A Comparison of Feature Extraction and Selection Techniques for Activity Recognition using Low-cost Sensors on a Smartphone

Sara Saeedi, Zainab Syed and Naser El-Sheimy

Abstract:	Recent popularity of mobile devices has resulted in considerable research efforts directed towards the 1 Mobile Multi-Sensor Systems is a research group at the University of Calgary that works on inertial sensors, integrated navigation systems and pedestrian navigation. recognition and monitoring of dynamic activity patterns using the low-cost sensors embedded in mobile devices such as smartphones and tablet personal computers (Tablet-PCs). In this research, a comparison has been conducted on different sensor data, feature spaces and feature selection methods to increase the efficiency and reduce the computation cost of activity recognition. The aim of this research is to find the best set of sensors and features needed to recognize the placement of the mobile device on the user as well as different user activities that are commonly used in everyday life. Using extensive experiments, the performance of various feature spaces have been evaluated. The results showed that the Bayesian Network classifier yields recognition accuracy of 96.2% using four features while requiring fewer computations.
Published in:	Proceedings of the 25th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2012) September 17 - 21, 2012 Nashville Convention Center, Nashville, Tennessee Nashville, TN
Pages:	3140 - 3146
Cite this article:	Saeedi, Sara, Syed, Zainab, El-Sheimy, Naser, "A Comparison of Feature Extraction and Selection Techniques for Activity Recognition using Low-cost Sensors on a Smartphone," Proceedings of the 25th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2012), Nashville, TN, September 2012, pp. 3140-3146.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In