TACTILE GLOVE FOR BREAST LUMP DETECTION AND HANDOVER-ORIENTED OBJECT CLASSIFICATION

Abstract

Breast cancer is one of the most common causes of mortality among women. Inspection of lumps in the breasts by palpation is the key to early detection and leads to a higher survival rate. Likewise, transferring an object between a robot and operator is one of the milestone problems that can be solved with proposed prior-to-pass classification-driven handover pipeline. Therefore, in this work, I present a flexible fabric-based and soft wearable tactile glove interfaced to the Robot Operating System (ROS) for detecting the lumps within custom-made breast prototypes and soft/rigid object classification prior to handover. The glove can be donned by the human hand or an anthropomorphic robot hand, and capable of measuring the force and mechanical vibrations at the points of the contact and finger motions while squeezing the breast models. Tactile glove incorporates 30 barometric pressure sensors with 1 kPa accuracy, five proprioceptive sensors, and two accelerometers with 0.1 g accuracy. The layered construction provides a pleasant look and protects the delicate wirings.