Development of multi-sensor system for dynamic analysis of prehension tasks

Abstract

Prehension involves coordination of two components, reach and grasp. Generally, studies about reach-to-grasp were made in static conditions, in which a person’s base of support is wide and stable. When performed combined with gait, the reach-to-grasp movement may be affected by dual task. Recently, hand gesture recognition has attracted a growing interest and can be acquired through a variety of sensors. The Leap Motion Controller introduces a new gesture and position tracking system with sub-millimeter accuracy. However, in a dynamic scenario, such device revealed an inconsistent performance. Therefore, the aim of this study is the improvement of Leap Motion system accuracy in the analysis of reach-to-grasp movements during walking combined with prehension. Two experimental protocols were defined for gait combined with the prehension task. The first one, using one Leap Motion Controller device, whereas the second uses the information obtained from two devices. Five young healthy subjects (24.8 ± 2.96 years) participated of this study. Five tests of each protocol were performed for each participant. The parameters analyzed in this work were the hand grip aperture, through the thumb and index tip positions, the peak grip aperture, the time to peak grip aperture and the stationary hand grip aperture. Results show acquisition failure and low accuracy in the first protocol. In the second, the errors are lower and the tests showed 100% acquisition success. Results show that the use of two Leap Motion Controller decrease the error mean from 47.75 to 18.87% in dynamic tests and reduces acquisition failures.