Search for: All records

In military operations, real-time monitoring of soldiers’ health is essential for ensuring mission success and safeguarding personnel, yet such systems face challenges related to accuracy, security, and resource efficiency. This research addresses the critical need for secure, real-time monitoring of soldier vitals in the field, where operational security and performance are paramount. The paper focuses on implementing a machine-learning-based system capable of predicting the health states of soldiers using vitals such as heart rate (HR), respiratory rate (RESP), pulse, and oxygen saturation SpO2. A comprehensive pipeline was developed, including data preprocessing, the addition of noise, and model evaluation, to identify the best-performing machine learning algorithm. The system was tested through simulations to ensure real-time inference on real-life data, with reliable and accurate predictions demonstrated in dynamic environments. The gradient boosting model was selected due to its high accuracy, robustness to noise, and ability to handle complex feature interactions efficiently. Additionally, a lightweight cryptographic security system with a 16-byte key was integrated to protect sensitive health and location data during transmission. The results validate the feasibility of deploying such a system in resource-constrained field conditions while maintaining data confidentiality and operational security.