Application of near real time artificial intelligence on soldiers' helmet for military training and safety

Abstract

The citizens of Singapore are subjected to hot and humid weather throughout the year. Soldiers are put in danger while undergoing military training as a result of this. Road marches, in which soldiers are expected to don a minimum of Standard Battle Order (SBO), are one of the most common military training exercises. The attire of SBO further reduces the heat dissipation of the body temperature to the environment, raising the risk of Exertional Heat Stroke (EHS). Therefore, this study is needed to introduce the usage of wearable heat stroke detection devices (WHDDs), which detect EHS and alert users with the necessary alarms and protection against it. The study was divided into two parts: determining helmet state and designing of WHDD. To determine whether the user is wearing the helmet or not, the author used a Logistic Regression (LR) model to predict it. The ability of the LR model to identify helmet state in a realtime application has been demonstrated in this study, where the highest mean goodness-of-fit and efficacy are 96% and 73% respectively. Meanwhile, individuals' EHS risk levels were calculated using Fuzzy Logic Inference. The model depicts the relationship between the quantity of vigorous activity performed by an individual and the calculated EHS risk level. Overall, this study demonstrates that the required notifications provide early warning to users, preventing overexertion of the individual body.