3D printing of self-powered triboelectric intelligent system with enhanced output performance for material identification via machine learning

Triboelectric nanogenerators (TENGs) have attracted global attention for energy conversion from various types of mechanical energy into electricity. Traditional approaches for improving triboelectric conversion efficiency generally involved complicated preparation methods and strict structural design, which cannot fulfill the stringent requirements of material adaptability and application scenarios. In this work, a flexible TENG was developed based on zirconium-doped barium titanate (BaZr x Ti 1−x O 3 , BZT) particles in polydimethylsiloxane (PDMS) matrix. Subsequently, BZT/PDMS composites with patterned structures were fabricated by direct ink writing (DIW) based three-dimensional (3D) printing to significantly enhance the electrical performance. It was demonstrated that the TENG could deliver a maximum open-circuit voltage (V OC ) of 350 V and a short-circuit current (I SC ) of 4.9 µA with a maximum power density of 2.4 W m −2 . Moreover, an effective utilization of the TENG was designed to detect the presence of humans, saving a huge amount of electrical energy. Finally, a TENG-based tactile sensor was integrated with machine learning to construct a material identification system, which was capable of recognizing different materials with a high accuracy of 98.54 %. The TENG-based tactile sensor with machine learning demonstrated a tremendous application in material perception, promoting the development of versatile sensors in the field of machine intelligence.