A wearable real-time fall detector based on Naive Bayes classifier

Author

Yang, Xiuxin ; Dinh, Anh ; Che, Li

Author_Institution

Dept. of Electr. & Comput. Eng., Univ. of Saskatchewan, Saskatoon, SK, Canada

fYear

2010

fDate

2-5 May 2010

Firstpage

1

Lastpage

4

Abstract

In this paper, we implement a wearable real-time system using the Sun SPOT wireless sensors embedded with Naive Bayes algorithm to detect fall. Naive Bayes algorithm is demonstrated to be better than other algorithms both in accuracy performance and model building time in this particular application. At 20Hz sampling rate, two Sun SPOT sensors attached to the chest and the thigh provide acceleration information to detect forward, backward, leftward and rightward falls with 100% accuracy as well as overall 87.5% sensitivity.

Keywords

Bayes methods; accelerometers; biomechanics; geriatrics; medical signal processing; signal classification; wireless sensor networks; Naive Bayes classifier; Sun SPOT wireless sensors; acceleration; accuracy performance; model building time; wearable real-time fall detector; Acceleration; Accuracy; Classification algorithms; Sensor phenomena and characterization; Testing; Training; Naive Bayes classifier; Sun SPOT; accelerometer; fall detection; machine learning;

fLanguage

English

Publisher

ieee

Conference_Titel

Electrical and Computer Engineering (CCECE), 2010 23rd Canadian Conference on

Conference_Location

Calgary, AB

ISSN

0840-7789

Print_ISBN

978-1-4244-5376-4

Electronic_ISBN

0840-7789

Type

conf

DOI

10.1109/CCECE.2010.5575129

Filename

5575129