Fuzzy Inference System for Physical Activity Recognition Based on General Features and PCA

DOI

10.2991/3ca-13.2013.13How to use a DOI?

Keywords

physical activity recognition; principal component analysis; fuzzy inference system

Abstract

This paper presents an approach of applying principal component analysis (PCA) and fuzzy inference system (FIS) to recognition of activity of daily life (ADL). To overcome the non-intuitiveness of single acceleration signal and difficulties of feature selection manually, 32 common features are computed and PCA is used for feature reduction. The membership functions are obtained from training data, and fuzzy rules are in same form for all classes. Thus, the FIS is not dependent on expert knowledge of physical activity. Thus, this system is extendable on new types of activity, new features or new locations. This system is designed for a real-time sensor-based monitoring system to recognize 6 types of daily physical activities. Sitting, standing, walking, going upstairs, going downstairs and running are classified with a precision of 99.78%, 90.78%, 91.89%, 89.72%, 91.28% and 100% for each type.

Copyright

© 2013, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Tongbi Kang
AU  - Lian Ying
AU  - Jiankang Wu
AU  - Yi Sun
AU  - Xiaoli Meng
PY  - 2013/04
DA  - 2013/04
TI  - Fuzzy Inference System for Physical Activity Recognition Based on General Features and PCA
BT  - Proceedings of the 2nd International Symposium on Computer, Communication, Control and Automation
PB  - Atlantis Press
SP  - 50
EP  - 54
SN  - 1951-6851
UR  - https://doi.org/10.2991/3ca-13.2013.13
DO  - 10.2991/3ca-13.2013.13
ID  - Kang2013/04
ER  -