Measurement of linewidth enhancement factor based on self-mixing interferometry and back propagation neural network

Line-width enhancement factor (α) is a fundamental parameter of semiconductor lasers (SLs). In this paper, we propose a method for measuring α of SLs. The method is based on back-propagations neural network (BPNN) for all feedback regimes. MATLAB was used to carry out the numerical calculations and simulations of the BPNN. We used the training set and the test set to train the prediction model, and then used the predictive model to output the predicated value. The results of the BPNN model showed that the R2 value was 0.99994, and the results were following the requirement model. The accuracy of the method has been confirmed and tested by computer simulations, which show that the method can estimate α with a relative error less than 2.5%.