Part 1: Upending the Defense Training Enterprise’s Data Sector

As noted in Halldale’s editorial departments for defense, civil aviation and safety critical industries, training enterprises have developed an insatiable appetite for data.

This is the first article of a two-part series that provides a deep dive on this learning enabler from the perspective of one of the sectors – the military. Insights for this piece were gained from one well-situated S&T organization, Orlando-based Naval Air Warfare Center Training Systems Division (NAWCTSD). The second, follow-on article to be posted the week of 3 February 2025 will provide perspectives on data from two S&T industry companies.

Data for an Increasingly Rigorous Training Environment

Military training is increasingly and inherently different from enterprises in adjacent high-risk sectors. In the case of the US Navy, its staffs, units and individuals are expanding their train-as-they-will operate aperture to an integrated battlespace model in a team-of-teams environment with ever-more sophisticated systems, networks and potential adversaries. At the same time, the sea service continues to push training to the point of demand – from ashore sites out to the fleet and beyond. The implications from these and other trends for using data to improve learner and instructor performance are huge. The Navy has moved beyond collecting every data byte generated in a training or education event to a more rigorous model, one with more precise and different requirements for industry S&T materiel and services suppliers.

It’s a Complex Data Environment

Dr. Heather Priest, Senior Scientific Technical Manager, LVC Training Solutions at NAWCTSD, built the case for the importance of data in Navy learning (training and education) when she observed, in part, “There is no Live Virtual Constructive training unless you can exchange the data needed to train,” and continued, “Data is key right now. Data is also necessary for everything we do in AI.”

And while data is important, it takes on different, evolving forms for the services.

In one case, the US Navy has a systems-of-systems construct on its aircraft, ships and submarines and shore-based operational nodes – all with diverse, embedded systems including sensors, data links, weapons, command and control and many, many others, and all producing data that is in play for operations and training. And then there are the layers upon layers of data generated by a training device and its embedded models, for instance, for weapons, the environment, constructive simulations and others. NAWCTSD’s Priest explained, “There is data that is involved in each of those components that all have to be brought together and communicated in order to create this ‘training environment.’” The research psychologist then emphasized key words that are important to understanding these relationships: standards; interoperability and security. “The big message is that each system has its own data – sometimes its own protocol, sometimes it’s proprietary – but all of this data must work together and simulate what the real world looks like for the warfighter and his or her team. And that is within one enclosed system, as you can have multiple simulators, which may all be ‘speaking’ the same protocols and language, but there are still sensor models, weapons models and others that all have to be integrated within that framework and be able to communicate across the network or wire to facilitate training.”

operating in that environment or doing what they do in the real world, but that we’re able to facilitate that training.” The migration from a training-to-operate construct to emphasizing more of the building blocks for training, provides the opportunity for more detailed discussions on: scenario; operational and training objectives; what is the person doing that is aligned to proficiency; and how they perform so that the mentor can make certain they are getting the right experiences to improve and get more proficient. “That means we measure them, look at that data and track that data – and that gets very complicated.”

Moving to support a more focused training construct also includes enablers beyond simulator-generated data, for instance, decisions, coordination, communications and other efforts “that are much harder as we need to make certain they are happening within the LVC environment and they have the right data sources to make sure we are capturing and tracking them, and looking at what the proficiency level of our warfighter is.” And that outcome, in turn, requires –you guessed it – even more data and data sources. Priest explained, “You have other data that needs to be considered and integrated into the environment that looks at not just the system, but the warfighter –for example, we have SHARP [Sierra Hotel Aviation Reporting Program] data in the Navy which has the readiness and experience levels of the warfighter, and communications in and between aircraft that are not easily measured. There are also things that are being looked at that are not based on system data. Things that we need to be able to collect and exchange that gives us knowledge of what the warfighter is doing, and why they are doing it and how they are doing it. This is the diagnostic piece of training – some are physiological and others are more on the subjective or typically on the observational side that we must explore potential data sources to facilitate assessment and training.”

The Navy also provides an example of efforts to more fully consider training data in terms of engineering-based business models.

Changing Data Management Model

Data collection is also an expensive proposition requiring additional decisions by program managers, according to Matt Adams, Lead Data Scientist, NAWCTSD. The command branch head initially observed a manger needs to make decisions on what he or she is going to collect and what is useful. “It is not practical to collect and store everything whether it is proprietary or available or not. And then there is the decision how do we make this useful for training, and that transfers over to additional work, where we will work with the operational side as they are developing new technologies, including autonomous vehicles, UAS swarming strategies and others. We can use that data to help understand how these technologies will be used, which will then help us understand how we’ll train the people operating them.”

have to pick, for financial and other reasons: which data is the most important; what do we want to collect and tag; and how do we want to store that data while protecting PII [personally identifiable information]. We have to make all those decisions and partner with industry to make certain we are collecting the right types of data and not just saying, ‘We want everything!’”

And here’s another trend impacting heretofore static data usage models – moving training from the traditional schoolhouse to forward-operating and other operational sites. Adams provided insights on the importance of data to support this development. “We’re trying to push the training from the schoolhouse and meet the sailors where they are and get that training out there. That also increases the challenge of data collection and more advanced data analysis technologies. This is a shift from, ‘You are going to do all of your training here in the schoolhouse.’ We now have training happening on a carrier or destroyer, and bases all across the world. And now you have to coordinate teams-of-teams training across all those systems. How you collect that data and provide quality training in this more distributed environment required considerable planning.”

The Right Data

There are significant changes occurring in the data sector for military training. Yet, NAWCTSD’s Priest said there is “good news” from the perspective that there is a coordinated focus to advance the science of data among different stakeholders at the operational and training levels at US DoD, from the department’s allies and partners, and industry. “There is a significant effort that recognizes the importance of data, the coordination that needs to happen in different areas – AI and Live Virtual Constructive environment – among data research scientists, research psychologists and others. The good news is as an organization we have recognized that we do need to get ahead of that and give industry the guidance that they need to develop these capabilities in a way that is driven by our requirements as the Department of Defense. This will help us focus on what our warfighters need to actually train.” Priest concluded, “There are parallel efforts to use the data we have now, not just to collect all the data, but the right data – the data that will be important to DoD and the fleet now and in the future.”

Future Topic on Data for Military Training

Halldale remains a conduit between the simulation and training industry and its customers in high-risk training markets to address community challenges and opportunities.

To that point, this author looks forward to providing two industry insights on military data in the second part of this article.