Human Activity Monitoring with Wearable Sensors and Hybrid Classifiers
Gamze Uslu, H. Ibrahim Dursunoglu, Ozgur Altun, Şebnem Baydere · 2013
Activity monitoring plays a crucial role in ambient living environments for assessing changes in the normal behav- ioral pattern of elderly people. In this paper, we present an ac- tion description and detection mechanism for real-time activity monitoring using wearable sensors and hybrid classifiers. F irst a single sensor single classifier model is presented (SSSC) f or the detection of simple and composite actions. Then the model is en- hanced with multiple sensors and classifiers for the purpose of real-time monitoring. The enhanced Multi-Sensor Multi Clas- sifier (MSMC) model uses two wearable TI Chronos watches with a built-in tri-axial accelerometer for data acquisiti on and a composition of naive Bayes, Susan Corner Detector(SCD) and Hidden Markov(HMM) classifiers for the detection of transi- tions between defined actions in real-time. A real testbed en vi- ronment is established to asses the success of real-time monitor- ing. The test results have revealed that SSSC model is highly successful in controlled tests when the burden of real-time sam- pling is ignored whereas MSMC model is fast and accurate for real time detection of transitions between actions. The proposed models are tested against the simple actions; walk, sit, stand, lie as well as walk-while-hands-in-pocket and walk-on-wheelchair. The unique feature of the selected actions is that the transi- tion between walk, sit and lie are the most likely causes of a fall event in a home environment for elderly people. The best achieved detection rates for simple actions range between 92- 100 % for SSSC model whereas MSMC model is 100 % success- ful in real-time detection of transitions with a slightly re duced achievement for individual actions.