Human activity classification with miniature inertial and magnetic sensor signals
Author
Yüksek, Murat Cihan
Barshan, Billur
Date
2011Source Title
19th European Signal Processing Conference, 2011
Print ISSN
2219-5491
Publisher
IEEE
Pages
956 - 960
Language
English
Type
Conference PaperItem Usage Stats
95
views
views
23
downloads
downloads
Abstract
This study provides a comparative performance assessment of various pattern recognition techniques on classifying human activities that are performed while wearing miniature inertial and magnetic sensors. Activities are classified using five sensor units worn on the chest, the arms, and the legs. Each sensor unit comprises a tri-axial accelerometer, a tri-axial gyroscope, and a tri-axial magnetometer. The classification techniques compared in this study are: naïve Bayesian (NB), artificial neural networks (ANN), dissimilaritybased classifier (DBC), various decision-tree algorithms, Gaussian mixture model (GMM), and support vector machines (SVM). The methods that result in the highest correct differentiation rates are found to be GMM (99.1%), ANN (99.0%), and SVM (98.9%). © 2011 EURASIP.
Keywords
Classification techniqueComparative performance assessment
Decision-tree algorithm
Differentiation rate
Gaussian Mixture Model
Human activities
Pattern recognition techniques
Sensor units
Tri-axial magnetometer
Triaxial accelerometer
Intelligent agents
Magnetic sensors
Neural networks
Pattern recognition
Signal processing
Support vector machines