Continuous capture/measuring of hand and arm vibration via smartwatches

Status:

completed 03/2016

Aims:

Long‐lasting and intense vibration of the hand and the arm can cause serious diseases. The estimation of risks with respect to the hand and arm vibration can be achieved by the consideration of exposure duration and Vibration intensity. Manual assessment of exposure durations and used tools is expensive, disturbs the workflow or may be performed only sporadically and rarely due to high costs.

Robust and cheap smartwatches including a plurality of integrated sensors are available on the market. Although smartwatches provide accurate acceleration, rotation rate, acoustic sensors, and a powerful processing unit, it is so far unclear whether these systems can be used to assess hand‐arm vibration (HAV).

Activities/Methods:

In a feasibility study by Institut für Arbeitsschutz IFA, St. Augustin, and Fraunhofer IGD, Rostock, it was evaluated if vibration pattern and sound features are suitable to identify the used tools under laboratory conditions and in field trials. As a result of the evaluation, the possibilities and constraints for an individual determination of handarm vibration have been worked out.

Results:

This study shows that a continuous monitoring and estimation of HAV by Smartwatches is possible. During the trials, the sampling rate of the acceleration sensor was 50 Hz, the sound data were recorded at 8 kHz with a window frame length of 1.28 seconds each. 71 features were selected and their relevance was examined. It was found that in field trials a subset of approx. 9 ‐ 15 features are relevant. In a field test of four different tools, the features were classified by a J48 decision tree, which performed with an accuracy of 72 percent. The A (8) – estimation showed an overvaluation of about 11 percent.

Last Update: