Article metrics

  • citations in SCindeks: 0
  • citations in CrossRef:0
  • citations in Google Scholar:[=>]
  • visits in previous 30 days:5
  • full-text downloads in 30 days:4
article: 1 from 1  
Zbornik radova, Elektrotehnički institut "Nikola Tesla"
2017, vol. 27, iss. 27, pp. 1-9
article language: Serbian
document type: Professional Paper
published on: 21/12/2017
doi: 10.5937/zeint27-13603
Creative Commons License 4.0
Using probability density function in the procedure for recognition of the type of physical exercise
aUniversity of Belgrade, Electrical Engineering Institute 'Nikola Tesla' + University of Belgrade, Faculty of Electrical Engineering
bUniversity of Belgrade, Electrical Engineering Institute 'Nikola Tesla'



This paper presents a method for recognition of physical exercises, using only a triaxial accelerometer of a smartphone. The smartphone itself is free to move inside subject's pocket. Exercises for leg muscle strengthening from subject's standing position squat, right knee rise and lunge with right leg were analyzed. All exercises were performed with the accelerometric sensor of a smartphone placed in the pocket next to the leg used for exercises. In order to test the proposed recognition method, the knee rise exercise of the opposite leg with the same position of the sensor was randomly selected. Filtering of the raw accelerometric signals was carried out using Butterworth tenth-order low-pass filter. The filtered signals from each of the three axes were described using three signal descriptors. After the descriptors were calculated, a probability density function was constructed for each of the descriptors. The program that implemented the proposed recognition method was executed online within an Android application of the smartphone. Signals from two male and two female subjects were considered as a reference for exercise recognition. The exercise recognition accuracy was 94.22% for three performed exercises, and 85.33% for all four considered exercises.


Accelerometer; physical exercise; signal descriptor; probability density function; exercises recognition


Ar, I., Akgul, Y.S. (2014) A computerized recognition system for the home-based physiotherapy exercises using an RGBD camera. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 22, (6), pp. 1160-1171.
Bonnet, V., Joukov, V., Fraisse, P., Ramdani, N., Venture, G. (2016) Monitoring of hip and knee joint angles using a single inertial measurement unit during lower limb rehabilitation. IEEE Sensors Journal, 16, (6), pp. 1557-1564.
Ermes, M., Parkka, J., Mantyjarvi, J., Korhonen, I. (2008) Detection of daily activities and sports with wearable sensors in controlled and uncontrolled conditions. IEEE Transactions on Information Technology in Biomedicine, Vol. 12, No. 1, pp. 20-26,
Guan, Q., Yin, X., Guo, X., Wang, G. (2016) A novel infrared motion sensing system for compressive classification of physical activity. IEEE Sensors Journal, 16, (8), pp. 2251-2259.
Phan, N., Ebrahimi, J., Kil, D., Piniewski, B., Dou, D. (2016) Topic-aware physical activity propagation in a health social network. IEEE Intelligent Systems, 31, (1), pp. 5-14.
Toth-Laufer, E., Varkonyi-Koczy, A.R. (2014) A soft computing-based hierarchical sport activity risk level calculation model for supporting home exercises. IEEE Transactions on Instrumentation and Measurement, 63, (6), pp. 1400-1411.
Wang, A., Chen, G., Yang, J., Zhao, S., Chang, C.Y. (2016) A comparative study on human activity recognition using inertial sensors in a smartphone. IEEE Sensors Journal, 16, (11), pp. 4566-4578.
Xu, J., Wang, Y., Barrett, M., Dobkin, B., Pottie, G.J., Kaiser, W.J. (2016) Personalized multilayer daily life profiling through context enabled activity classification and motion reconstruction: An integrated system approach. IEEE Journal of Biomedical and Health Informatics, 20, (1), pp. 177-188.