The rapid advancement of intelligent materials and bio-inspired engineering has opened new frontiers in human-robot interaction (HRI), enabling robots to perceive, respond, and adapt to human touch in real time. In this context, auxetic materials—characterized by their negative Poisson’s ratio—have emerged as promising candidates for enhancing the tactile responsiveness and mechanical adaptability of robotic skins. Their unique property of lateral expansion under tensile strain, coupled with superior energy absorption and strain amplification capabilities, makes them ideal for high-sensitivity tactile interfaces.

In this study, we introduce a real-time tactile detection and localization system tailored for human-robot interaction, built upon an auxetic sensor-embedded multilayer structure. Inspired by the hierarchical organization of biological tissues such as human skin, the proposed architecture integrates mechanical compliance with intelligent sensing. The multilayer system comprises a rotating square patterned auxetic framework that facilitates mechanical transformation, a distributed matrix of piezo-resistive sensors for capturing contact stimuli, and a low-power embedded signal processing module optimized for real-time operation.

The auxetic structure serves not only as a mechanical interface but also as a strain-amplifying scaffold that enhances the detection sensitivity of the tactile layer. This synergy allows the system to accurately interpret tactile events—including pressure magnitude, contact location, and stimulus type—at high temporal and spatial resolutions. The real-time signal processing pipeline employs adaptive noise filtering, centroid-based localization algorithms, and convolutional neural networks (CNNs) to classify tactile interactions, even in the presence of dynamic and unpredictable user inputs.

Experimental evaluations in HRI scenarios demonstrate that the system can detect tactile forces as low as 1 N, with a localization error below 2.5 mm and a latency under 15 ms. It exhibits consistent performance under repeated human touch interactions, confirming its robustness and durability in continuous-use applications.

This platform is particularly suited for soft robotic systems, robotic prosthetics, and socially interactive service robots where tactile perception is essential for safe, responsive, and intuitive engagement with humans. By integrating auxetic mechanics with bio inspired sensor design, our work contributes a novel approach to achieving perceptual intelligence in robotic systems, bridging the gap between artificial tactile sensing and human-like touch perception.