ÓÐÁϺÐ×ÓAPP

I'm Joanna Guy, lead associate in ÓÐÁϺÐ×ÓAPP's global defense sector focused on accelerated readiness. And this is Master the Mission with Wes Haga. Wes is ÓÐÁϺÐ×ÓAPP's senior leader for C6ISR in the Pacific. Prior to joining ÓÐÁϺÐ×ÓAPP, Wes had a combined service career of 35 years in the Air Force and intelligence community. Wes served as a chief of mission applications and infrastructure at Air Force Research Laboratory and then as the chief scientist of data and infrastructure at the Defense Intelligence Agency. Wes, thank you for joining me. In your role with the Air Force A2, you were involved with the development of the first Tactical Assault Kit. What is the Tactical Assault Kit (or TAK) and what was that experience like? So the Tactical Assault Kit it is basically a small Android device that has been reloaded with a bunch of DOD software that provides near real-time or real-time situational awareness to war fighters in disconnected and connected operations. They're getting their feeds, they're getting their information from the TAK or the ATAK, and the ATAK means it's an Android Tactical Assault Kit with an Android operating system. My role with the A2 was actually at the 497th Intel Group, and that's where we actually provided the requirements to the Air Force Research Lab up in Rome, New York, to where that they could actually turn those operational mission requirements into actual capabilities that could then be fielded. The Army actually took those fielded capabilities from the Air Force Research Lab and turned that into a program of record, and it is common space throughout our warfighting community today. How is the attack still relevant on the battlefield today and what is ÓÐÁϺÐ×ÓAPP's role now in furthering the development of that technology? Well, the relevance continues to become greater and greater, as our need to understand what's going on around us becomes more and more in a highly complex and congested environment, so we, our soldiers, now have the ability to get more and more feeds with greater and greater fidelity, accuracy, and timeliness to be able to make decisions. ÓÐÁϺÐ×ÓAPP? We just acquired Par. Par was the original creators of the TAK from the Air Force Research Lab. That was the contractor underneath the Air Force Research Lab. We acquired Par Government, and that whole team has actually joined our digital battlespace team and our aerospace team up in Rome, New York. So it's extremely exciting. Why is ÓÐÁϺÐ×ÓAPP uniquely positioned to deliver acceleration of these technologies and to improve C6ISR and networking systems? I think what makes us unique is the way that we organize ÓÐÁϺÐ×ÓAPP and the way that inside of our DNA in ÓÐÁϺÐ×ÓAPP is to be able to go find the problem and solve it. For example, in a live fly operation that we just did in upstate New York, we kind of got after a problem in ÓÐÁϺÐ×ÓAPP called Fusion and we wanted to reduce the timeline to do process exploitation, dissemination of an intelligence product from 24 hours down to less than 5 minutes. So we got after that problem, and we actually got with a teammate, we got an airborne asset with a bunch of collectors on it. We created a mothership environment. We put some automatic target recognition algorithms that were pre-positioned to the battlefield that we thought we understood. Then when the battlefield changed, we were able to sense those changes through the mothership. Process those on a high-performance computer airborne, send that to a ground computer, retrain the algorithms, and put them back up in flight to be able to identify the changes in the battlefield. And then we were able to take a full motion video and a frame, a full motion video, and then an ELINT or a COMMINT hit, which is a form of an ellipse, and burn that ellipse right onto that frame and then submit that, transmit that into the existing comms infrastructure in INDOPACOM. We transmitted that to an ATAK to where now when you're collecting information in theater, you have to get it back to the national agencies. And that's doctrine. It's a beautiful thing. We did that in three weeks, and we proved that it was not under 5 minutes, it was under one minute that we can do the collection, do the fusion, do the dissemination. And I was actually down in SOFWeek when this happened, and we had another team in upstate New York, and I was actually watching the live fly from the console and seeing the fused intel on the screen down in SOFWeek in Florida. So it's pretty cool. That's incredible. Wes, thank you so much for sharing that. What was the time to decision before you implemented that solution and how much greater than 24 hours to actually do PED? And that's what our people are telling us today that are sitting with the clients and understanding the pain points that they have, and they need this information quicker. So we brought that to under a minute. Wow, that's incredibly impressive. And for DOD clients who want this technology implemented today, what is your advice to those leaders to get this into rapid development and out on the battlefield? But there are ways that the U.S. government has the ability to acquire these very rapid solutions at a low volume to be able to put them to work quickly. So organizations like your special operations community that have the desire to work these high-impact, low-volume type environments, I think they work with the research lab today. Why is it so important to have this technology implemented rapidly in today's landscape in the Pacific? It comes back to time. How do we make that time window go for our time window from decision, from understanding what's happening to a decision that a leader needs to make to counteract what's happening, or move something so they can get a better look at what's happening? Thank you so much, Wes. I appreciate your time today. ÓÐÁϺÐ×ÓAPP is at the intersection of mission technology, and your team is in the heart of the mission. It's inspiring to hear about your work and I look forward to seeing what you do next.

We are here today to do a limited objective experiment called Edge PED. PED stands for processing, exploitation and dissemination. We're taking a lot of the ÓÐÁϺÐ×ÓAPP capabilities that we have with Bighorn, our computer vision, our devs like OPS Pipeline and doing some data fusion along with some of our partners to provide intelligence targeting. We are starting with one of our partner teams. They have developed a software defined radio that can send decoy messages to an aeroscope. This is simulating drone swarms which would stimulate an adversary to react so that we would be able to collect on them. We are then collecting with our other partner, MAG Aerospace with their Sustenna Caravan, the MX-15 full motion video sensor. The MAG team is here providing the video data stream between the aircraft and the ground station here, where we're then feeding that data into the ÓÐÁϺÐ×ÓAPP model. It basically feeds across to our different teams with the Bighorn AI kit and they're able to train the information and we'll be able to pass it through to our video fusion team. We're taking in a live video stream from an aircraft and fusing it with multiple data sources and producing an output video stream that has information overlaying onto the video. That information is going to be ellipses that indicate the locations of signal intelligence that's gathered from the areas from other sensors and object detection using computer vision models to identify vehicles that are being viewed in the frame. As adversaries would modify vehicles, you would be able to retrain a model to not only recognize the old vehicle as well as the new modifications. Really we need to keep pace with our adversaries and a lot of that is about data and it's about collection. It's really about how do you win the information battle so that we can keep a free and open Indo Pacific. How do we use AI and sensor data to reach conclusions more rapidly and then make decisions. That is at the heart of so many of our government clients problems and being able to do this demo will help them realize that this is technology that can solve it for them today. It's proven that we can do it in a short time frame in the field, in realistic type conditions with internet going down, data feeds not working, really troubleshooting on the fly. They're really stress testing not only the capabilities of the technology, but our people. Everything just changes. There's different CODECS, different MIME types, streams break all the time. After this field exercise, when we actually make changes and updates to our our code, we're better prepared in making the technology more robust. Of course, with any live event, there is some debugging to be done. Being able to get some of that live data, see how we actually have to process that real time in order to be able to ingest that into the stack. But we've been able to work through that and get things trained pretty quickly all the way through to inference. The end of this is really proving that it can be done. But we've been able to develop this capability in three weeks, fusing the full motion video and signals intelligence to get data to reduce that PED life cycle, which normally takes about 24 hours down 2 minutes to help target for the operator to know what to look at. It has a lot of application moving forward. We're just scratching the surface right now.