Automated Refueling of UAVs
If unmanned air vehicles (UAVs) are to become contenders for manned aircraft roles, they need to be refuelable by air. The Automated Aerial Refueling project, an advanced technology demonstration led by the U.S. Air Force Research Lab (AFRL) under the Joint Unmanned Combat Air System (J-UCAS) program, is studying the challenges of UAV aerial refueling. AFRL is conducting flight tests to prove the ability to fly in close formation, while running simulations of more demanding scenarios. We talked with Jake Hinchman, AFRL's program manager, to find out more about this work.
Avionics: What is your program about?
Hinchman: The main purpose is to look at how we would automate the air refueling process. Refueling is probably the most strenuous task Air Force pilots do. The boom length is only about 40 feet [12 m]. We are focusing on UAVs, but the work could apply to manned systems, too. It would help with pilot workload relief in refueling in turbulence and storms, when the aircraft bounce around a lot. Our program is looking at how an unmanned vehicle would fly in formation with a manned vehicle, developing the technology core, which can be tailored and applied to different systems.
Avionics: What's new here?
Hinchman: Very close formation flight with two different sized aircraft is something completely new with this program. This is a unique application of close formation flight.
Avionics: What's the status?
Hinchman: The program started in 2003. So far we've identified requirements and two preliminary UAV formation-keeping designs--one based on a sensor coupled with GPS and the other based on differential GPS. We are in the requirements development phase for the sensor. We are looking at the whole bandwidth, from EO [electro-optic] to IR [infrared]. What you're looking at is either light or heat from the tanker. We are now doing a sensor selection trade study, looking at trades such as resolution and wavelength.
On the differential GPS side, it's basically a combination of GPS and INS [inertial navigation system] on each vehicle. You'd use a data link to trade position information back and forth to be able to get real-time GPS position between the two aircraft, the UAV and the tanker. The actual approach is called carrier phase differential GPS. JPALS [Joint Precision Approach and Landing System] technology is the core system we're using. Boeing is developing the control laws, the Naval Air Systems Command is developing the GPS positioning system, and Northrop is developing the sensor-based positioning approach.
Here at the labs we've developed a generic UAV aerodynamic model, including the control system. The control laws are representative of a UAV, but not specifically of either J-UCAS vehicle.
Avionics: What are the challenges?
Hinchman: One is the "see near" problem--detecting where the tanker is, what it's doing, and flying in close formation with it. Another issue we have is collision avoidance. We're looking at leveraging other sense-and-avoid technologies such as NASA's Access 5 and a separate AFRL program. You need submeter accuracy to fly in close formation with a tanker. Another problem is if another manned aircraft decides to cut across the airspace. We need an avoidance capability to say, "don't go there." Another operational requirement is to fly in light turbulence and stay connected.
The technical focus of the program is how you do close formation flight. That's the hardest problem we're dealing with. The refueling box we'd like to hold is about 3 cubic meters--1.5 meters [4.9 feet] in each direction. We're trying to develop sensors and positioning systems to get within that accuracy.
We're doing a lot of trade studies in the simulator. We've simulated a package of four UAVs in turbulence. The tanker is about 130 feet [40 m] long, so when it rocks, it really moves. How much of that motion do you simulate or emulate in your vehicle?
The other area we're looking at is command and control. Most of our UAV systems are controlled from the ground or via satcom link, and you suffer some kind of delay there. If you're within 30 feet [9 m] of an aircraft, that delay is not going to be acceptable in an emergency situation. So we have the ability in our system to directly command a breakaway, or emergency procedure commands, from the tanker. We are evaluating the tactical targeting network technology [TTNT] data link to trade the GPS positioning information.
Avionics: What's your refueling style?
Hinchman: The focus is on boom/receptacle refueling. Our conops [concepts of operation] are built off the F-16. But decisions we make under the demonstration will be compatible with probe and drogue refueling.
The idea is the UAVs use the same tanker and take the same time as manned systems. We don't have to have special UAV refueling tankers and special UAV refueling tracks.
Avionics: How about flight tests?
Hinchman: We're in iterative flight tests with a variable-stability Learjet [surrogate, manned aircraft] from Calspan. We completed the first flight test in September 2004.
We will show that the generic UAV model can be retailored to other vehicles by tailoring it to the Learjet. So all the development is done on the that generic model.
Avionics: How reliable is GPS?
Hinchman: You're flying underneath another aircraft. As you move up under the tail, it may block the GPS satellites, and you lose your precision. The blockage depends on what configuration you're in.
The first flight test, last September, was "open loop," where a pilot flies the surrogate aircraft. We were trying to gather data on how the GPS system degrades. We wanted to determine whether the blockage was so bad that automated refueling was not feasible. We also collected electro-optical data, so that we could go in and try imaging algorithms and positioning algorithms. We determined that both of these approaches are feasible.
Avionics: Will you test the automation?
Hinchman: The first closed-loop flight test is scheduled for the summer of 2006. Closed loop means that the algorithms, the computer, is flying the aircraft. We will hand-fly the jet into the contact position--right up underneath, within 40 feet of the tanker--where you take in fuel. We're demonstrating that the positioning system and the control laws work in formation flight. This will be the first autonomous close formation flight.
The graduation flight test will take place in the summer of 2007. It will show that we can change formation and move around the tanker. We'll be able to command breakaways from the tanker and show that the algorithms can handle that and move the vehicle.
Avionics: Then what?
Hinchman: We take all these lessons learned and move into our graduation simulation of a four-ship package in the fall to winter of 2007. The simulation will include a rendezvous, where all four UAVs go from separate airspace to the tanker. This is where we simulate emergency procedures and failures in order to show the system works.
We use the flight test to demonstrate close formation flight, which is a key aspect and very hard to simulate. The simulation is where we add turbulence and all the high-fidelity GPS data link models. We bring in a tanker pilot and a boomer. In one scenario the tanker turns in front of the vehicles and everything lines up. But what if the tanker turns early or late? How do we compensate? We're going back into simulation to test all these things out. At that point we'll be pretty well done and transition this to a system program office.
Avionics: Is there a special simulator?
Hinchman: We have a test bed that brings in a boom station--the back of a KC-135 that we've put projectors behind. And we have a UAV operator station. We have the ability to "fly" around the tanker and then bring other manned receivers in. This allows us to test these conops out and different models we bring in--like the control laws.