With the rapid development of smartphone, human activity recognition based on acceleration sensors attracts much attention in the academic and industry recently. However, the recognition accuracy is not ideal due to the diversity of human activities and other environmental factors. A real-time user activities monitoring system is developed on android, and comparison of several feature extraction and classification algorithms is carried out. Based on the monitoring system, a feature called (TF4+FFT10) is proposed. Experiment result shows that the recognition accuracy rate of feature (TF4+FFT10) with the adopted KNN algorithm is 98.6%.
CITATION STYLE
Cai, S., Shan, Z., Zeng, T., Yin, J., & Ming, Z. (2017). Human activity recognition based on smart phone’s 3-axis acceleration sensor. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10135 LNCS, pp. 163–172). Springer Verlag. https://doi.org/10.1007/978-3-319-52015-5_17
Mendeley helps you to discover research relevant for your work.