mHealth Research Digest with Tim Bredrup


A variety of body sensor network systems have been proposed to detect health related problems in real time.

However the focus of many of these systems is on the gathering and presentation/storage of data, rather than on autonomous real-time decision making.

A study conducted at Coventry University in the UK focused on the implementation of on-body prediction system for reducing health risks caused by Uncompensable Heat Stress (UHS) such as hyperthermic exhaustion.

It targets the monitoring of Explosive Ordnance Disposal (EOD) operatives during missions through the gathering of physiological data such as multi-point skin temperature, as well as postural information (multi-point body acceleration).

According to previous research, real-time autonomous processing of skin temperature and body acceleration data are both required for the prediction of heat stress.​​ ​

Data from a total of 26 trials were used to evaluate the accuracy of the UHS prediction system. The study protocol was as follows:

…twelve subjects underwent a mission-like protocol while wearing the EOD suit at 40°C ambient temperature and three different in-suit cooling variations – no cooling, chest cooling, and head cooling. The trials consisted of four identical back-to-back cycles of: walking on a treadmill (3 mins), kneeling while moving weights (2 mins), crawling (2 mins), postural testing (2.5 mins), arm exercise while standing (3 mins), and cognitive tests while sitting (6 mins).

As reported by the study’s researchers,

“Posture classification (was) performed with an accuracy of 96.1%, and a heat stress prediction algorithm (was) demonstrated with an overall accuracy of 88.5% when predicting the occurrence of heat stress within the next 2 minutes. This demonstrates that the model is a usable predictor of whether the danger threshold will be exceeded.”

The evolution of human-centric monitoring systems toward machine driven learning, as demonstrated in this study, could present a range of benefits going forward for a variety of applications.

Authors: Ramona Rednic, John Kemp, Elena Gaura, James Brusey

Institution: Cogent Computing Applied Research Centre, Faculty of Engineering and Computing, Coventry University, UK

Original Abstract: PDF