arrow
Return

Conformal Tactile Sensor Array-Based Hemispherical Detector for Robotic Spatial Exploration

delete2026-08-24
delete0
PRE
AI
W
Weijie Liu
S
Shihang Wang
P
Pan Chen
J
Jie Jin
梅
梅德庆 (Deqing Mei)
Y
Yancheng Wang
DOI:10.1109/tase.2026.3726432delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
Tactile perception plays a critical role in spatial exploration for robotics when visual sensing is limited by extreme lighting, occlusion, and dust. Conventional tactile sensors, constrained by their planar substrates, cannot be conformally mounted onto curved surfaces commonly used in robotic probes, thus being unable to accurately capture spatially distributed tactile information. In this work, we present a fully conformal flexible tactile sensor array based on a surface-unfolding design, which can be integrated onto a hemispherical detector for spatial surface morphology exploration. Building upon polyhedral discretization and geometry-driven unfolding principles, the unfolded geometry was translated into a functional multilayer tactile sensing array with regularized sensing units and narrow-bridge interconnects for curved-surface integration. The conformal tactile sensor array was mounted onto a probe to construct a hemispherical detector and integrated with a robotic manipulator to conduct robot-assisted tactile exploration along a prescribed trajectory, thereby evaluating the complete sensing-to-reconstruction pipeline. The results showed that the tested continuous freeform surface profile could be reconstructed using the measured tactile signals and robotic kinematic information, achieving an average reconstruction error of 1.81 mm and a maximum error below 3.50 mm. The proposed conformal tactile sensor array-based detector demonstrates a complete pathway from curved-surface tactile array design and distributed sensing to robotic morphology reconstruction, providing a practical tactile perception interface for robotic exploration under visually constrained conditions. Note to Practitioners—This work is motivated by the need for robotic systems to perceive and explore physical environments when visual sensing becomes unreliable, such as when lighting, dust, or occlusion limits camera-based perception. Under such conditions, tactile sensing provides a direct means of obtaining contact and surface-geometry information through physical interaction. However, many existing tactile sensors are fabricated on rigid or planar substrates and are therefore difficult to conformally attach to the curved probe surfaces commonly used in robotic end-effectors. This geometric mismatch limits sensing coverage and may reduce the completeness of tactile information acquired during surface exploration. To address this challenge, we developed a conformal flexible tactile sensing array that can be integrated onto a curved carrier, as demonstrated using a hemispherical detector. Building upon established surface-unfolding principles, the target curved surface is discretized into planar segments that are implemented as flexible sensing units connected through narrow bridges. This configuration allows the array to follow the hemispherical carrier while maintaining distributed tactile sensing and electrical interconnection. When integrated with a robotic manipulator, the detector can perform tactile scanning along a predefined trajectory and reconstruct a continuous surface profile by combining tactile responses with robotic kinematic information. The present system establishes a practical sensing-to-reconstruction pathway in which a conformal hemispherical detector acquires distributed contact information that is subsequently combined with robotic kinematics for surface morphology reconstruction. From an engineering perspective, this work demonstrates a practical route for translating conformal geometric design into a curved tactile sensing interface. Similar design principles may be applicable to other curved robotic interfaces, such as soft grippers, prosthetic devices, and surgical probes, although redesign and recalibration would be required for different geometries and mounting conditions. Current limitations include manual layer alignment, screen printing, wiring complexity for high-density arrays, curvature-dependent sensing characteristics. uture efforts will focus on automated fabrication, integrated readout electronics, curvature-aware calibration, computational acceleration, and adaptive tactile sampling.
Keywords:
Tactile sensors
spatial exploration
robotic detection
flexible electronics
tactile exploration

Journal

IEEE Transactions on Automation Science and Engineering cover
IEEE Transactions on Automation Science and Engineering
IF:
6.4
Papers:
5.1K
Citations:
1.6W

Organization

Z
zhejiang university
Scholars:
6.3K
Papers: 1.8K
Citations: 0
Cited Papers

Cited Papers

No cited papers available