Enhancing the Roughness Recognition Accuracy of Tactile Sensors Based on the Triboelectric Material Surface Structure Management Strategy

Currently, triboelectric nanogenerator (TENG)-based tactile sensors demonstrate considerable effectiveness in material type recognition. However, challenges remain regarding the stability of triboelectric materials and low precision when applied to the recognition of material surface roughness. To address these issues, the present study focuses on the preparation of a template of unconventional micro-nanostructure using recycled polylactic acid (PLA) materials. By meticulously designing and fabricating polydimethylsiloxane (PDMS)-based elastic triboelectric materials with well-defined concave-convex structures by the template method, a suitable micro-nanoarchitecture was chosen to enhance the triboelectric effect and improve sensing accuracy. To be precise, enhancing the roughness recognition accuracy of the tactile sensor through a triboelectric material surface structure management strategy was developed. Among the various concave-convex structures, the TENG device with a specific micro-nanostructure surface (roughness: 0.18 μm and maximum depth: 0.78 μm) exhibited the highest output voltage and charge density, reaching values of 4.9 V and 3.9 μC/m², respectively. More importantly, the mapping relationship between the strain field distribution (or signal waveform characteristics) and the surface characteristics of the contacting material becomes clearer and more accurate with the establishment of special micro-nanostructures on the triboelectric material surface, thereby improving the recognition accuracy of the material surface roughness. Hence, by integrating deep machine learning and triboelectric effects, and combining signal collecting, data processing, and display modules, a material sensing system integrating high-precision TENG sensors was developed. It can monitor the different roughnesses of the object surface in real time in the natural environment, approximately 93% (0.027 mm), 78% (0.10 mm), and 82% (0.12 mm). Finally, this strategy can provide robots with richer and more accurate sensing capabilities.