Cold-start of coherent optical receivers with decision-aided maximum likelihood phase estimation scheme

In coherent optical communication systems adopting decision-aided maximum likelihood (DA-ML) phase estimation, receiver initialization requires transmission of a training sequence followed by data symbols. However, such a start-up scheme is infeasible under the scenarios like burst-mode transmission and multi-receiver scheme. We propose a novel cold-start concept for DA-ML receiver without using a training sequence. By adopting the Mth-power scheme, initial data symbols are estimated and further utilized to start up the DA-ML receiver effectively. Extensive numerical simulations are conducted under different situations for comparison. Results verify the feasibility of the proposed cold-start scheme. Based on conventional DA-ML, adopting cold-start can effectively start up the DA-ML receiver as well as achieve comparative bit error rate (BER) performance under burst-mode transmission and multi-receiver scenarios, with good robustness and nearly the same system complexity.