W.W. Pettine, M. Christenson, P. Koirala
Mountain Biometrics, Utah, United States
Keywords: Sepsis, blood loss, AI, wearable device, near-peer
In a near-peer conflict, warfighters may be in the field without support for days to weeks. This is especially problematic for managing casualties. In such scenarios, rationing limited quantities of blood or antibiotics requires predicting the individual progression of blood loss, or the likelihood of sepsis. While military medics are highly skilled at initial treatment of traumatic injuries, they are not trained to track the health trends of wounded warfighters over extended periods. Blood loss and sepsis produce typified patterns in heart rate, blood oxygen and other biometrics commonly collected from wearable devices. These patterns are individual-specific and emerge over hours to days. Recent advances in machine learning methods are only now making detecting such patterns possible. We are developing a generalized artificial intelligence (AI) for interpreting individual-specific, long-timescale biophysical data from wearable devices. While the AI can be used to detect and track a wide variety of health conditions, we are particularly focused on managing warfighter casualties in the field. The AI integrates into commonly used hardware and software, acting as a “force multiplier,” for existing equipment. Our AI will aid medics manage field casualties, saving lives when warfighters are cut off from advanced medical care.