Best listening experience is on Chrome, Firefox or Safari. Subscribe to Federal Drive’s daily audio interviews on Apple Podcasts or PodcastOne.
Manned and unmanned aircraft operating safely together is still an air traffic control challenge. It’s a step closer to a solution thanks to an award-winning project at a federally funded research and development center know as FFRDC. The Federal Drive with Tom Temin talked to someone who shed light on the project, Wes Olson, a retired Air Force...
Manned and unmanned aircraft operating safely together is still an air traffic control challenge. It’s a step closer to a solution thanks to an award-winning project at a federally funded research and development center know as FFRDC. The Federal Drive with Tom Temin talked to someone who shed light on the project, Wes Olson, a retired Air Force officer, who is now a group leader at MIT’s Lincoln Laboratory.
Tom Temin: Mr. Olsen, good to have you on.
Insight by Splunk: During this exclusive webinar, moderator Tom Temin will discuss data and the concept of observability with agency and industry leaders.
Wes Olson: Well, thank you, Tom. It’s good to be here.
Tom Temin: And let’s begin with the problem that you were trying to solve here. Because the FAA at this point, I think keeps drones out of the airspace altogether. But it sounds like inevitably, we’re going to have to be able to find a way for those types of aircraft to coexist?
Wes Olson: Exactly. So the whole goal of this project is to increase aviation safety, we in the U.S. have one of the safest air transportation systems in the world. We’re a world leader. And our goal as we introduce new entrants, is what we call them, whether they’re larger uncrewed aircraft, smaller uncrewed aircraft drones, as the public would call them, or advanced or urban air mobility, the new concept, we’re going to have highly automated vehicles transporting the public to reduce congestion and drive time. And all of these need to be integrated safely. And our experience in aviation is that although the air traffic control system is very safe, there needs to be a final safety net at that last ditch your moment to prevent a collision between two vehicles. Whether those vehicles are a large transport category aircraft or smaller ones, this is essential to maintain aviation safety. So we were trying to develop a flexible, adaptable system that would meet the needs of current aircraft, but also future aircraft with new methods. We call it surveillance of knowing the location and identity of those other aircraft and have a common framework that’s easy to implement, and easy for the FAA to certify and for the international community to accept because this is a worldwide system, right?
Tom Temin: It’s called the Airborne Collision Avoidance System X ACAS I guess, or ACASX. Just in layman’s terms. Briefly, how does it work?
Wes Olson: Well, it ACASX works. It’s a modular system. So the first part of the system is determining the location of the other aircraft. Now for larger aircraft, there is a transponder on board, much like the toll transponder on your car. So one aircraft would send out a signal to all the nearby aircraft, they’d respond. And based on that information on the altitude and location, the system could decide whether or not an alert is necessary. For smaller aircraft, there might be an onboard radar to detect these aircraft that don’t have a transponder. Also, with advanced technology, using GPS aircraft can broadcast their position twice a second, so that information can be used. So our system takes in that surveillance information, correlates it if there’s multiple sources, determines which is the best source, and also determines how certain that is, just because I think I know where the other aircraft is, doesn’t mean that’s actually true. There’s some level of uncertainty even with GPS. So we calculate that uncertainty, pass it to what we call the threat logic, which determines whether or not there’s a threat, and takes into account multiple competing objectives. So we want to be safe, but we don’t want to alert all the time. If you’ve ever had an alerting system in your car that went off all the time, it’s very annoying, it’s annoying air traffic controllers and pilots. So we use machine learning techniques to come up with the optimal solution that balances safety. What this needs to save, you know, to integrate these vehicles in the air and that information, what to do, whether it’s to climb or descend, turn left, turn right, or potentially even in the future speed up or slow down, is passed to either the flight crew operating the vehicle, or to the automated systems that control the vehicle. And that information, these maneuvers are also coordinated. So if you’ve ever been walking down a hallway and encountered somebody else, you both go left you go right? Well, we want to avoid that in the air. So these systems also send a message to the nearby aircraft saying, I’ll go this way and then the other aircraft would make a complementary maneuver. It also takes into account multiple aircraft. So we could solve, you know, 5, 10, which is very uncommon, very uncommon, but we can solve multiple competing conflicts at the same time. So that’s, in essence, how this system works.
Tom Temin: I sense a lot of artificial intelligence, and just a lot of mathematics running very fast in all of this.
Wes Olson: It is, it does take advantage of modern computer science techniques. This is a very complicated problem. And if we were just to solve this via classical methods, where we tried to look at the optimal action at every point in time that would exceed the number of atoms and molecules in the universe. So we use advanced computer science techniques, as well as machine learning techniques to make this a tractable problem, which means we can solve it rapidly. And the end product can fit on an aircraft. So it doesn’t take massive computing power, especially for the smaller aircraft, the drones. And it’s also the case that, you know, with these smaller vehicles, it can operate on a ground system, so it’s flexible. It could operate in the ground or in the air to provide that same level of capability.
