Real-time data visual monitoring of triboelectric nanogenerators enabled by Deep learning

Abstract

The rapid advancement of smart sensors and logic algorithms has propelled the widespread adoption of the Internet of Things (IoT) and expedited the advent of the intelligent era. The integration of triboelectric nanogenerator (TENG) sensors with Deep learning (DL) leverages unique advantages of TENG such as self-powered sensing, high sensitivity, and broad applicability, along with DL's robust data processing capabilities to effectively, efficiently, and visually monitor various relevant signals. This amalgamation exhibits significantly superior sensing performance and immense developmental potential, finding extensive utility in domains like smart homes, healthcare system, environmental monitoring, among others. Currently, the synergistic working principle of integrating these two technologies remains insufficiently elucidated. This review presents a comprehensive overview of cutting-edge DL techniques and related research aimed at enhancing real-time visual monitoring of TENG. Specifically, it focuses on DL algorithms such as Convolutional Neural Networks (CNN), Recurrent Neural Networks (RNN), and Long Short-Term Memory (LSTM) for processing intricate TENG-generated datasets. Furthermore, this review outlines the advantages and synergistic mechanisms resulting from the integration of DL with TENG sensors, providing a comprehensive summary of their latest applications in various fields requiring real-time data visual monitoring. Finally, it analyzes the prospects, challenges, and countermeasures associated with the integrated development of TENG and DL while offering a comprehensive theoretical foundation and practical guidance for future advancements in this field.