Starting from the maximum likelihood criterion, we derive a novel blind chromatic dispersion (CD) estimation method in the presence of unknown data, propagation delay, polarization state, and differential group delay. By using CD estimation, electronic dispersion compensation (EDC) can be carried out without prior knowledge of the amount of accumulated CD. This adds flexibility to the EDC, which may prove valuable in reconfigurable optical networks. Using numerical simulations, we compare the suggested algorithm with a well-known CD estimation algorithm based on the constant modulus algorithm. We find that the proposed algorithm has better estimation performance and lower computational complexity. Furthermore, the impact of differential group delay is small. The derivation of the algorithm also shows the close connection between CD estimation, clock recovery, and polarization effects.
© 2012 IEEEPDF Article