Advances in Wearable Sensor-Based Automatic Tremor Classification

This work reports advances in tremor classification based on a wireless wearable sensor presented by the authors in a previous paper. This device was characterized in order to determine the reliability of the platform when used as a frequency measurement unit. Results confirm the system accuracy in the estimation of the dominant frequency of the measured vibrations. In addition, device's orientations, accelerations and angular speeds were collected from 15 patients affected by parkinsonian tremor during the execution of four standardized tasks: those data were analyzed in order to assess the equivalence of the results obtained from different data sources. The procedure highlights both a strong correspondence between the different data sources in the task that emphasizes the parkinsonian tremor, and similar trends in the results.