.png){kind=icon}
.png){kind=icon}
Return
Aligned electrospun nanofiber-engineered triboelectric interface for non-acoustic human-robot communication and material-adaptive manipulation
.png){kind=icon}
PRE
AI
P
Y
B
A
W
H
Z
DOI：10.1016/j.compositesb.2026.114060.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Advanced human-robot interfaces require the seamless integration of intention decoding, environmental perception, and adaptive actuation within an energy-efficient framework. Herein, we present a multimodal triboelectric interface engineered with aligned electrospun nanofibers that enables both non-acoustic human-robot communication and material-adaptive robotic manipulation. The system integrates a flexible facial triboelectric nanogenerator (TENG) sensor array for articulatory motion sensing with a gripper-mounted tactile TENG sensor matrix for material perception. To overcome the inherent poor spinnability of PDMS arising from its low molecular weight and insufficient chain entanglement, coaxial electrospinning with a sacrificial PVP shell is employed to fabricate structurally uniform and directionally aligned PDMS nanofibers. These are paired with aligned Nylon 6/6 nanofibers to form an optimized triboelectric interface with enhanced dielectric properties and superior surface charge density, yielding significantly higher electrical output than randomly oriented counterparts. The facial TENG array achieves real-time decoding of nine silent speech commands with 95.56% accuracy, while the tactile sensor matrix performs six-class material recognition with 99.44% accuracy under varying force, temperature, and humidity. These sensing modalities are fused within a closed-loop control framework, wherein inferred material properties dynamically modulate robotic gripping force for adaptive, damage-free object manipulation. Despite these advances, the current proof-of-concept study is limited by the user-dependent nature of silent speech signals and the sensitivity of static thresholds to variations in wearing conditions, articulation intensity, and background noise. Future work will address cross-user generalization through multi-participant transfer learning and adaptive thresholding strategies. By combining nanofiber interface engineering, multimodal triboelectric sensing, and closed-loop adaptive control, this work establishes a scalable pathway toward intelligent, context-aware, and autonomous human-robot interaction systems for wearable electronics, assistive robotics, and complex environment manipulation.
Journal
IF:
14.2
Papers:
1.2W
Citations:
8.9W
Organization
Cited Papers