A smart finger for robotic tactile sensing of surface/subsurface patterns based on a high-density piezoresistive sensor array.

The human finger, with its high concentration of sensory receptors, excels at sensing both surface patterns and subsurface properties within soft tissue. However, replicating this dual capability in artificial systems poses significant challenges. This study presents a smart finger system based on a high-density piezoresistive sensor array, which demonstrates high sensitivity, fast response, and the ability to recognize both surface and subsurface patterns. The smart finger system integrates a flexible high-density piezoresistive sensor array (PRSA), a miniaturized circuit board for collecting distributed pressure signals, and convolutional neural network algorithms. The enhanced performance is attributed to the cross-striped nanocarbon-polymer active material, which improves sensor sensitivity and stability. Additionally, machine learning algorithms, particularly convolutional neural networks, are employed to process the tactile data and improve pattern recognition, allowing for advanced tactile perception in robotic applications. Characterization test results indicate that our fabricated PRSA possesses a 32 × 32 pixels, adjacent pixel spacing of only 0.6 mm, and high flexibility. The smart finger demonstrates impressive performance, with high sensitivity (10.69 mV/kPa), low fluctuation, long-term durability, an ultra-fast response time of approximately 3 ms, and a two-point threshold of 1.8 mm, surpassing human fingertip capabilities. We showcase applications of the smart finger system in robot-assisted tactile recognition of both surface and subsurface patterns. Experimental results indicate that the smart finger system recognizes surface patterns, such as embossed letters, with significantly higher accuracy than human touch (95.5 % vs. 26.9 %) and effectively captures subsurface patterns with varying softness. This innovative smart finger holds substantial promise for advancing robotic tactile sensing technologies.