An improved step counting algorithm using classification and double autocorrelation

Abstract

The objective of this paper was to develop an end-to-end algorithm that would improve the step counting accuracy in regular walking/running data and also meet the ANSI/CTA-2056 standards. The ANSI/CTA-2056 standards are to achieve an error rate of less than 10% on treadmill data on at least 20 participants. Our UWF-algorithm (UWFv1) has an improved step counting accuracy and also performs well below the acceptable ANSI/CTA-2056 error rate, using both treadmill data and non-treadmill data, hence our UWFv1 algorithm also meets the ANSI-CTA-2056 standards. For the end-to-end algorithm, the random forest model, trained on a feature engineered dataset, was chosen as the walking/running detection classifier, and double autocorrelation was recommended in the process of determining the step counts.

KEYWORDS:

• Step counting
• regular walking/running data
• autocorrelation
• time-series data
• machine learning algorithms
• ANSI/CTA standards

Acknowledgments

This work has been partially supported by the Askew Institute of the University of West Florida and Actigraph.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Sikha Bagui

Dr Sikha Bagui is Professor and Askew Fellow in the Department of Computer Science, at The University West Florida, Pensacola, Florida. Dr Bagui is active in publishing peer reviewed journal articles in the areas of database design, data mining, BigData and Big Data analytics, and machine learning. Dr Bagui has worked on funded as well unfunded research projects and has over numerous peer reviewed publications. She has also co-authored several books on database and SQL. Bagui also serves as Associate Editor and is on the editorial board of several journals.

Xingang Fang

Xingang Fang has received his PhD degree in Chemistry from Emory University and MS degree in Computer Science from University of West Florida and has worked as researcher and teacher for many years. He is currently a PhD student in Computer Science at Florida State University, focusing on research in high performance computing and machine learning.

Subhash Bagui

Subhash Bagui received his BSc in Statistics from University of Calcutta, M. Stat. from Indian Statistical Institute and PhD from University of Alberta, Canada. He is currently a University Distinguished Professor at the University of West Florida. He has authored a book titled, “Handbook of Percentiles of Non-central t-distribution”, and published many high quality peer reviewed journal articles. He is currently serving as associate editors/editorial board members of several statistics journals. His research interests include nonparametric classification and clustering, statistical pattern recognition, machine learning, central limit theorem, and experimental designs. He is also a fellow of American Statistical Association (ASA) and Royal Statistical Society (RSS).

Jeremy Wyatt

Jeremy Wyatt is the President and CTO of ActiGraph, an industry leading global provider of physical activity and sleep monitoring solutions for the pharmaceutical and academic industries. Jeremy’s 20 years of experience in working with low-power micro-electro-mechanical systems and related cloud technology have given him a unique perspective on the challenges and opportunities of deploying wearable technology to produce meaningful patient data. He has an undergraduate degree in Electrical Engineering from the University of Florida and an MBA from the University of West Florida and is a member of the Digital Medicine (DiMe) Society.

Patrick Houghton

Patrick Houghton is the Senior Vice President of Operations at ActiGraph. Patrick has a demonstrated history of working in the medical device industry for nearly 10 years and has a Master’s degree in Software Engineering from the University of West Florida. He has led software development and engineering projects with ActiGraph for nearly 10 years and has over 14 years of combined experience in both government and private sector. His experience includes database design, web application development, data integration & management, software-hardware interface, hardware design/development, software development life-cycle implementation, project management, cloud-based systems architecture, virtualization, health care information systems, engineering process improvement, manufacturing, logistics, and inventory management. Key projects include development of ActiGraph’s CentrePoint ecosystem, CentrePoint Insight Watch re-design, and DataHub LTE cellular integration. Mr Houghton oversees day-to-day operations at ActiGraph.

Joe Nguyen

Joe Nguyen graduated in 2004 from the University of Florida with a Bachelor of Science in Electrical Engineering and Computer Engineering. The bulk of his career has been in printed circuit board design, microcontroller firmware, and industrial sensors. He joined ActiGraph in 2015 where his primary duties involve digital signal processing, algorithm analysis, and data integrity testing. In addition to his main duties, his interests include deep learning, metaheuristic optimization, and error correction coding for digital communications.

John Schneider

John Schneider is the former Senior Vice President of Research & Development at ActiGraph. For more than 25 years, he worked closely with ActiGraph’s hardware development team in the selection and execution of several native sensors and a new wireless sensor interface. A major focus of his work was in the development, refinement, and implementation of various algorithms and equations that allow end users to derive meaningful endpoints from raw accelerometer output. He has developed or collaborated with others to develop equations related to physical activity energy expenditure, heart rate energy expenditure, cut point values, step counts, body position and sleep/wake behavior. Mr Schneider earned a BS in Electrical Engineering from the University of Wisconsin, Madison.

Tyler Guthrie

Tyler Guthrie is the Director of Scientific Affairs and Data Management at ActiGraph. His responsibilities focus around educating clients on how ActiGraph’s devices collect data and how this is transformed into multiple meaningful endpoints. He also works with data teams to determine data formatting, optimum endpoints, and most applicable algorithms for their studies in the therapeutic area. He received his BSc in Exercise Science from the University of North Florida and MSc in Exercise Physiology from Florida State University.