What if, during a battle at sea, the commanding officer on a ship could tell whether a key officer, such as the TAO, is crumbling under stress—and may start making bad decisions—based on that officer’s heart rate, blood pressure, galvanic skin response and other stress measures?
While wearable devices are commonly used to improve performance in sports, the joint forces may soon have the opportunity to use wearables—along with AI—to determine whether officers and others are both mentally and physically sharp in critical situations.
Because this information can be presented to the commanding officer on a laptop, he or she could, from anywhere on a ship, tell whether the CIC watch officer, for example, was in danger of making a serious mental error in battle, and perhaps would need to be quickly replaced.
Wearable data and AI can aid in peacetime as well, helping to make sure officers can stay focused on preventing potential disasters, such as collisions, fires, flooding, and man overboard—and can quickly make the right decisions if such a situation occurs.
How Wearables Measure Stress
The latest wearable devices—including watches, chest straps, rings, headbands and earpieces—can generate a host of metrics to show a person’s stress levels. For example, some devices measure a person’s heart rate variability—the time between each heartbeat—which fluctuates during the day. Heart rate variability can show whether a person’s nervous system is in fight-or-flight mode, which indicates stress, or is leaning more toward recovery and healing.
Other devices estimate stress levels by measuring galvanic skin response, which can indicate when sweat glands are triggered by emotions—even in small ways we may not be aware of. These and other metrics, such as resting heart rate and blood pressure, are combined to create a full picture of how well a person is coping with stress.
The Role of Training
The key question, of course, is not whether a person is stressed—which could be the case with anyone in battle—but whether the stress is interfering with his or her ability to perform mission tasks, and could lead to poor decisions. There are several steps to determining this.
It begins during training. Outfitted with wearables, officers and others go through various drills that mimic battle conditions. As trainers add stressors—such as unpredictable complications and increased tempo—they can monitor how well individuals perform as their stress levels increase.
Data from the wearables might indicate, for example, that individuals experiencing heightened stress have slower reaction times, or less working memory, or perhaps mental tunnel vision, in which they’re focusing on a single threat or goal without seeing the larger picture. All these can lead to poor decisions.
This approach has another benefit, helping to pinpoint whether a person is making mistakes because of stress, or because he or she needs more training. This might be revealed, for example, when an individual is making mistakes during training, but is showing no signs of increased stress.
Bringing in AI
By correlating stress levels and decision-making during intense training, defense organizations can begin to predict how well an individual will perform in real-world conditions. But training alone can’t tell the whole story—it is unlikely to show whether a person can perform every possible mission task at every possible stress level. This is where AI comes in.
Machine learning, a form of AI, has the ability to find patterns in large datasets. The first step is to use machine learning to find patterns in how well an individual performed different tasks at various stress levels during training. Next, defense organizations can bring together the data from large numbers of people who were monitored for stress during training—and look for those individ- uals who showed the same patterns. With enough such individuals, most if not all possible combinations of stress levels and mission tasks will likely be covered. This provides a greater ability to predict how well an individual will perform a particular task at a particular stress level—even if he or she was not in that exact situation during training.
Wearables in Battle
Here’s how this might apply in an actual battle: An officer of the deck, for example, is outfitted with wearables. Data from the wearables show that the officer’s stress levels are skyrocketing. Based on the training data from the officer as well as the relevant individuals in the larger group, a machine learning model might predict that the officer can still handle some critical mission tasks, but is at risk of making serious mistakes with one or two others.
At the same time, the machine learning model can track how an officer’s ability to perform a task is rapidly changing as his stress is increasing. For example, the data might show that a few minutes ago, the officer was doing fine, but now his decision-making ability is suddenly deteriorating.
Such information—on key officers throughout the ship—can be conveyed instantly to the commanding officer and the executive officer through dashboards on their laptops or tablets, enabling them to take timely action.
The information can also be sent to the officers themselves, so that they can try to bring down their stress levels through techniques they learned during training. An officer might do some quick deep-breathing exercises, for example, or he might recall a time when he performed well in a high-stress training situation—giving him confidence he can do it now.
While the technology for such an approach is currently available, several obstacles would need to be overcome to make it feasible. For example, policies would need to be changed to allow TAOs and other officers to use wearables in secure spaces. Infrastructure changes would be needed as well, such as sensors that would enable data to be transmitted across decks and compartments. Ships would also need edge computing with AI to collect and analyze the data.
With this approach, data from wearables is kept private and secure—it is deidentified until it reaches the commanding officer or other authorized person. That way, if the data is intercepted, it can’t be connected with a specific individual.
In addition to its use in wartime, data from wearables can also be valuable in peacetime situations where there is little stress. Data can show, for example, whether officers or others aren’t getting enough regular sleep, or aren’t drinking enough water, or for other reasons may not be mentally sharp and might make mistakes that could endanger the ship or its crew.
Irik Johnson ([email protected]) integrates wearable technologies, data science and virtual reality to improve training and performance for Booz Allen’s DoD clients. As an expert on sports science, he has optimized athletic programs for the NFL, NBA and MLB.
Commander Alan Kolackovsky ([email protected]) is a retired Naval Limited Duty Officer, whose assignments
included Executive Officer NIWC PAC. He leads Booz Allen’s 5G/ CBRS infrastructure deployment, delivering emerging technical solutions including unmanned systems capabilities.
Ken Kryszyn ([email protected]) is a retired Navy Force ISSM who oversaw cybersecurity for all surface ships in the Pacific Fleet. Ken is now a senior lead technologist at Booz Allen, where he has been delivering cybersecurity engineering and RMF automation to the Navy for 20 years.
Booz Allen subject matter experts Commander Jarrod (JRod) Groves,U.S. Navy (Retired) and Maggie Corry contributed to this article.