Post-FEC performance evaluation of optical SDM systems with mode-dependent loss

Area of Science:

• Optical communication systems
• Digital signal processing
• Information theory

Background:

• Spatial Division Multiplexing (SDM) systems are crucial for increasing optical network capacity.
• Mode-dependent loss (MDL) in SDM channels introduces random bit error rates (BER), impacting system design and forward-error correction (FEC) strategies.
• Traditional methods for calculating post-FEC BER are inadequate for MDL-impaired systems due to the random nature of pre-FEC BER.

Purpose of the Study:

• To analyze the bit error rate (BER) performance of spatial division multiplexing (SDM) systems affected by mode-dependent loss (MDL).
• To evaluate two distinct approaches for applying low-density parity check (LDPC) forward-error correction (FEC) in MDL-impaired SDM systems with optimal MIMO receivers.
• To determine the most effective LDPC-FEC application strategy for mitigating MDL impairments.

Main Methods:

• Numerical simulations were employed to analyze post-FEC BER performance.
• Two LDPC-FEC application schemes were simulated: independent application per mode and joint application across all modes.
• The study considered an optimal linear multiple-input multiple-output (MIMO) receiver architecture.
• Investigated simplifications in log-likelihood ratio (LLR) computations for efficiency.

Main Results:

• The probability density function (PDF) and coding gain of the pre-FEC BER are dependent on the signal-to-noise ratio (SNR) in MDL-impaired SDM systems.
• Applying LDPC-FEC independently to each spatial mode demonstrated superior performance compared to applying it jointly across all modes.
• The first LDPC-FEC application strategy significantly reduced the post-FEC BER.

Conclusions:

• Independent application of LDPC-FEC per mode is the optimal strategy for mitigating MDL impairments in SDM systems.
• The findings provide crucial insights for designing robust SDM systems and selecting appropriate FEC schemes.
• Further research into simplified LLR computations can enhance the practical implementation of these systems.