Classification of Daily Activities Using an Intelligent Tip for Crutches

Abstract

Individualization of rehabilitation therapies for gait recovery is usually guided by subjective assessments based on the perception of the patient or the therapist. The use of sensorized devices can provide more objective data to adapt the therapy to each patient. In this work, an approach to recognise different daily activities is proposed based on the data captured by an instrumented tip for crutches. The developed approach is based on two steps: a pre-processing based on obtaining a set of statistical indicators, and an artificial neural network that process them. The proposed methodology has been tested in a group of 13 volunteers, allowing to recognize three critical daily activities (walking, standing still and going up/down stairs) with a success rate of 88%.