Home // ADAPTIVE 2020, The Twelfth International Conference on Adaptive and Self-Adaptive Systems and Applications // View article

What Are You Doing? Real-Time Activity Recognition Using Mobile Phone Sensors

Authors:
Bernhard Hiesl
Marc Kurz
Erik Sonnleitner

Keywords: Activity recognition; Mobile sensing; Self-adaptation; Adaptive application; Adaptive real-time strategies

Abstract:
This paper focuses on recognizing different activities of people by utilizing the smart-phone as sensor delivering unit. For the sake of simplicity, five different activities (i.e., modes of locomotion) are considered: (i) standing, (ii) walking, (iii) running, (iv) walking upstairs and (v) downstairs. The research challenge is the fact that the phone is placed on the body of the subjects dynamically and orientation independent - thus the system has to adapt autonomously to these characteristics. The sensor data was collected from the built-in accelerometer, gravity and gyroscope sensors of a common smartphone (i.e., Google Pixel XL). The data collection procedure is part of the dynamic placement of this paper addressing position and orientation independent recording and recognizing of activities. Additionally, to acquire full orientation independence, data transformation (horizontal and vertical movement) is applied on the gathered data. Within the framework of this study, five different phone positions are taken into account and are therefore considered in the classification process. To achieve the best approach concerning performance and recognition, different classifiers are evaluated (i.e., (i) k-nearest neighbours (KNN), (ii) Naive Bayes, (iii) Decision Trees and (iv) Random Forest).

Pages: 1 to 6

Copyright: Copyright (c) IARIA, 2020

Publication date: April 26, 2020

Published in: conference

ISSN: 2308-4146

ISBN: 978-1-61208-781-8

Location: Nice, France

Dates: from October 25, 2020 to October 29, 2020