A wearable sensor and framework for accurate remote monitoring of human motion

Abstract

Remote monitoring and evaluation of human motion during daily life require accurate extraction of kinematic quantities of body segments. Current approaches use inertial sensors that require numerical time differentiation to access the angular acceleration vector, a mathematical operation that greatly increases noise in the acceleration value. Here we introduce a wearable sensor that utilises a spatially defined cluster of inertial measurement units on a rigid base for directly measuring the angular acceleration vector. For this reason, we used computational modelling and experimental data to demonstrate that our new sensor configuration improves the accuracy of tracking angular acceleration vectors. We confirmed the feasibility of tracking human movement by automatic assessment of experimental fall initiation and balance recovery responses. The sensor therefore presents an opportunity to pioneer reliable assessment of human movement and balance in daily life.