Resumo

CARBO PEREZ, Alain J.; SALVADOR FIGUEROA, Elizabeth  e  LOPEZ DELIS, Alberto. Evaluation of an algorithm for threshold-based fall detection from inertial signals. EAC [online]. 2023, vol.44, n.1, pp. 58-65.  Epub 06-Dez-2023. ISSN 1815-5928.

The aging of the population and the significant health problem represented by falls in the elderly population, promotes a growth in the limitations of the activities of daily living (ADL) of this sector of the population. The development of the medical internet of things, wireless sensors and inertial measurement units (IMU) make it possible to directly and optimally sense patient data in their natural environment without invading their privacy. The purpose of this study is to verify if, through a threshold-based fall detection algorithm, it is possible to accurately distinguish between older people, falls, and the development of ADL, based on the analysis of the linear acceleration and angular velocity variables obtained of the UMI sensor. This information is used to calculate the magnitude of the vector sum (SVM) of the acceleration and angular velocity, as well as the pitch and roll angles. Based on this information, a pilot version of an experimental protocol for the detection of falls in healthy subjects was carried out at the Center of Medical Biophysics. The performance of the algorithm was evaluated by implementing it in the public database CGU-BES Dataset, obtaining a sensitivity of 91.6%, a specificity of 88.3% and an accuracy of 89.4%.

Palavras-chave : fall detection; inertial measurement unit; wireless sensors; threshold-based methods; activities of daily living.

        · resumo em Espanhol     · texto em Espanhol     · Espanhol ( pdf )