Haptic Glove for Object Recognition and Hand Motion Detection

Abstract

Haptic perception plays an important role in robotics research area. In recent years, various tactile sensors have been developed to simulate human skin. In this paper, the authors improved the tactile gloves to detect different objects by haptic touch, compared to the tactile sensors made by other groups. To evaluate the performance of the tactile gloves presented in this paper, the small-scale 2D and 3D convolutional neural network that are for classifying the objects and detecting motion based on pressure map has been set. The evaluation results indicate that this methodology proved the practicality of the pressure sensor array. It will act as the promising foundation for the subsequent human-computer interaction experiments.