Classification Accuracies of Physical Activities Using Smartphone Motion Sensors

By Wanmin Wu, Sanjoy Dasgupta, E. Ramirez, C. Peterson, Gregory J. Norman. Journal of Medical Internet Research, 2012

334 citations recorded by the source corpus.

Background Over the past few years, the world has witnessed an unprecedented growth in smartphone use. With sensors such as accelerometers and gyroscopes on board, smartphones have the potential to enhance our understanding of health behavior, in particular physical activity or the lack thereof. However, reliable and valid activity measurement using only a smartphone in situ has not been realized. Objective To examine the validity of the iPod Touch (Apple, Inc.) and particularly to understand the value of using gyroscopes for classifying types of physical activity, with the goal of creating a measurement and feedback system that easily integrates into individuals’ daily living. Methods We collected accelerometer and gyroscope data for 16 participants on 13 activities with an iPod Touch, a device that has essentially the same sensors and computing platform as an iPhone. The 13 activities were sitting, walking, jogging, and going upstairs and downstairs at different paces. We extracted time and frequency features, including mean and variance of acceleration and gyroscope on each axis, vector magnitude of acceleration, and fast Fourier transform magnitude for each axis of acceleration. Different classifiers were compared using the Waikato Environment for Knowledge Analysis (WEKA) toolkit, including C4.5 (J48) decision tree, multilayer perception, naive Bayes, logistic, k-nearest neighbor (kNN), and meta-algorithms such as boosting and bagging. The 10-fold cross-validation protocol was used. Results Overall, the kNN classifier achieved the best accuracies: 52.3%–79.4% for up and down stair walking, 91.7% for jogging, 90.1%–94.1% for walking on a level ground, and 100% for sitting. A 2-second sliding window size with a 1-second overlap worked the best. Adding gyroscope measurements proved to be more beneficial than relying solely on accelerometer readings for all activities (with improvement ranging from 3.1% to 13.4%). Conclusions Common categories of physical activity and sedentary behavior (walking, jogging, and sitting) can be recognized with high accuracies using both the accelerometer and gyroscope onboard the iPod touch or iPhone. This suggests the potential of developing just-in-time classification and feedback tools on smartphones.

View Paper: https://www.semanticscholar.org/paper/84dfc06c63b6768d684e747807742b18ff2c2959

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

A Review of Wearable Technologies for Elderly Care that Can Accurately Track Indoor Position, Recognize Physical Activities and Monitor Vital Signs in Real Time (2017, methodology): “Some commonly used data-driven approaches are the Hidden Markov Model (HMM) [57], k -Nearest Neighbor algorithm (KNN) [13,60,68,74,85,86], Principal Component Analysis (PCA) [87], Naïve Bayes (NB) [13,59,60,74,86], Gaussian Mixture Model (GMM) [58], Random Forest (RF) [85], Decision Tree…”
Increasing Physical Activity With Mobile Devices: A Meta-Analysis (2012, methodology): “Recently, Wu et al [47] used accelerometer and gyroscope data obtained by smartphone devices to classify activity types using machine learning”
A robust human activity recognition system using smartphone sensors and deep learning (2018, methodology): “Gyroscopes have also been employed for HAR and have demonstrated to improve the recognition performance when used in combination with accelerometers [25], [26].”
Human Activity Recognition Using Recurrent Neural Networks (2017, methodology): “Various approaches have been explored for activity recognition, among them the majority of the techniques focuses on classification algorithms such as Naive Bayes (NB) [12], Decision Trees [13], HMM [5] , CRF [14], Nearest Neighbor (NN) [15], Support Vector Machines (SVM) [16] and different boosting”
A Survey on Activity Detection and Classification Using Wearable Sensors (2017, methodology): “There are systems using single sensor accelerometer-based system [25], [39], [159], [160] and those fusing several sensors [63], [161]”
Smartphone Sensors for Health Monitoring and Diagnosis (2019, methodology): “8 • Activities were classified using supervised machine learning (SVM, Decision tree, KNN and discriminant analysis) algorithms.”
Transition-Aware Human Activity Recognition Using Smartphones (2016, methodology): “Gyroscopes have also been employed for HAR and have demonstrated to improve the recognition performance when used in combination with accelerometers [18,19].”
Deep Learning Models for Real-time Human Activity Recognition with Smartphones (2019, methodology): “Gyros are also used in HAR and have been shown to improve recognition performance when used in conjunction with accelerometers [35–39].”
A Public Domain Dataset for Human Activity Recognition using Smartphones (2013, background): “Consequently, in the last few years, some works aiming to understand human behavior using smartphones have been proposed: for instance in [6], one of the first approaches to exploit an Android smartphone for HAR employing its embedded triaxial accelerometers; additional results have also been presen”
Can Smartphone Apps Increase Physical Activity? Systematic Review and Meta-Analysis (2019, unknown): “They are equipped with advanced technological features, most notably their connection to the internet, global positioning system, and inbuilt accelerometers [9], and offer the capacity to create individualized and interactive apps that collect real-time data [10].”
Behavioral Functionality of Mobile Apps in Health Interventions: A Systematic Review of the Literature (2015, background): “Most published literature regarding health apps has focused on preventing and managing chronic disease [6,7], monitoring health behaviors and vitals [8], content analyses of health and fitness apps [9-11], app acceptability and utility [12-15], and qualitative studies of user experience and desired…”
Fusion of Smartphone Motion Sensors for Physical Activity Recognition (2014, unknown): “Moreover, in the state-of-the-art, frequencies lower than 50 samples per second have been shown to be sufficient for activity recognition [3,4].”
Just a Fad? Gamification in Health and Fitness Apps (2014, background): “This technology has the capacity to provide for easy collection of personal health-related data and providing timely behavioral cues [4,5].”