Monitoring I/Q data and pulse carving misalignments in RZ-DQPSK transmitters using a neural network approach

We propose a technique using artificial neural networks (ANNs) to simultaneously identify I/Q data misalignment and data/carver misalignment in both parallel-type and serial-type RZ-DQPSK transmitters. A correlation coefficient of 0.99 is obtained by using a 3-input ANN for the parallel case and a 2-input ANN for the serial case.