Omar Morales and Jarron Lembke

Omar Morales: Good afternoon, everyone. My name is Omar Morales.

Jarron Lembke: And I'm Jarron Lembke. We both work for the 309 Software Engineering Group, otherwise known as 309 SWEG. 309 SWEG is one of three primary U.S. Air Force software groups. It employs over 1,900 personnel and is currently working around 90 projects. The primary operating location for 309 SWEG is Hill Air Force Base, although there are three other locations on other bases around the United States.

Some of the primary workloads that we work on, if you can see on the little images at Hill Air Force Base, we do F-16 and A-10 software development and integration testing, F-22 software integration testing, and we are standing up capabilities to do F-35 software development.

Omar Morales: So we're here to talk about Project FOX. It's a project that started up in 2019. We started having discussions on how we could build software, test software, deliver software better for the F-35. In spring of 2020 we had a discovery and framing event with the 309th, with Edwards Air Force Base personnel, and also people from the F-35 Joint Program Office, the JPO. There we identified a lot of those problems and challenges we were having with F-35, as well as some solutions for how we could get around that and do stuff better. We immediately started to work on an MVP, minimum viable product, to realize some of those solutions.

As we were working that, F-22 came along saying, hey, we want Project FOX working on F-22 as well. So that brought in also the F-22 Program Office and AFRL to join up with Project FOX, and that kind of makes up the core Project FOX team as it is today.

Jarron Lembke: As far as the problem statement that Project FOX is attempting to address, the National Defense Strategy calls for a rapid, iterative approach to capability development to reduce costs, technology obsolescence, and reduce acquisition risks. In contrast to that, the annual report to Congress on China states that China has achieved parity or even exceeded the United States in military modernization areas. Additionally, fighter programs have shown that they're unable to deliver capabilities in relevant timelines and costs.

In speaking to flight test pilots at Edwards Air Force Base, we learned that it takes years to bring a new capability to flight testers after it becomes a program of record. After that capability comes to flight test, it takes several months to fix bugs that they found during initial testing. In addition to taking several months to fix bugs, they're very limited in their ability to refine and improve the capability because of a rigid software lifecycle.

Omar Morales: That's the main goal of Project FOX: how do we speed up that process? How do we sidestep it? What in the world can we do to make this better? We're doing that with Project FOX, providing an open software enclave, kind of a sandbox, if you will, that we can connect to these systems, F-35 or F-22 and potentially others, and allow other third-party developers to come in and bring their applications, bring their technologies, and utilize the sensors and the systems on board these to deliver additional capability in a faster, more streamlined way.

This essentially removes that vendor lock. When our prime contractors make us defense systems and retain a lot of that control, and the program offices retain a lot of that control, it's really hard to break through and get your cool stuff on some of these systems. Project FOX is another avenue of how to do that.

We work fully with the program offices. We work fully with the prime contractors. They're supporting us in this effort. They're doing a lot of cool stuff they've already got planned out, doing projects. We plan to do some other cool stuff as well, just providing an additional avenue to bring capability to these jets.

Jarron Lembke: Some of the advantages that FOX brings: it utilizes ghost mode testing, which is non-interference testing where FOX can ride along with an existing flight test mission while still bringing flight test benefits. It also reduces software integration risk because it brings software to the pilot sooner. It also provides an opportunity for iterative improvements to reach a better product at the end. It allows for integration of new, emerging technologies such as AI or machine learning from groups that have already developed the capability. It also allows for crowdsourced testing, where testing doesn't have to be performed on a single platform or by a single test group.

Omar Morales: Use cases. We've identified three use cases for Project FOX. There's tactical, training, and test. As far as tactical, these are applications that would be beneficial in an operational or wartime environment, such as manned-unmanned teaming, which is the name for an application that gives a pilot the ability to easily control and interact with drones that are performing functions related to that pilot's mission. There's also cyber defense: advanced cybersecurity applications could be brought into the aircraft. Rapid hardware integration: we'll speak more on this later, but Project FOX is being used to integrate a helmet-mounted display into F-22.

