No abstract

No abstract

We propose a deep learning approach based on an autoencoder for identifying and localizing fiber faults in passive optical networks. The experimental results show that the proposed method detects faults with 97% accuracy, pinpoints them with an RMSE of 0.18 m and outperforms conventional techniques.

No abstract

A novel approach based on an artificial neural network (ANN) for lifetime prediction of 1.55 µm InGaAsP MQW-DFB laser diodes is presented. It outperforms the conventional lifetime projection using accelerated aging tests. © 2020 The Author(s) OCIS codes: (140.0140) Lasers and laser optics; (200.4260) Neural Networks.