Feedforward equalizer with selective noise decorrelation for bandwidth-limited signal

Area of Science:

• Optical communication systems
• Signal processing

Background:

• High-speed optical transmission systems face performance limitations due to transceiver bandwidth constraints.
• Linear equalization combats inter-symbol interference (ISI) but amplifies high-frequency noise, degrading receiver performance.
• Conventional noise whitening methods often require computationally intensive maximum likelihood sequence estimation (MLSE) equalizers.

Purpose of the Study:

• To introduce a novel selective noise decorrelation (SND) equalizer for high-speed optical transmission.
• To address the challenge of noise enhancement caused by linear equalization without resorting to complex MLSE techniques.
• To offer a computationally efficient alternative for mitigating noise in optical receivers.

Main Methods:

• A selective noise decorrelation (SND) equalizer is proposed, operating in a feedforward manner after a linear equalizer.
• The SND equalizer predicts symbol error likelihoods based on noise patterns.
• It applies weighted averaging of neighboring noise for decorrelation selectively, only when necessary.

Main Results:

• The SND equalizer effectively corrects errors introduced by bandwidth limitations and noise enhancement.
• Experimental results demonstrate superior performance compared to feedforward and decision feedback equalizers.
• The proposed SND equalizer achieves performance comparable to MLSE-based approaches with significantly reduced computational complexity.

Conclusions:

• The SND equalizer presents a viable and efficient solution for mitigating noise enhancement in optical systems.
• It offers a practical alternative to complex MLSE equalization for cost-sensitive applications.
• The SND equalizer requires only 6.25% of the computational load compared to PF-MLSE equalizers.