A Low-Cost Tactile Fingertip Design for Dexterous Robotic Hands

Zhuowei Xu, Zilin Si, Kevin Lee Zhang, Moonyoung Lee, Shashwat Singh, Oliver Kroemer, Zeynep Temel

Keywords: Tactile sensing, dexterous manipulation, bio-inspired robot design

TL;DR: low-cost multi-modal tactile fingertip

Abstract:

Touch sensing is a key modality that humans utilize to perform dexterous
manipulation tasks in complex environments. However, most robotic hand
systems do not have tactile sensors due to their high costs, difficulty of
integration, and lack of reliability and robustness. In this work, we
introduce a low-cost, sensorized, bio-inspired fingertip design for robotic
hands. Our proposed fingertip design leverages strain gauges and a contact
microphone to capture static and dynamic contact information. In addition,
the design features a fingernail that enhances small object manipulation
and acts as a vibration-sensitive intermediate to improve dynamic contact
detection. The fingertip is compact, with a minimal sensor footprint and
can be easily customized to be mounted on robotic hands. Through sensor
characterization, we demonstrate the high repeatability of the sensors with
a range of 0 to 5 N detection on 2-DoF forces and its ability to reliably
distinguish different materials. This fingertip design provides a simple
yet effective solution for sensing contacts with robotic hands for fine and
dexterous manipulation.

Submission Number: 28