Recognition of human fall events based on single tri-axial gyroscope

摘要

Falls are a critical public health issue that requires continuous monitoring, especially for the elders. This paper proposed a method based on a tri-axial gyroscope for fall events recognition. A tri-axial gyroscope is placed at the user's waist to collect tri-axial angular velocity information. In order to facilitate data processing and extract features, real-time data are divided into a set of consecutive and partially overlapping windows. Three time-domain features that reflect the differences between the falls and other movements in our daily lives are extracted from these consecutive data windows. Then, each of these windows is classified as representing either a fall or a non-fall event by using a trained machine learning classifier. Decision Tree is chosen as the classifier because of its low algorithm complexity and easy implementation on embedded systems. Experimental results have shown that our proposed method can effectively differentiate the fall events from other human daily activities in spite of their high similarity in some cases, with the Accuracy of 99.52%, Precision of 0.993, Recall of 0.995 and F-measure of 0.994.