RemGlove: A Three-Degree-of-Freedom E-Textile Robotic Hand with Computer-Vision-Guided Control and Vibrotactile Haptic Feedback

Abstract

Wearable robotic hands that translate natural human gestures into robotic motion, while returning a sense of touch to the operator, remain an open challenge at the intersection of soft robotics, electronic textiles (e-textiles), and computer vision. This paper presents RemGlove, a three-degree-of-freedom (DOF) tendon-actuated robotic hand system that is teleoperated through a stretch-sensing e-textile glove, tracked in real time with a monocular red-green-blue (RGB) camera pipeline, and closed with vibrotactile haptic feedback delivered to the operator’s fingertips. The mechanical subsystem couples a bank of servo actuators to an underactuated tendon network; the sensing subsystem embeds piezoresistive e-textile strain channels directly into a fabric glove; and the perception subsystem uses a landmark-based hand-tracking pipeline, implemented and evaluated in MATLAB, to estimate joint targets from video. The design draws on published soft-robotics and functional-fiber research, including work associated with the Advanced Functional Fabrics of America (AFFOA) institute, and is oriented toward defense-relevant wearable applications. We report the mechanical and electrical design, derive the forward kinematics and control equations governing the system, and present calibration data, closed-loop response curves, pipeline latency measurements, and grasp-success statistics collected from a benchtop prototype. The prototype achieves an average end-to-end control latency of approximately 34 ms and a mean grasp success rate of 84% across six object categories. We conclude with a discussion of the system’s broader impact and directions for future work.