How the Air Force data strategy is evolving

Source Link

By: Valerie Insinna

The Air Force’s ambitious new intelligence, surveillance and reconnaissance strategy calls for a sensing grid that fuses together data from legacy platforms (such as the RQ-4 Global Hawk), emerging technologies (like swarming drones), other services’ platforms and publicly available information. Artificial intelligence will decipher that data. Such a system may sound like science fiction, but the service believes it could be up by 2028.

Lt. Gen. VeraLinn “Dash” Jamieson, the deputy chief of staff for ISR, explained the genesis for the Air Force’s new “Next Generation ISR Dominance Flight Plan,” which lays out the service’s goals for the next 10 years. She spoke recently with Valerie Insinna of sister publication Defense News.

C4ISRNET: First, why does the Air Force need a new ISR plan?

LT. GEN. “DASH” JAMIESON: This Flight Plan really does go out for 10 years. We did it because, primarily, we have a National Defense Strategy that was written, and crafted, and came out in January. It looks at a changing complex world with a great power competition. We see that the character of war is potentially changing based off of technologies that are being fielded and that are under development today. So that is one reason.

The other reason really predates the NDS. When I took over the A2, the chief of staff, Gen. [David] Goldfein, really looked at me and said, “Dash, your ISR Enterprise is very airman intensive.” So I took an evaluation with my team, and it is extremely airman intensive. The airmen are applying new things to old tools.

How we share the outcomes of our sensing capability is via PowerPoint that our airmen construct using Excel spreadsheets to look at the data, identify what is the data, and try to then manually layer the data in this construct.

To get at some fused data, to get at what are the trends, that approach is not going to give us the ability to actually conduct our operations at the speed of relevance across the entire spectrum of conflict. More importantly, it drags out our decision cycle for our war fighter. When you drag out your own decision cycle, the adversary has the ability to get inside of your decisions and to disrupt those decisions.

C4ISRNET: What can you do?

JAMIESON: Our intent is to actually get inside the adversaries’ decision cycle and create chaos. Once you do that, that really is a tough, tough problem to get out of. We established a framework and we have two major efforts.

One was how we integrate and balance our ISR portfolio. We take a look at what we have today, what we see are our seams and our capability gaps, and we determine how we make investments on that. But our other major effort is, “What is my future pathway, what are those lines of effort that are going to give us an advantage?”

We came up with three macro categories. First, it’s disrupt the technologies, and see what capabilities and options that brings you. The second is how we bolster lethality and readiness with what we’re going to do to the enterprise. Then, third, it’s how we establish foundational capabilities that transcend this entire framework.

C4ISRNET: How does that manifest itself?

JAMIESON: You have to have a data strategy because you actually have to have standards on how you are going to condition your data. How are you going the access your data? How is your data gonna move for your infrastructure? How are you gonna secure your data? How do you ensure your data is not up for malicious attack?

We did talk to industry. Then, we also worked with our acquisition professionals and said we really need to have an agile capability development concept annex. That really gets at how we prototype, how we do DevOps, what is the environment and how we get at acquisition of software in a very different paradigm.

C4ISRNET: What does that mean for the force?

JAMIESON: In 10 years, our digital airmen will be the preponderance of the force. Our airmen right now, a majority of the ones 24 and under, come in already knowing how to code. So what skill sets do we know come up for ISR where coding is fundamental? How does that affect our retention capability? Because we want to empower and unleash them to develop new skill sets that will complement where we’re going with the ISR enterprise.

Finally, we want to partner much deeper with our think tanks, our academia, our labs so that we are sharing right up front.

C4ISRNET: You’ve talked about a collaborative sensing grid that uses advanced technology. What does that look like?

JAMIESON: No longer are we going to invest primarily in just the air domain. We’re gonna look at capabilities in and from space.

We’re also working with our joint partners to integrate in surface and subsurface capabilities, so that our sensing grid of tomorrow is no longer a sensor looking in a specific domain with a specific in. If we initially look at the first Predator, we had motion video that was EOIR, electro optical infrared. Then, maybe we need to have SAR. Then when we went to hyperspectral, we went to EO/IR and [synthetic aperture radar], because we’re getting multiple ins in a domain.

C4ISRNET: Is this a place to use AI?

JAMIESON: We want to have algorithms to get at ensuring that the data is pure and not malicious or false. But we are going to take that, and we’re going to use that with our exquisite capability to really flesh out that sensing grid. We’re going to do it so now that we now have resiliency, it’s not just in one domain or one capability.

If we look at high altitude, what can I do from a manned and an unmanned capability? From a persistent, standoff and stand-in capability? We’re going to look at swarming, we’re going to look at even hypersonic capability — that give us a multitude of new capabilities to form this fencing grid.

