An F-35 Joint Strike Fighter pilot will one day control a small fleet of nearby drones from the cockpit while in flight — according to a new Air Force report on autonomous systems, Air Force Chief Scientist Mica Endsley said.
The Air Force is poised to unveil a new strategy for unmanned aircraft systems next month. The report will discuss more teaming with manned aircraft such as the F-35, greater levels of automation and a wider scope of missions for UAS — such as transporting cargo.
“We see unmanned vehicles being used for a much wider variety of missions,” Endsley said in an interview with Military.com. “Today they are primarily used for ISR, long duration missions where we want to collect information. In the future, they will be moving cargo and more manned-unmanned teaming where they are acting as extensions of a manned aircraft.”
The new Air Force report, called “Autonomous Horizons,” will highlight plans to improve sensors, develop new algorithms and introduce new unmanned platforms.
The Air Force currently flies MQ-1 Predator, MQ-9 Reaper and RQ-4 Global Hawk drones remotely using pilots to navigate from a ground control station. The new strategy calls for additional unmanned platforms and also explains that existing UAS will be increasingly engineered to perform a wider range of functions without needing human intervention – such as data analysis.
“They are going to be smarter in terms of algorithms to handle things like mission planning and collecting data and analyzing that data to take the load off of the human component of a system,” Endsley added.
Endsley said the Air Force will likely begin developing the C-17 cargo planes for unmanned missions, allowing the aircraft to reach high-risk forward locations with supplies, weapons and ammunition.
Manned-unmanned teaming wherein manned aircraft control the flight path and sensor payload of a nearby UAS while in flight is emphasized in the report as critical to the Air Forces’ future plans.
For instance, an Air Force F-35 Joint Strike Fighter might have several UAS assigned to it to perform a variety of missions from ISR to off-board weapons delivery in dangerous or hard to reach areas, Endsley explained.
“We are setting up the ability for an aircraft to take high-level command of UAS. Those unmanned aircraft will have to be capable of flying in concert in a safe manner. They will need to be capable of taking high level commands and be able to execute those effectively,” she added.
The Army has advanced manned-unmanned teaming technology in its helicopter fleet -successfully engineering Apache and Kiowa air crews to control UAS flight paths and sensor payloads from the air in the cockpit.
Senior Air Force leaders have said that the services’ new next-generation bomber program, Long Range Strike Bomber or LRS-B, will be engineered to fly manned and unmanned missions.
Also, in September of 2013, the Air Force and Boeing flew an unmanned F-16 Falcon at supersonic speeds for the first time at Tyndall Air Force Base, Fla. The unmanned fighter was able to launch, maneuver and return to base without a pilot.
Despite these developments, Endsley emphasized that software algorithms have not yet progressed to the point such that a remotely flown fighter jet can maneuver and react to fast-changing dynamics in a combat environment anywhere near as effectively as a manned jet.
There is often a two-second long lag time before a UAS in the air can respond to or implement directions from a remote pilot in a ground station, a circumstance which underscores the need for manned pilots when it comes to fighter jets, she said.
“One of the reasons you would go to an unmanned technology is to be able to go into more dangerous areas than you want to send humans or to fly longer duration missions. You have to weigh what the right kinds of things to fly unmanned are. Not every mission should be done unmanned. There is a long time lag with remote control. Those time lags can be very difficult for rapid response flight dynamics,” Endsley said.
As a result, Endsley explained that the Air Force is much more likely to use autonomy for ISR and cargo missions as opposed to fighter aircraft missions.
“I don’t think that fighter aircraft are a good target for that kind of autonomy,” she said.
While computer processing speed and algorithms continue to evolve at an alarming pace, it still remains difficult to engineer a machine able to instantly respond to other moving objects or emerging circumstances, Endsley argued.
“I don’t believe we will see fully autonomous systems overnight. We are going to see a slow evolution in that direction as we add autonomy to different functions in the cockpit for different functions in the analysis process or in the cyber arena. We want to be sure that we have effective human-autonomy teaming so that people are still going to be able to do their jobs – automation can increase workload if it is not easy to use,” Endsley said.
In contrast to the Air Force’s apparent position that growth in unmanned technology is not expected to replace pilots of fighter aircraft anytime soon, Navy Secretary Ray Mabus recently said the F-35 Joint Strike Fighter will likely be the last manned strike fighter ever bought by the Navy.
However, Endsley and other experts in autonomous navigation technology made the point that it is very difficult to engineer a machine able to quickly react to unanticipated circumstances.
“Trying to teach a computer to have the same kinds of perceptual capabilities that people have is very difficult. They have gotten better at object recognition but understanding the context in which that object is operating could be difficult,” she said.
For example, an aircraft might succeed in being programmed to locate a specific target but might lack to ability to properly interpret the surrounding context and civilian casualties, Endsley explained.
Also, when it comes to the potential use of lethal force, an existing DoD policy directive requires that a human always be in the loop – regardless of how quickly autonomy develops.
“You can have a lot of variability in situations and it is very hard to program systems to handle every situation. People, on the other hand, are much more able to deal with novel or unforeseen circumstances,” she said.
One analyst agreed with Endsley.
“You need humans for situational awareness,” said Richard Aboulafia, vice president of analysis at the Teal Group, a Virginia-based consultancy.
— Kris Osborn can be reached at email@example.com