A wearable system based on a breathable cloth wristband equipped with stretchable strain gauge sensors were assembled and tested to detect a set of 16 different hand gestures. The sensors embedded on the wristband prototype do not require a direct contact with the skin, thus maximizing comfort. To evaluate the performance of the developed band, different gestures were labelled by using grasping information detected in real-time by commercial Force-Sensing Resistor (FSR) sensors. Signals recorded by the wristband were processed through two machine-learning algorithms, i.e. Linear Discriminant Analysis (LDA) and Support Vector Machine (SVM), reaching accuracies of 87% and 95% respectively.
9 Nov 2016
2016 IEEE Healthcare Innovation Point-Of-Care Technologies Conference (HI-POCT)