Comprehensive evaluation of human activity classification based on inertia measurement unit with air pressure sensor

This paper focuses on accuracy improvement of human activities detection and classification by using single Inertia Measurement Unit sensor (IMU sensor: an acceleration sensor, a gyro sensor, a magnetometer, and an air pressure sensor) which is a type of the wearable sensors. Generally, performance of classification model is determined by these methodologies; number and type of sensors, coordinate transformation, time window, time-frequency domain analysis, and machine learning algorithms. The contributions of this paper are summarized in the following three points. Firstly, a pressure sensor is additionally utilized to improve the accuracy of human activities estimation. This information is effective to estimate up/down motion by stair and elevator. Secondly, comprehensive evaluation of the combinations using different methodologies is conducted to find an optimal classification model. Thirdly, ensemble learning is performed to improve estimation accuracy. It shows superior performance with over 95 % accuracy of human activity estimation.

[1]  Davide Anguita,et al.  Energy Efficient Smartphone-Based Activity Recognition using Fixed-Point Arithmetic , 2013, J. Univers. Comput. Sci..

[2]  S Leonhardt,et al.  Human motion classification based on a textile integrated and wearable sensor array , 2013, Physiological measurement.

[3]  Michael Zyda,et al.  NPSNET: Flight Simulation Dynamic Modeling Using Quaternions , 1992, Presence: Teleoperators & Virtual Environments.

[4]  A. Enis Çetin,et al.  Human activity classification using vibration and PIR sensors , 2012, 2012 20th Signal Processing and Communications Applications Conference (SIU).

[5]  Nader Karimi,et al.  Automatic Monocular System for Human Fall Detection Based on Variations in Silhouette Area , 2013, IEEE Transactions on Biomedical Engineering.

[6]  Bernt Schiele,et al.  Weakly Supervised Recognition of Daily Life Activities with Wearable Sensors , 2011, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[7]  Marko Topič,et al.  Calibration and data fusion solution for the miniature attitude and heading reference system , 2007 .

[8]  Wei Pan,et al.  SoundSense: scalable sound sensing for people-centric applications on mobile phones , 2009, MobiSys '09.

[9]  Sebastian Raschka,et al.  Python Machine Learning , 2015 .

[10]  Vigneshwaran Subbaraju,et al.  Energy-Efficient Continuous Activity Recognition on Mobile Phones: An Activity-Adaptive Approach , 2012, 2012 16th International Symposium on Wearable Computers.

[11]  Jeen-Shing Wang,et al.  A Wearable Sensor Module With a Neural-Network-Based Activity Classification Algorithm for Daily Energy Expenditure Estimation , 2012, IEEE Transactions on Information Technology in Biomedicine.

[12]  Senem Velipasalar,et al.  A Survey on Activity Detection and Classification Using Wearable Sensors , 2017, IEEE Sensors Journal.

[13]  Trevor Hastie,et al.  An Introduction to Statistical Learning , 2013, Springer Texts in Statistics.