Application of neural network in sliding mode control

In this paper, a combination of neural network with sliding mode control (SMC) is proposed. As a result, the chattering is eliminated and error performance of SMC is improved. In such an approach two parallel neural networks (NNs) are proposed to realize the SMC. The equivalent control and the corrective control term of SMC are the outputs of the NNs. The training algorithms applied to NNs are based on the SMC equation with a gradient descent method to minimize the control and chattering while optimizing the error performance. In this paper, a sliding mode neurocontroller in power systems is proposed, and experimental results are presented. Two parallel NNs are used to realize the neuro-SMC. To increase the first neural network structure flexibility hidden layer neuron pruning and node splitting algorithms are considered.