A Smart Finger for Soft Material Identification Based on a Multimodal Tactile Sensor

Existing artificial tactile sensing technologies face significant challenges in addressing signal instability caused by the complex mechanical responses of soft materials. Here, a multimodal smart tactile finger based on a triboelectric nanogenerator (TENG) and a piezoelectric nanogenerator (PENG) is developed. The smart finger system integrates a stable contact mechanism with a multimodal sensing module, enabling the synchronous acquisition of composite responses from four TENG channels and one PENG channel. For different soft materials, the TENG and PENG signals generated during the detection process are significantly different. The former is generated by the contact and separation of the TENG module and the soft materials, while the latter is generated by the elastic collision between the PENG module and the soft materials. The acquired signals are standardized and processed using a linear discriminant analysis (LDA) model to achieve accurate classification of soft material types. Testing with 23 soft materials covering broad ranges of modulus and adhesion strength demonstrates that the classification accuracy with all five channels exceeds 83%, significantly outperforming results from any single sensor channel. This smart tactile finger offers promising technological prospects for advanced human-machine interaction, bioinspired robotics, and medical tactile systems.