Voltages prediction algorithm based on LSTM recurrent neural network

Abstract Because of the characteristic of delay of adaptive optics system, the performance of adaptive optics system can be improved effectively by predicting the deformation mirror voltages. This paper presents a nonlinear voltages prediction algorithm based on LSTM recurrent neural network. It is found that the predicted voltages by LSTM neural network are consistent with the actual voltages. By comparing the prediction results of the proposed algorithm with those of other prediction algorithms, the prediction algorithm based on LSTM has higher prediction accuracy. This algorithm has strong applicability and higher accuracy in adaptive optics voltages prediction.