In the training realm, we've already demonstrated the ability to do some live, virtual, constructive capabilities, where we're taking live data from the aircraft, we bring in our own synthetic data or tracks, entities, whatever they may be, and bring those in together and provide a training scenario environment for pilots to train with. What we bring in synthetically can be high fidelity or low fidelity. It's really whatever people can think up and what we need to bring in for the situation. Not only is it for live aircraft; also in labs, simulators, trainers, Project FOX can connect to all these different systems and provide an ecosystem of training capabilities for all those systems.

In the test world, that was kind of how we really started: how can we test these things better? Project FOX allows that little sandbox where we can bring in smaller pieces of software, maybe possibly still in development. We want a quick look, we want an early look at how something performs, maybe how it looks, maybe it's a new symbology. We want to get the pilot's feedback on what something looks like in an earlier state and bring the pilots further up in that development process.

We have to do real-time data analysis. We can do machine learning and help tune those models a little faster with real data, live real data in these test scenarios. We're also helping the test organizations test more efficiently. We have a Ready to Test app that we're developing with test engineers. We're also talking about writing a control room augmentation for all these test centers as well, where you can have a Project FOX unit there in the control room or even a mobile control room system using Project FOX capabilities.

This is what Project FOX looks like. We have three configurations that we're working toward. Right now, all the work we've done, all of our software, all applications reside on a tablet and operate on the tablet, and you walk out and plug that tablet into the aircraft and go from there. That's worked well, but for a lot of applications and a lot of algorithms, they need a lot more computing power than what is on a small form-factor tablet.

Option two is we're going to take our software and re-host it inside the weapon system itself. F-22 and F-35 have an open compute environment where they allow us to bring our software in and run it on those platforms. You'll still bring the tablet and you'll plug it in. The tablet is just the UI; the apps themselves, all the heavy lifting, will be done inside the weapon system.

For platforms that don't have an open compute environment or don't have enough computing environment, we're producing a FOX box, which is just an external mobile device that has CPU, GPU, some memory, that kind of thing. Pretty basic, still small form factor. The pilot can carry it out with his tablet and then just connect the whole system into the jet itself. All the heavy lift is done in the FOX box. The tablet again is just your UI or applications. We're working on both those now. We hope to demo those by the end of this year.

Jarron Lembke: This figure depicts the general concept of Project FOX in orange. You can see that there's an open software enclave that's partitioned within the aircraft's avionics. FOX software would receive proprietary aircraft data and translate it into a universal, platform-agnostic interface. Third-party applications will be given the boundary within the open software enclave where they can operate and receive platform-agnostic data, which makes it possible that that same application can be run on a variety of platforms that host Project FOX.

There are four key components that make up Project FOX. The first one is government-owned software architecture, which is designed to allow for the same application to work on different weapon systems, as I just spoke to. This provides cost savings because multiple platforms can share in the cost of the development of a capability, versus each platform developing or paying for the unique development of a single capability.

There's also a government-sanctioned software boundary. As you can see in the figure, FOX software, the open software enclave, is kept separate from safety-critical code. This would reduce airworthiness requirements and allows for user software updates. Third-party applications are given defined ingress and egress rules where communication is limited to just the data that's applicable to the purpose of that application. This allows for user approval for prototyping, testing, and the addition of new applications. It also allows for increased ability to update software in between flight tests.

Another component is open standards. The desire is that we have flexibility to move with market forces and not be tied down to a rigid system. We use a DOD-managed interface which is not tied to a specific platform or specific vendor. This will allow for more open competition and allow access to third-party developers.

The last component is cybersecurity. There is a lot of worry about allowing third-party applications onto a weapon system such as F-35 or F-22. To address that, we make sure that we do incorporate cybersecurity into Project FOX. Software is developed on DOD-approved pipelines, DOD-approved networks, with software that has a proven pedigree. Third-party developers can develop software within the boundaries and integrate it into these weapon systems.

Omar Morales: So how do we deliver it, how do we maintain it? Like a normal software DevSecOps pipeline, we develop locally, do all the testing we need to do, do all the security scanning and everything there, and then we push out new versions or new capabilities out into the cloud. Right now that is an F-35-provided cloud.

