Fault diagnosis for manipulators based on NeuCube

In this paper, a new fault diagnosis approach for manipulators based on spiking neural network is investigated. The newly proposed evolving spiking model is named NeuCube, it can be employed for classification, pattern recognition, and other kinds of problems. As NeuCube is good at processing spatio-temporal data, we apply it to fault diagnosis for manipulators. And receive better results than other traditional methods. This paper analyses the basic concepts of the spiking neural network and several spiking neuron models existed, then introduced the spiking neuron models, such as NeuCube, in detail about the structure and feature of the model. Finally, after experiment concluded the characteristics and effectiveness of NeuCube.