Mitigation of fiber nonlinearities in multilevel coded-modulation schemes
View Patent ↗A receiver and method are provided for mitigation of finer non-linearities in multilevel coded-modulation schemes. The receiver includes a multilevel Bahl-Cocke-Jelinek-Raviv (BCJR) equalizer configured to receive channel samples of an input signal, partially mitigate fiber non-linearities, and provide symbol reliabilities. The receiver further includes a bit probabilities module coupled to the BCJR equalizer configured to calculate soft bit reliabilities from the symbol reliabilities. The receiver also includes one or more low-density parity-check (LDPC) decoders coupled to the bit probabilities module and the BCJR equalizer, configured to receive the soft bit reliabilities and output code words. The one or more LDPC decoders iteratively provide extrinsic soft information feedback to the BCJR equalizer to compensate for the fiber non-linearities.
1. A receiver, comprising:
a multilevel Bahl-Cocke-Jelinek-Raviv (BCJR) equalizer configured to receive channel samples of an input signal, partially mitigate fiber non-linearities, and provide symbol reliabilities;
a bit probabilities module coupled to said BCJR equalizer configured to calculate soft bit reliabilities from the symbol reliabilities; and
one or more low-density parity-check (LDPC) decoders coupled to said bit probabilities module and said BCJR equalizer, configured to receive the soft bit reliabilities and output code words, said one or more LDPC decoders iteratively providing extrinsic soft information feedback to said BCJR equalizer to compensate for the fiber non-linearities.
2. The receiver of claim 1 , wherein the soft bit reliabilities are in a form of log-likelihood ratios (LLRs).
3. The receiver of claim 1 , wherein the one or more LDPC decoders stop iterating if a valid code word is reached or a number of iterations is exceeded.
4. The receiver of claim 1 , wherein said BCJR equalizer is maximum a posteriori probability (MAP) equalizer.
5. The receiver of claim 1 , wherein said BCJR equalizer operates on a discrete dynamical trellis description of an optical channel upon which the input sequences are received.
6. The receiver of claim 5 , wherein the discrete dynamical trellis description comprises a triplet: a previous state, a next state, and a channel output.
7. The receiver of claim 5 , wherein the discrete dynamical trellis description is based on at least one of one or more histograms and an instanton-Edgeworth expansion method.
8. The receiver of claim 1 , wherein said one or more LDPC decoders are based on a min-sum-with-correction-term algorithm.
9. The receiver of claim 1 , wherein said BCJR equalizer is configured to perform turbo equalization.
10. The receiver of claim 1 , wherein said BCJR equalizer is configured to simultaneously mitigate the fiber non-linearities and linear channel impairments relating to intra-channel four-wave mixing (IFWM), intra-channel cross-phase modulation (IXPM), nonlinear phase noise, self-phase modulation (SPM), residual chromatic dispersion, and polarization-mode dispersion (PMD).
11. The receiver of claim 1 , wherein said BCJR equalizer is configured to decode different convolutional and block codes and as a maximum a posteriori probability (MAP) detector.
12. The receiver of claim 1 , further comprising a coherent detector coupled to said BCJR equalizer for initially detecting the input signal.
13. A method, comprising:
receiving and sampling an input signal using a sampler;
partially mitigating fiber non-linearities using a multilevel Bahl-Cocke-Jelinek-Raviv (BCJR) equalizer configured to receive channel samples of the input signal and provide symbol reliabilities;
calculating soft bit reliabilities from the symbol reliabilities;
decoding the input signal to output code words using one or more low-density parity-check (LDPC) decoders based on the soft bit reliabilities; and
iteratively feeding back extrinsic soft information to the BCJR equalizer to compensate for the fiber non-linearities.
14. The method of claim 13 , wherein the soft hit reliabilities are in a form of log-likelihood ratios (LLRs).
15. The method of claim 13 , wherein the one or more LDPC decoders stop iterating if a valid code word is reached or a number of iterations is exceeded.
16. The method of claim 13 , wherein said BCJR equalizer is maximum a posteriori probability (MAP) equalizer.
17. The method of claim 13 , wherein said BCJR equalizer operates on a discrete dynamical trellis description of an optical channel upon which the input sequences are received.
18. The method of claim 17 , wherein the discrete dynamical trellis description comprises a triplet: a previous state, a next state, and a channel output.
19. The method of claim 17 , wherein the discrete dynamical trellis description is based on at least one of one or more histograms and an instanton-Edgeworth expansion method.
20. The method of claim 13 , wherein said one or more LDPC decoders are based on a min-sum-with-correction-term algorithm.
21. The method of claim 13 , wherein said BCJR equalizer is configured to perform turbo equalization.
22. The method of claim 13 , wherein said BCJR equalizer is configured to simultaneously mitigate the fiber non-linearities and linear channel impairments relating to intra-channel four-wave mixing (IFWM), intra-channel cross-phase modulation (IXPM), nonlinear phase noise, self-phase modulation (SPM), residual chromatic dispersion, and polarization-mode dispersion (PMD).
23. The method of claim 13 , wherein said BCJR equalizer is configured to decode different convolutional and block codes and as a maximum a posteriori probability (MAP) detector.
24. The method of claim 13 , further comprising initially detecting the input signal using a coherent detector.