notes.
The Phase 1 operating area will look like rural, Class G airspace, compatible with agricultural, mapping and survey applications, for example. Phase 1 is intended to show whether the participating vehicles can be operated as safely and intelligently as aircraft that have human pilots on board, Qualls says.
Competitors will fly one at a time on courses tailored to their vehicle performance characteristics. They will have to sense and avoid cooperative aircraft — those employing ADS-B-Out — through use of ADS-B-In on board. They will try to avoid deliberately inserted traffic snafus and occasionally contend with faulty or absent GPS data to prove the robustness of their navigation solutions.
While the rules have yet to be finalized, Qualls plans to use a lost C2 link scenario as part of Phase 1 qualification, although, for safety reasons, it may be a simulated rather than an actual lost link.
Phase 1 contestants will also have to have a traffic display as part of their ground stations, Qualls says. Participants won’t be able to score highly simply by observing their aircraft visually. They will have to use the information on these displays to make the correct decisions about how to respond to traffic situations.
NASA is putting together a “traffic squadron” of unmanned aircraft which it can use “to create controlled, repeatable air traffic situations,” Qualls says. These vehicles will “closely resemble general aviation aircraft to make it clear, both to the public and the competitors, that the main point of the competition is to determine whether UAS can operate safely around manned aircraft.”
Phase 2, slated to take place about a year after the successful completion of Phase 1, will focus on demonstrating sense-and-avoid technologies to maintain safe separation from cooperative and non-cooperative air traffic. Competitors will be expected to have ADS-B-In and –Out on board. They will also be expected to have “an onboard system capable of communicating verbally with air traffic control and surrounding aircraft during a lost link situation,” Qualls says. This would involve voice recognition and voice generation software on board the aircraft.
UAS in NAS Program
The $160 million UAS in the NAS program, on the other hand, is attempting to reduce or eliminate barriers to civil and commercial use of UAS in the national airspace, says Chuck Johnson, program manager.
In the sense-and-avoid area, NASA is trying to help determine how safe a system needs to be for UAS vs. manned aircraft since the former have such different performance characteristics. The agency will collect data based on human-in-the-loop simulations and flight experiments, Johnson says. He predicts UAS will end up having to have some kind of active collision avoidance system on board.
The program also is looking at C2, trying to validate that the civil frequencies recently approved by the World Radiocommunication Conference (WRC) are usable, secure and scalable, Johnson says. The agency has contracted with Rockwell Collins, which is designing a waveform and a prototype radio.
A third area is the ground control station (GCS). GCS issues may be the biggest challenge in the program since not very many people are working on them, Johnson says. Military ground stations are all proprietary and vehicle-specific, he adds.
A GCS is similar to a cockpit, but aspects such as the feel of an aircraft as it approaches stall speed, are quite different. What kind of indicator would capture the remote pilot’s attention? The good news is that there is more real estate in a ground station than a cockpit, and Johnson expects that “ultimately, unmanned aircraft probably will be much safer than manned aircraft.”
Command and Control
C2 is a critical area that must be understood and dealt with before unmanned aircraft can be injected routinely into the airspace: there is no analog to it in directly piloted aviation. So it is essential to develop a reliable, robust, low-latency, scalable, secure and FAA-certifiable C2 data link to maintain safety as large numbers of commercial unmanned aircraft enter the airways.
In 2011 NASA awarded Rockwell Collins a three-year cooperative agreement to develop a prototype waveform and communications architecture for potential use as a C2 data link for commercial UAS ground stations and vehicles. In this partnership NASA is working on higher-level issues such as encryption and network access control.
Last year the WRC allocated two blocks of spectrum — in the L-Band and C-Band — for the command and control of commercial UAS from terrestrial stations, explains John Moore, Rockwell Collins’ principal investigator for the project. There is still an open agenda item to allocate spectrum for SATCOM-based C2 which will be addressed at the next WRC, in 2016.
The dual spectrum allocation for ground-based C2 could potentially improve availability, Moore says. If there’s trouble receiving on one band, maybe the other will have better reception. Having two blocks of spectrum could also enhance reliability, fault tolerance and single-mode fault rejection, he adds.
Some, however, are suggesting that the C-Band spectrum be reserved for airport areas since it can handle higher data rates and loading, he notes. In the airport environment, many aircraft will be asking for bigger bandwidths at the same time. While a final decision has not been made for an operational system, Rockwell Collins is evaluating both the L-Band and C-Band spectrum in all potential modes of operation, he says.
“We’re trying to maintain as much commonality as possible” between the L-Band and C-Band versions of the waveform architecture, Moore says. But there will be differences. In addition to the higher data rates possible at C-Band, “when you move up in frequency, it’s more complex — for example, there are different … Doppler shift problems at the higher frequencies.”
The first prototype waveform was delivered to NASA in February 2013, and the agency finished the first phase of flight testing in May. The critical design review of the second prototype — which includes both L-Band and C-Band — was scheduled for June 2013. The third and final configuration of the prototype waveform is slated for delivery to NASA in 2014.
What’s in a Waveform?
