Concussive and subconcussive head impacts during athletics have received a significant amount of attention in regard to their possible contribution to chronic neurological effects. Numerous studies have adopted inertial body sensors to collect data to discover the relationship between head impacts and these brain injuries. However, the sensor data provided by the inertial body sensors is confounded with many artifacts because of the chaotic motion of the athletes. Previous studies developed simple metrics (peak detection, manual threshold, and change in head acceleration) to identify the real head impact data, but the supervised identification process requires a substantial amount of manpower and time to achieve.
The objective of the project is to develop an efficient, robust and unsupervised method to identify the real head impact data from the inertial body sensors, and ultimately, wirelessly provide the real-time feedback of head impact to the athletes.