A shear-sensitive artificial Meissner sensor for event-driven dynamic tactile sensing

초록

Rapidly adapting tactile cues are essential for detecting contact transitions, incipient slip, and texture. However, many artificial tactile systems either prioritize static normal forces or depend on dense arrays with continuous, high-bandwidth processing. To address this limitation, we present a shear-sensitive artificial Meissner sensor (AMS) that mechanically prefilters contact inputs prior to piezoelectric transduction. The sensor integrates an upper structure of stiff Mold Max 60 with a base of compliant Mold Max 10 T, featuring fingerprint-like ridges and an embedded 80-µm poly(vinylidene fluoride) (PVDF) film. This stiffness- and geometry-guided architecture attenuates slowly varying normal loading while preferentially transferring shear-induced deformation, vibration, and contact-transition events to the PVDF layer. Finite-element analysis verified a robust shear-selective design, with the selected structure retaining 98.3% of the maximum shear-to-normal stress ratio obtained in the design space. Experimentally, the AMS showed a low quasi-static normal sensitivity of 7.70 × 10⁻² mV kPa⁻¹ and a velocity sensitivity of 15.8 mV (mm s⁻¹)⁻¹ , confirming its preference for dynamic contact inputs. In a gripper-relaxation and load-perturbation test, the AMS generated a 26.0-mV transient near the onset of object motion, demonstrating its potential for slip-aware contact monitoring. For texture recognition, the AMS classified 4763 fabric segments with 97.93% segment-level accuracy using a random forest model. These results demonstrate mechanical prefiltering as a simple route to directional, dynamic tactile sensing that preserves both contact-transition and texture information with a single compact transducer. © 2026 Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.

키워드