CushSense: Soft, Stretchable, and Comfortable Tactile-Sensing Skin for Physical Human-Robot Interaction

TL;DR

CushSense presents a soft, stretchable, fabric-based tactile-sensing skin for whole-arm human-robot interaction, enabling safe and comfortable contact through capacitive taxels arranged across a Kinova Gen3 arm. The design emphasizes low cost (~US$7 per taxel), modularity, shielding to suppress noise, and open-source fabrication, achieving a relative force-sensing error of $0.58\%$ with durable performance over $1000$ cycles. A physics-based capacitance model addresses deformations from axial, lateral, and bending motions, and a user study demonstrates superior perceived safety and comfort compared with a Scuba Fabric variant. The work advances practical pHRI by delivering a scalable, open hardware skin with strong performance and clear deployment guidance for researchers and developers.

Abstract

Whole-arm tactile feedback is crucial for robots to ensure safe physical interaction with their surroundings. This paper introduces CushSense, a fabric-based soft and stretchable tactile-sensing skin designed for physical human-robot interaction (pHRI) tasks such as robotic caregiving. Using stretchable fabric and hyper-elastic polymer, CushSense identifies contacts by monitoring capacitive changes due to skin deformation. CushSense is cost-effective ($\sim$US\$7 per taxel) and easy to fabricate. We detail the sensor design and fabrication process and perform characterization, highlighting its high sensing accuracy (relative error of 0.58%) and durability (0.054% accuracy drop after 1000 interactions). We also present a user study underscoring its perceived safety and comfort for the assistive task of limb manipulation. We open source all sensor-related resources on https://emprise.cs.cornell.edu/cushsense.

Figures (8)

• Figure 1: Whole-Arm Soft and Stretchable Robot Skin. We cover a robot arm with CushSense to perform limb manipulation. CushSense comprises taxels which use soft and stretchable materials, allowing each layer to stretch while maintaining reliable sensing and comfort.
• Figure 2: Whole-arm Skin Structure. (a) A taxel with 9 layers acts as a capacitive sensor. (b) An array of taxels forms a skin section, where each taxel measures a capacitance value. (c) We daisy-chain these sections to build a whole-arm skin comprising 56 taxels. (d) An Arduino board collects data and publishes it on a ROS topic, used to estimate contact forces, as shown on a 3D robot model in RViz.
• Figure 3: Sensor Characteristics and Performance Evaluation. The top-left corner of (a) shows the experiment setup comprising F/ T Sensor (in light gray), Indenter (in dark gray), and the taxel (in purple). Experiment (a) validates the force-sensing capability of our skin, (b) exhibits minimal hysteresis error, (c)-(d) shows reduction of noise due to electromagnetic interference, and (e)-(f) retains sensitivity under compression and deformation.
• Figure 4: Comparative Analysis with Alternative Designs. Experiments (a)-(b) illustrate that Squishy as a dielectric maximizes sensitivity. (c) Smaller taxels exhibit further increases in sensitivity. We chose a taxel size of 3x3 cm based on a tradeoff between sensitivity and ease of construction.
• Figure 5: Math Model Validation. Comparison plots validate the efficacy of the proposed mathematical model for different actions, where Measured Capacitance represents experimental data, and Analytical Capacitance is based on the developed mathematical model. The fitted model closely resembles the measured data for all actions.