Deep learning assisted mat-type triboelectric nanogenerator for personnel monitoring

With the development of sensing technology and artificial intelligence, the smart system has attracted increasing attention in personnel security for the perception and acquisition of various information of personnel. However, high cost, huge computing resources,and personnel privacy concerns restrict the traditional camera-based monitoring system applied in smart buildings. Here, a triboelectric smart mat system based on deep learning (DL) for personnel status monitoring, identification, and positioning is proposed. The smart mat unit comprises the fluorinated ethylene propylene (FEP) membrane and flexible conductive sponge electrodes with different filling rates. Meanwhile, the special connection of the smart mat unit leads to a higher resolution. The minimal structure smart mat system combines real-time sensory data analysis to extract gait-related information, allowing accurate and comprehensive personnel perception. Thus, this smart mat system develops a new field for researching artificial intelligence-assisted Internet of Things (AIoT) and smart buildings.