Eventually, once it gets into the cloud, we're going to create an app store where pilots or users or test engineers, whoever they may be, can just grab their tablet and download applications to run on a jet that day. That is what we're working toward. That's our target.

For us, with Project FOX, we're really trying to stay in front of the latest technologies, utilizing the newest technologies, the newest processes for developing software. In defense systems, a lot of times we build something and then we have to maintain it and upgrade it for decades and decades and decades. A lot of times when we build software, those tools and languages, those processes get stuck when they got developed. Maybe there was one upgrade instance over decades and we get stuck using these old processes and these old tools. With Project FOX, we don't want to be that way. We don't do that crap. We're always looking to improve ourselves. We're always looking for new technologies. We're already re-architecting our architecture for the fourth or fifth time in a couple years. We're always trying to learn. We're always trying to do better.

For us it's not just DevOps, it's DevSecOps. We need that security piece built in there. Make sure our software is secure and safe to run on these weapon systems. We have technologies built in as we develop our software to handle that. All that together, our approved tools, DOD-approved tools and processes running in the approved environment, the F-35-provided cloud, we're granted a continuous authority to operate, meaning our software is always ready to be built and delivered and it's always got the stamp of approval. It's safe to run on an F-35 system.

Jarron Lembke: 309 SWEG started Project FOX back in March of 2020. We defined a minimum viable product, or an MVP, that included a back end that received and parsed F-35 air-to-ground track data that came from F-35 UDP messages. It also included a front end that displayed that track information.

This MVP was expected to be completed in six months, but it took one year to receive approval to connect a tablet to the F-35. In March of 2021, we successfully received, parsed, and displayed F-35 data on a Panasonic Toughpad.

Omar Morales: So next, we had F-22. They came in a year later at the end of 2021. We were having discussions to legitimately put Project FOX on F-22 around some capabilities there. At the beginning of 2022, we were given the green light to go ahead and develop that platform adapter for F-22 and do a demo on an F-22 system.

We built that software in April. We went out to Edwards Air Force Base, took our tablet, plugged into an F-22 in a hangar, ran our software, and within one hour we had our software working 100%. This was a huge success for us.

Not only was it a new thing for F-22, for us to do an F-22, we had never plugged into an F-22 before. We didn't have a lab. We didn't have a simulator. We were given one example of one instance of one message, and the documentation was almost good enough to help us. We went out there, high-risk event, and our team nailed it. We had two bugs we were able to fix while we were connected to the aircraft. In an hour, all of our stuff was working 100%. That was a big confidence boost for us and for the whole team to show we have a good quality product, and also our processes and tools and the direction we're going are also solid.

Several months later, in August this year, we did the flight test. So the same software, essentially the same demo, but now we gave it to a pilot to plug into the F-22 and go fly around and play with it. It was also a very successful test event for us.

We tested three applications during that time. One of those applications was the same application that we tested on the F-35 MVP, just showing that with no code changes needed, an application could run on F-35 and could run on F-22. The third party didn't have to do anything, just that platform adapter swap it in and out and the back end handled all the changes.

Another application we tested on that flight test didn't quite work 100%. We flew it on that first flight and it kind of worked but it kind of didn't. So we used that record-replay feature that we have built into the back end. We had the pilot record some data while he was flying around. We landed, we took the tablet back to the back room, replayed the data, debugged, made some fixes. We were ready to fly the next morning, where we successfully demonstrated that application in flight.

Normally for defense systems, that fly-fix-fly process takes weeks, possibly even months, to do that whole turnaround and test a fix. If you find a fix and you get it back to the contractor, they will fix that and rebuild it for you in a month, you would be happy. That was a good job, doing it in a month. Well, with Project FOX we did that in another day. So that was a big win for us and the whole Project FOX team.

Jarron Lembke: What are our current efforts? We're currently working really hard to develop a FOX box that will be connected to a maintenance port in the F-35 cockpit. Following that effort, we're going to transition or re-host FOX software onto the F-35's open compute environment. In a very similar effort with F-22, we're transitioning from running FOX software exclusively on a tablet on that F-22 to re-hosting that software to run within its open compute environment.