The beauty of the sensing grid is it doesn’t forget what we have already fielded. Our airmen are looking at why is something happening, what are the trends. In other words, we fielded a sensor to answer a question. What we’re trying to develop is how I get the data so that I can fuse it, look at it, then ask the right questions.

C4ISRNET: What goals do you have over the next 10 years to really bring that into the ISR infrastructure?

JAMIESON: The real importance of cloud computing ... and when I talk about cloud computing, I’m really talking about it as a service. The service that we’re trying to get is really a platform, infrastructure and software. I’m not just looking for a data-storage hub. I’m looking for a partnership with industry.

I’m not necessarily looking for just one industry partner; I’m looking for multiple industry partners in a multi-cloud concept, because each one of the big five, if we just were to use that, has their expertise. We want to be able to take that expertise and use it. When I say platform, infrastructure, software, I’m really talking about the capability that they give me to field at scale.

Every time you turn on your Tesla updates are instantaneous, and I want my ISR enterprise to have those type of instantaneous updates on the infrastructure so that I can continue to prototype my applications.

The data is what we’re going to use to develop those applications. We’ve told industry, “You can make applications and we’re going to buy applications from you. The data is ours. But we want to partner with you on the ability to create algorithms, applications, different software packages.”

C4ISRNET: You talked about private sector partnerships, but how do you overcome or address the fallout from Project Maven?

JAMIESON: We have been talking a lot about this. I don’t see this as a problem for the ISR enterprise, I don’t really see this as a problem for DoD. I see this as a U.S. public issue that needs to be debated and discussed. Because when we look at artificial intelligence and you look at how you develop algorithms, there’s always a bias when we put the math together. The bias is formed by the humans that are putting that math together.

What I mean by that is, the bias right now is whether we are going to use Western values. We value privacy, we value life, liberty and the pursuit of happiness. Our competitors, as we’ve seen, do not value privacy. Nor do they value life, liberty or the pursuit of happiness. They have a completely different construct.

I think the American public needs to have a discussion on where do we want to go. How do we want to approach this? What does it mean to us as a society? How are we going to protect our privacy? How are we going to protect our values? In lieu of that, then how does that apply to our national defense? We absolutely welcome that debate. We want to have that dialogue.

C4ISRNET: Do you envision capability upgrades for the Global Hawk or Reaper? Are we going to see certain platforms phase out, or new platforms built into the budget?

JAMIESON: As we look at where we are with our [remotely piloted aircraft, or RPA], U-2, our Rivet Joint capability now, it was pretty airmen intensive. What we are doing is developing algorithms to take the data off those platforms in a much faster cycle.

I don’t want to do processing, exploitation and dissemination in a reach-back mode in the future. I want to process, to exploit right on the aircraft or right on the sensor so that I can actually take that data, condition it, and then use it with other data so that I can get out better quality of information into the joint war fighter.

Think of the Reaper. Today, we take the data off. If it’s full-motion video, my airmen are actually identifying the object and looking at patterns of life after staring at that video for hours on end. What we’re gonna do is automate that entire process and that is what Maven is doing. But we are developing algorithms much faster.

In the next two to four years, it will be processed at the sensor. So that allows the airmen to no longer take the hours to do the processing. It takes a lot of bandwidth, it takes a lot of time. We want to do all that onboard the sensor so that I can fuse the data from the sensing grid. The reason I want to fuse the data from the sensing grid is because I want to identify certain characteristics.

C4ISRNET: Did the Flight Plan address capabilities at the platform level, such as whether new systems were needed? Or were you purely focused on exploiting data?

JAMIESON: We have to go back and identify first where are our big gaps, because we are in a cost-effective modernization way forward.

Do I look at what I need from a space capability? Do I look at what I need from a swarming RPA? Do I look at autonomous remotely piloted aircraft, whether they are high altitude or minis that go for a specific length? We want to have a balanced portfolio of standoff, penetrating and persistent capabilities.

C4ISRNET: Are you going to invest in swarming capabilities? Or some sort of hypersonic vehicle that can do ISR?

JAMIESON: It is in the mode of being thought about. But you have to look at what technologies are real today and what technologies are really going to be there for tomorrow. You don’t want to commit early to something that isn’t gonna give you the best payoff.

C4ISRNET: Are there any concrete goals that you guys are looking at to make sure the department is staying on track?

JAMIESON: I’m pretty direct, and I’m pretty blunt. In our classified ISR Flight Plan, I’ve got an implementer for every single annex with milestones, goals, objectives and pathways so that, starting in 2018, we have the deliverables to ensure that we do stay on path.

We don’t know what’s going to happen with the internet of things. That could be just as disruptive as the internet was.