Eye-width and Eye-height Estimation Method based on Artificial Neural Network (ANN) for USB 3.0

Abstract

As technology develops, the amount of data used increases exponentially and the operating speed of electronic devices increases accordingly. As the operating speed increases, the transmission speed of data exchanged through the channel also increases. To get an eye-diagram, a time-domain simulation is required, which consumes a lot of time and computer power. However, a channel simulation in frequency-domain is fast and it consumes less computer power. Therefore, using the frequency-domain simulation data is more efficient than the time-domain simulation. This paper proposes an eye-width and eye-height estimation method using an artificial neural network (ANN). The input of the ANN is insertion loss and the outputs of the ANN are eye-width and eye-height. Finally, the performance of the proposed method is verified by transient eye-diagram simulations with arbitrarily-selected channel parameters.