The source selection for a helmet-mounted display for F-22 has begun. As far as future efforts, AFWERX is working with eVTOL companies and helping them come to market. eVTOL stands for electric vertical takeoff and landing aircraft. The United States has interest in supporting these companies for economic and for national security reasons. And so we are working with AFWERX in integrating Project FOX into some of these eVTOL aircraft.

We're also having discussions with A-10, which is very seriously looking into integrating Project FOX onto its platform. Last, we are working with several companies in developing a variety of third-party applications.

Omar Morales: Here's some example sample set of applications we're using on Project FOX. The middle one, the big dark table one, is what we used on the F-35 MVP. It was just to show, yeah, we could plug into an F-35, we can read sensor and system data and just display it in a usable format to the pilot.

The one in the middle there in the back, that's the one built by Lockheed Martin ADP that we ran on F-35 and F-22. It's kind of a route-planning application. It worked well for us there. Other examples of applications that 309 is working on are currently in a design-development phase.

The list of organizations on the side are some of the organizations we're working with to bring their applications and algorithms and host them on Project FOX on these fighter jet platforms. You'll notice, yes, we have the large defense contractors working with us, providing new applications to run on Project FOX. We also have small companies who also have some really cool tech and some really cool ideas, really cool software, where before it was kind of hard to break into that program office, contractor, acquisition cycle. Project FOX gives them an avenue to bring in their cool software and host it on a weapon system.

As the third-party integrators, 309th, we help them do that. As I mentioned before, we're writing an SDK and supporting and helping them bridge that gap from their applications into a military system. Then the government labs, 309th SWEG, Air Force Research Laboratory: really Project FOX is open. We're opening up these weapon systems to anybody who has really cool stuff they want to bring on these fighter jets. We can add it. We can now do that.

Jarron Lembke: We're really excited to be helping F-22 integrate a helmet-mounted display. F-22 is the only DOD fighter platform without a helmet-mounted display. Back in the early days of F-22, this was less of a problem because it was the air-dominant platform of the Air Force. There weren't many aircraft that could come close to it. But nowadays, with more parity among advanced fighters from other countries, it's more of a problem. So pilots have a need to be able to look out at an aircraft and, via their visor, cue a missile to objects that they're looking at.

The F-22 Program Office has attempted to integrate a helmet-mounted display into the F-22 three different times, but it has not been able to because of cost. FOX is providing an avenue that's cost-effective enough that they actually could accomplish this. There's a National Guard Squadron at Hickam Air Force Base that is providing funding to outfit their squadron with helmets. We expect this effort would take us six months of development time.

Omar Morales: The challenges we're facing today and in the future: number one, two-way communication. Right now all of our demos and tests, everything we've done, is one way, meaning we're just reading stuff off of these sensors and off these systems and making use with them. We can do some cool stuff with that, but to realize the full vision of Project FOX, we really need that two-way where we can send data, tracks, whatever information into these weapon systems, have them react appropriately, display something appropriately, et cetera. We've had conversations with contractors and the program offices. It was not in their heads that we're going to do this, but we expect some hurdles to be able to do that.

Authority to operate: I mentioned we have that continuous authority to operate because of how we're developing our software, which is great. Unfortunately, that only applies to our software. When we bring other software, we bring hardware in, now we need additional approvals from the program offices to bring that all together, to connect it into a fighter jet, to fly around in the air, and do all sorts of stuff. That's really been a long pull for us to get those together. It took us a year on the first F-35. It's complicated and challenging.

Multi-level, cross-platform classification: we want Project FOX to be ubiquitous from unclassified commercial platforms like eVTOL to a top-secret platform, and across different classified systems, F-35 to F-22. How do we do that? How do we create software that will run across all those environments? Where do we develop that software? How do we deploy it to all those classification levels? That is something we still need to do.

With our hardware, we've got the FOX box. We're trying to get through some of the hardware limitations on platforms. And then our app store: we want people to be able to download applications anywhere in the world they are, whether a test center here at Nellis, an operational squadron in Europe. They can go on the cloud, download an application, put it on their tablet, run it that day, and everything works awesome.

That's Project FOX in a nutshell. It's a blue ocean capability, so we see it as a lot of interest both on platforms and other organizations bringing their apps. Thank you.