Comparison of hand gesture inputs of leap motion controller & data glove in to a soft finger

Abstract

Gesture recognition devices in the market are getting popular today. Many of these devices are used different technologies to recognize gestures and generate an output to control different mechanisms. In this research, a data glove has developed to track the motion of the hand & compare its performance against Leap Motion Controller to control a Soft Finger mechanism. A data glove has developed to track the motion of the human hand using flex sensors, gyroscopes and vision data. Position, orientation, velocity & acceleration, bending angle of the fingers has extracted from the data. Similar data has extracted from the Leap Motion controller and then performance has compared. Then required parameters has extracted from the data set and fed into the virtual elastomer simulation and bending angle of a single Soft Finger has studied. The average percentage error between Leap Motion and the Data Glove for the bending angle was found to be 26.36% & 18.21% with respect to the standard finger behavior. Then the average standard deviation of the orientation has obtained for Yaw, Pitch & Roll separately for Leap Motion and Data Glove. The Leap Motion & Data Glove bending angle data has the fed to the finite element simulation and the average percentage error of the response generated has found to be 10.13% for the Leap Motion and 33.03% for the Data Glove. Therefore, Leap Motion Controller shows a high repeatability and high potential in using for Soft Finger type applications. Improvements to this system and material optimization could lead this mechanism to high precession applications.