Associative dynamics and its control of chaotic neural network

In this paper, the Nagumo-Sato model is used to construct a chaotic neural network (CNN). Each reference sample is stored in the chaos attractor that is formed by the associative dynamics. When the inputs sometimes deviate obviously from its original attractive region and the correct association is not realized by itself,the feedback pinning is use to control the associative dynamics. The lost original memory will be retrieved quickly considering the fact that the CNN is a spatiotemporal system. The simulation experiments of both the associative dynamics and the retrieval process are done for the faults of broken rotor bars of an induction motor. The results show that the feedback pinning control is a simple and effective control method to the CNN.