Introducing a Modular Activity Monitoring System
Attila Reiss, Didier Stricker
Conference Proceedings of the 33rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC-11), 33rd, August 30 - September 3, Boston, MA, USA

Abstract:

In this paper, the idea of a modular activity monitoring system is introduced. By using different combinations of the system�s three modules, different functionality becomes available: 1) a coarse intensity estimation of physical activities 2) different features based on HR-data and 3) the recognition of basic activities and postures. 3D-accelerometers � placed on lower arm, chest and foot�and a heart rate monitor were used as sensors. A dataset with 8 subjects and 14 different activities was recorded to evaluate the performance of the system. The overall performance on the intensity estimation task, relying on the chest-worn accelerometer and the HR-monitor, was 94.37%. The overall performance on the activity recognition task, using all three accelerometer placements and the HR-monitor, was 90.65%. This paper also gives an analysis of the importance of different accelerometer placements and the importance of a HR-monitor for both tasks.

Introducing a Modular Activity Monitoring System
(Hrsg.)
Conference Proceedings of the 33rd Annual International Conference of the IEEE Engineering in Medicine and Biology Society International Conference of the IEEE Engineering in Medicine and Biology Society (EMBC-11), 33rd, August 30 - September 3, Boston, MA, USA

Abstract:

In this paper, the idea of a modular activity monitoring system is introduced. By using different combinations of the system�s three modules, different functionality becomes available: 1) a coarse intensity estimation of physical activities 2) different features based on HR-data and 3) the recognition of basic activities and postures. 3D-accelerometers � placed on lower arm, chest and foot�and a heart rate monitor were used as sensors. A dataset with 8 subjects and 14 different activities was recorded to evaluate the performance of the system. The overall performance on the intensity estimation task, relying on the chest-worn accelerometer and the HR-monitor, was 94.37%. The overall performance on the activity recognition task, using all three accelerometer placements and the HR-monitor, was 90.65%. This paper also gives an analysis of the importance of different accelerometer placements and the importance of a HR-monitor for both tasks.