Dig information of nanogenerators by machine learning

Nanogenerators (NGs) are one of the promising energy solutions, which collect different forms of energy in the environment, and have great potential applications in self-powered systems, the Internet of Things, and blue energy. NGs are commonly applied as power supplies or transducers. To be designed into an outstanding power supply, the output performance should be evaluated by a standardized indicator series. Despite the existing standard, the output performance evaluations in the field are still using a series of indicators such as charge density, current density, output power density, and root mean square power density. Meanwhile, the non-linear time-strain-electric signal relation requires an up-to-date quantitative method. After more than half a century of development, machine learning (ML) has been able to adapt to the research needs of many disciplines. ML was proven applicable for multi-index evaluation, non-linear high-dimensional information processing, and distributed data analysis. In this article, interdisciplinary connection is established. The current works can be divided into eight categories: (1) Equipment optimization and fault diagnosis, (2) Information security protection, (3) Intelligent devices for sports and healthcare, (4) Human-machine interaction, (5) Object recognition, (6) Intelligent transportation system, (7) Intelligent Sound System and (8) Environmental protection. According to the existing literature, principal component analysis (PCA), random forest (RF), support vector machine (SVM), and a variety of deep learning (DL) and related algorithms are used, producing better numerical results than traditional algorithms in certain scenarios. Based on the analysis of the current reports, the development trends and challenges of NGs and ML are discussed. In the future, much more comprehensive application of ML and NGs will certainly cause great repercussions in the research field.

Automated summary

Nanogenerators (NGs) are a promising energy solution for self-powered systems, IoT, and blue energy, and their output performance can be evaluated using standardized indicators. Machine learning (ML) has been proven applicable for multi-index evaluation and distributed data analysis. Current research on NGs and ML includes equipment optimization, information security, intelligent devices, human-machine interaction, object recognition, intelligent transportation, intelligent sound systems, and environmental protection. Advanced algorithms like PCA, RF, SVM, and deep learning show better results than traditional algorithms. The development trends and challenges of NGs and ML are discussed, indicating their future extensive applications in the research field.