Tom Temin: We’re speaking with Wes Olson. He’s a group leader of the Surveillance Systems Group at the MIT Lincoln Laboratory. And what is the status of this with respect to it becoming operationalized? Because FAA has a very long process before things are actually relied on in day to day use. So where do we stand in all of this?
Wes Olson: So that’s an excellent question. And you know, people from the outside looking in and think it’s a very slow process, but I would frame it, that it’s a very deliberate process to ensure that new systems are very safe, and that they’re really examined and scrutinized to a high level of detail to ensure that that safety can be maintained across a wide variety of conditions. And it’s also the case as you pointed out earlier, that these are international systems. You know, for example, even traffic within the U.S. often flies over Canadian airspace, or if you go overseas, you want everything to work the same. So the Federal Aviation Administration has been coordinating with the International Civil Aviation Organization, ICAO, to implement ACASX. And there are a number of different variants. So ACASX started to improve collision avoidance for transport category aircraft, the aircraft that you know, you’re listening public flies on every day, and that system was finished in 2018. It improves safety by over 20%, while reducing nuisance alerts by over 60% that has been approved by the FAA. It was recently approved by ICAO for international use, the European rulemaking is a little bit longer than U.S. rulemaking. So it’s still being finalized within Europe. But that system should be flying very shortly. We also developed a system for the larger uncrewed aircraft. Think about Global Hawk predator size aircraft, that was finished in 2020. And we are currently working within the International Civil Aviation Organization to get acceptance of that safety standard to allow worldwide operations. So that is in progress, it should be approved in the next six to 12 months. We just finished the standard for the system for smaller uncrewed aircraft or drones, if you will to facilitate commercial operations. The FAA will need to implement that guidance this small drone community is things are evolving rapidly in terms of what operations are intended to take place. The companies are developing concepts, but this capability would be usable. We are currently also working with some other regional partners like the Massachusetts Department of Transportation, to integrate ACAS SXU for the small unmanned aircraft into their operations. And finally, we’re currently in the middle of developing a variant for we call it ACAS XR for Rotorcraft to include the advanced and urban air mobility aircraft. Right now that it’s scheduled to be completed in 2025. And I would just mention that this is not MIT Lincoln Laboratory doing this, we are sponsored, you know, really under the visionary leadership, in this case of the Federal Aviation Administration, who has funded this work for over a decade and is strongly championed in the community. And we’ve had strong participation from the aircraft manufacturers, the avionics manufacturers, the pilot unions, the air traffic control unions, to make sure that this is an acceptable system. So we should see that system for advanced air mobility and time for their operations. And that should be finished in around the 2025.
Tom Temin: And just a detail here, some aircraft that is uncrewed is nevertheless piloted, but remotely, others really are drones are operating more or less autonomously. Big difference, correct?
Wes Olson: That is correct. There is a big difference, almost all of the uncrewed aircraft right now are operated by a pilot. And that’s a requirement, there needs to be a human in the loop to oversee safety and be legally responsible for the flight. These more autonomous operations where no human is involved are in the future, and the Federal Aviation Administration, and the aviation community is working to understand how to do that safely. But all of these operations currently do involve a human, whether they’re in the vehicle or on the ground. And that could be fairly remote.
Tom Temin: And the final question, does the next gen air traffic control system and I’m not positive of the total status of that rollout? It seems like it was kind of stuck in two zones for a few years, does that figure into the deployment of this system of the avoidance?
Wes Olson: So this system was designed to be compatible with the next gen concepts, and many of these have rolled out over time, it takes a long time to implement something in the aviation system. Unlike a computer upgrade, we can’t just shut down the airspace for a week and switch everybody over to a new procedure. It has to be rolled in incrementally, which is a challenge for many of the systems to accommodate aircraft that aren’t equipped with new technology and those that are, but the compatibility is very important. We don’t want to implement, go to the effort of implementing new procedures and then find out that they’re not viable because other safety critical systems like collision avoidance fail too often, or don’t support those operations. So in ACASX, we’ve taken into account all of the future operations that we know about, but designed it to be flexible to provide variable learning based on a specific type of operation. So as new operations are implemented, and the objectives about the target level of safety or the reduced separation can be defined, we can tune that system using our machine learning algorithms relatively quickly, to adapt and provide operations to support those new operations in an effective manner.
Tom Temin: Wes Olson is group leader of the surveillance systems group at the MIT Lincoln Laboratory. Thanks so much for joining me.
Wes Olson: Thank you very much, Tom. It’s been my pleasure