A Review of Spiking Neural Network Research in the Field of Bearing Fault Diagnosis

In view of the bearing fault diagnosis problem, in order to better deal with the temporal data, extract nonlinear features, enhance the model robustness and adaptability, and realize efficient distributed computing, the spiking neural network is introduced to improve the accuracy and efficiency in the process of bearing fault diagnosis. In recent years, after a series of studies, the spiking neural network has developed in supervised learning, unsupervised learning, semi-supervised learning, and some other derived methods, such as reinforcement learning and transfer learning, for its applicable scenarios with different advantages. From the perspective of learning principle, method advantages and practical applications, we introduced several spiking neural network for bearing fault diagnosis, discussed the advantages, disadvantages and applicable scenarios of each learning method, discussed the application prospect of spiking neural network for bearing fault diagnosis, and proposed the spiking neural network for bearing fault diagnosis in future research, so as to provide reference for subsequent research.