A bioinspired pneumatic finger pad sensor for simultaneous recognition of coupled tactile features

Mimicking the human sense of touch is essential for endowing robots with the capability to interact intelligently with their surroundings. However, effectively recognizing the inherently coupled features of objects (e.g., stiffness and shape) remains a significant challenge, as traditional sensors often struggle with signal ambiguity under complex contact conditions. Inspired by the multilayered structures of the human finger pad, this study developed a novel bioinspired pneumatic finger pad sensor (PFPS). The PFPS features a vertically integrated design with a pneumatic cavity that provides mechanical compliance for surface adaptation. Crucially, it achieves stimulus detection via a top resistive layer and a bottom capacitive layer that simultaneously achieve local strain sensing and global pressure monitoring, respectively. By fusing these complementary modalities, the PFPS effectively decouples mixed mechanical stimuli. A robust tactile sensing system was established by integrating the PFPS with machine learning algorithms, thereby achieving high-precision recognition for different coupled features, including moving direction and shape (100%), stiffness and shape (95.5%), and contact conditions (94.4%). This work offers a structurally integrated solution for resolving signal ambiguity in tactile perception, significantly enhancing the capability of robots to understand and manipulate their surroundings.