Inertial Sensor Based Human Activity Recognition via Reduced Kernel PCA
In the past decade, wearable inertial sensor based human activity recognition has attracted lots of attention from researchers in the world. High-dimensional feature set will increase the computation and memory cost. In this paper, kernel PCA has been utilized for dimensionality reduction to deal with inertial sensor based human activity recognition. However, kernel method may increase the computation and memory cost. Thus, reduced kernel method is proposed. The real dataset has been utilized to evaluate the proposed reduced kernel PCA (RKPCA) method. Experimental results demonstrate the efficacy of the proposed method, which achieves better results than traditional PCA method.
KeywordsHuman activity recognition Inertial sensor Kernel PCA Reduced kernel method
This work was supported by Fundamental Research in cutting-edge technologies in the project of Henan province (162300410070), Key Science Technology Program of Henan Province, China (19A413013), Foundation of Henan Educational Committee (162300410070), and Ph.D. early development program of Zhengzhou University of Light Industry (13501050009).
- 1.Avci, A., Bosch, S., Marin-Perianu, M., Marin-Perianu R., Havinga, P.: Activity recognition using inertial sensing for healthcare, wellbeing and sports applications: a survey. In: Proceedings of 23rd International Conference on Architecture of Computing Systems (ARCS), Feb 2010, pp. 1–10 (2010)Google Scholar
- 6.Cheng, W.C., Jhan, D.M.: Triaxial accelerometer-based fall detection method using a self-constructing cascade-AdaBoost-SVM classifier. IEEE J. Biomed. Health Inf. 17(2), 411–419 (2013)Google Scholar
- 7.Sirichana, W., Dolezal, B.A., Neufeld, E.V., et al.: Wrist-worn triaxial accelerometry predicts the energy expenditure of non-vigorous daily physical activities. J. Sci. Med. Sport (2017)Google Scholar
- 11.Maurer, U., Smailagic, A., Siewiorek, D.P., Deisher, M.: Activity recognition and monitoring using multiple sensors on different body positions. In: Proceedings of International Workshop Wearable Implantable Body Sensor Networks (BSN), Apr 2006, pp. 113–116 (2006)Google Scholar
- 15.Gao, L., Bourke, A.K., Nelson, J.: Activity recognition using dynamic multiple sensor fusion in body sensor networks. In: Conference on Proceedings of the IEEE Engineering in Medicine and Biology Society, p. 1077 (2012)Google Scholar
- 16.Ertuǧrul, Ö.F., Kaya, Y.: Determining the optimal number of body-worn sensors for human activity recognition. Soft Comput., 1–8 (2016)Google Scholar
- 20.Chen, Y., Guo, M., Wang, Z.: An improved algorithm for human activity recognition using wearable sensors. In: 2016 Eighth International Conference on Advanced Computational Intelligence (ICACI), pp. 248–252. IEEE (2016)Google Scholar