The command and non-payload communications (CNPC) waveform is by definition a clean design from the ground up, Moore explains. NASA looked at more than 70 legacy waveforms, but found nothing off-the-shelf that would meet the requirements without modifications.
Rockwell Collins has developed a prototype waveform, using information from the International Telecommunication Union report (ITU-R) M.2171 — on UAS characteristics and spectrum requirements — that was the basis of the WRC allocation. The lengthy report contained, for example, estimates of the numbers of UAS expected in coming decades and considerations regarding cell size.
The issue of scalability is the biggest challenge, Moore says, leading, among other things to a multiple access uplink technique in high-density areas that is relatively more sophisticated than what’s used in civil aviation data links today.
“It’s going to be a different class of data link from anything in the civil world before,” he says. It will be quite unlike ACARS, for example, in reliability, data throughput and total system capacity.
Moore describes Rockwell Collins’ work as a bottoms-up approach, designing the CNPC waveform based on sizing numbers — estimates of how big the UAS fleet might get in the next several decades. “Then we looked at how much data it would take for command and control.”
The requirements called for a waveform that could provide sufficient data bandwidth on both uplink and downlink to support a maximum density of 20 UAS simultaneously in airspace cells about 60-70 miles in radius. The “cellular lattice” will probably also be layered by altitude, Moore says, since high-flying aircraft can transmit for significant ranges and still be heard — and possibly interfere with other transmissions.
At the high end, each cell would contain an average of 10 UAS at a time, with a “peak pulse” of up to 20 in certain sectors for short periods of time, Moore explains. The figure of 20 probably won’t be reached for some time, he emphasizes. In some cells it may never be reached.
The final size of the cells is yet to be determined, and more work is likely to be done. However, there are significant trade-offs between larger and smaller cells, Moore explains. Smaller cells mean lower-power systems but complex frequency planning. Larger cells mean easier scalability and easier frequency planning but higher-power systems and potentially more issues with terrain masking, he says. Most of the work so far suggests that a 60-70-mile radius is the likely balance point in benefits, he says.
The company has developed a physical-layer signal in space and is verifying that simulated and predicted behavior is matched out in real life, he says. In the near term the data link will also include pilot-to-controller voice communications, he adds.
This waveform approach is intended to accommodate both individual direct line of sight operations — when densities are low — and a networked configuration when densities are higher. In the latter case, the ground stations would transmit to all the aircraft in one cell in time slots, using time division multiple access (TDMA). All the aircraft would be tuned to one frequency for the uplink. Orchestrating the time slots would require synchronization via a time source, such as GPS, with a backup mitigation in case of signal loss.
|Rockwell Collins engineers set up equipment as
part of the company’s project to prototype UAS
waveform and communications architecture.
On the downlink, conversely, each aircraft would talk at the same time on its own allotted frequency, using frequency division multiplexing. The asymmetrical structure of the data link maximizes efficiency and minimizes the transmit power, weight, size and expense of the aircraft radios.
It is not clear yet who will build the various elements of the ground infrastructure, Moore says. But it will probably be a mix of individual ground control stations, local/regional networks and possibly a third-party service provider, comparable to ARINC, for aeronautical operational control communications.
In the prototype communications architecture the maximum required uplink data rate will be about 7 Kbits/sec. The bandwidth of the downlink will vary from vehicle to vehicle, depending on their phase of flight. A vehicle in taxi, for example, may need a relatively high-bandwidth video link — not for the payload but for pilot situational awareness in reading runway signs and watching for other vehicles. The architecture is intended to allow this flexibility, so a vehicle in taxi may use as much as 235 Kbits/sec on the downlink, while a vehicle in cruise might get by with 14 Kbits/sec.
The waveform also supports a 20-Hz frame rate, which effectively means that downlinked data could be updated 20 times a second. This update rate, at the limit of human cognizance, therefore allows “real time,” hands-on control. The update rate is high enough and the latency of the link is low enough to allow an operator to hand-fly a vehicle against instruments, Moore says. Although hand-flying is something most vehicle operators don’t do today, the waveform and architecture need to include that possibility.
Generally, the architecture deals with vehicles that are too large to fit the FAA’s forthcoming small UAS rule, he says. The small vehicles (less than 55 pounds) covered in that rule will be operated at visual and electrical line of sight. But Rockwell Collins feels it’s important to include smaller vehicles, such as the Scan Eagle, that will be operated beyond visual line of sight.
RTCA has just announced a new Special Committee, SC-228, to continue the work of SC-203. The new panel, which was slated to hold its first public meeting in July, is charged with developing an initial set of standards for certification of a C2 data link and detect-and-avoid systems.
To be certifiable by the FAA, it’s a good idea for the waveform to be as simple as possible, Moore says. “Reliability and predictability of performance are far more important to the FAA than the level of complexity.” But the need for simplicity is driven even more by economics. The more complex the system, the more costly it will be to prove its reliability and predictability. And since many people are looking for UAS to be more cost-effective than manned aviation, “the market needs the most cost-effective solutions that can meet the FAA’s performance requirements,” he says.
Rockwell Collins is using a “constant envelope with binary order Gaussian Minimum Shift Keying,” which is “pretty vanilla,” Moore says.