Fall Prediction Based on Head-Mounted IMU Sensor System

Le résumé fourni par la source

Owing to the danger of falls, many fall detection systems have been proposed. However, fall detection cannot avoid the injuries and discover the falls in advance. This work proposes the algorithm of the fall prediction, which alarms before the impact from the fall. In order to identify the forecast of the fall, this algorithm, based on the preset threshold, analyzes the inertial information from the accelerometer and the gyroscope. In the experiment, 5 subjects carried out the simulated falls and the daily activities, which were predicted correctly by the proposed algorithm with the sensitivity of 81.4% and accuracy of 79.4%. Moreover, the time interval between the alarm and the impact was 0.287 seconds in average from the experiment.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.