Biomimetic Tactile Systems of Electronic Skin: Sensing, Processing, and Feedback

Abstract

Electronic skin (e-skin) represents a transformative technology aimed at replicating the intricate tactile capabilities of human skin, enabling humanoid robots and immersive interfaces to achieve safe, intuitive, and empathetic interactions. Inspired by the complex network of mechanoreceptors in biological skin, e-skin are explored with various materials and structure based on the mechanisms like piezoresistive, capacitive, piezoelectric, and triboelectric-to detect a wide range of stimuli, including pressure, strain, and dynamic forces. However, most of the present works relies on simple biomimetic implementations of individual skin mechanoreceptors. This approach lacks in-depth research into the comprehensive architecture of skin structures and the underlying neural structure for tactile signal transmission, which limits the ability of electronic skin to achieve more complex tactile capabilities and to realize advanced perceptual functions based on tactile sensing. In this thesis, we aim to advance the development of e-skin by replicating the synergistic effects of human mechanoreceptors through innovative multi-unit sensor designs with asymmetric configurations, enabling precise 3D force detection. Furthermore, we focus on constructing hierarchical neuromorphic signal transmission networks within e-skin architectures to facilitate efficient data processing via neural pulse-like signals, thereby supporting advanced functionalities such as active pain and injury sensing. Additionally, we strive to develop on-body e-skin systems with haptic replication capabilities, utilizing shear actuators and programmable drive patterns to selectively stimulate specific mechanoreceptors, achieving naturalistic tactile reproduction for enhanced human-robot interactions and immersive virtual reality applications.

Our research from a soft e-skin with a novel tactile sensor design featuring four asymmetrically arranged sensing units. Drawing inspiration from human mechanoreceptors, this system decodes force direction, magnitude, and type from 3D forces simultaneously. The sensing units use constrain strips to reduce noise, achieving a high gauge factor of up to 1492, with response and recovery times of 127 ms and 117 ms, respectively. Tested on flat and curved surfaces, including a 5×5 array and a hand-shaped array, the e-skin accurately tracks dynamic forces and distinguishes normal and shear force distributions. This innovation holds promise for human-machine interfaces, intelligent robots, and VR/AR applications.

Next, the tactile signal transmission network also plays a great role in optimizing the electronics robotic sensing ability. Building on the e-skin’s sensing ability, we present a neuromorphic robotic e-skin (NRE-skin) with a hierarchical, neural-like receptive field structure. This design enables basic pressure sensing with a resolution of 10 mN and advanced features like active pain and injury detection. The NRE-skin converts tactile data into pulse trains, reducing data dimensionality while retaining key information, and consolidates signals into a single channel. It includes an active pain-sensing mechanism that triggers local motor responses to excessive stimuli, mimicking a human reflex arc, and supports wound detection to locate and alert for injuries, with a Lego-like design for easy replacement. This enhances robotic perception and supports empathetic human-robot interactions, ideal for household assistants.

To further enhance the tactile experience and bridge the gap between robotic and human interactions, we developed a wearable shear actuator array delivering shear forces up to 2 N. This system replicates sensations like gentle touch, caress, squeeze, and stretch, surpassing traditional vertical or rotational actuators. User studies confirmed its superior realism and tactile fidelity, making it a strong candidate for robotic remote tactile feedback systems.

In conclusion, this thesis significantly contributes to the advancement of e-skin technology by developing a 3D force-sensing e-skin system, a neuromorphic signal transmission network capable of detecting pain and injury, and a wearable shear actuator array for immersive tactile feedback. These innovations collectively enhance e-skin’s capabilities in tactile perception and reproduction, addressing critical challenges in biomimetic design. We believe that these contributions, though incremental, will accelerate the practical application of e-skin technology, ultimately fostering the development of humanoid robots with improved tactile sensitivity and interaction finesse, enabling them to serve humanity in a more intuitive and effective manner.

Date of Award	18 Sept 2025
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Xinge YU (Supervisor)