Virtual Constellation Points for Phase Noise in Optical Receivers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Phase noise in optical communications systems leads to overlapping detection regions in coherent receivers, causing errors in Maximum Likelihood (ML) decoding, especially in modulation schemes like QPSK that are more susceptible to phase distortion.
Innovation Solution
The implementation of an asymmetrical constellation with virtual constellation points in the receiver, allowing for accurate determination of the most likely transmitted symbol by modeling each symbol as a set of virtual points that reduce overlap and are robust against phase noise, even in high error conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional Maximum Likelihood decoding based on minimum Euclidean Distance is used, then decoding is simple and fast, but detection accuracy deteriorates when phase noise causes overlapping detection regions
Solution Approach 1:
The patent segments the constellation space by introducing virtual constellation points that divide overlapping detection regions into distinct segments. Each received signal is evaluated against multiple virtual points associated with each constellation symbol, allowing the system to handle phase noise by testing multiple hypotheses rather than relying on a single minimum distance calculation.
Solution Approach 2:
The patent changes the decoding parameters by replacing the conventional single-point Euclidean distance metric with a multi-point likelihood evaluation. Instead of finding the minimum distance to one constellation point, the system calculates likelihoods based on distances to multiple virtual constellation points and uses these to determine the most probable transmitted symbol, thereby adapting to phase noise conditions.
2Productivity
If QPSK modulation is used to increase bits per symbol, then spectral efficiency improves, but noise tolerance deteriorates due to smaller Euclidean distance between constellation points
Solution Approach 1:
The patent introduces an additional dimension to the decoding process by considering multiple virtual constellation points for each symbol rather than a single point. This multi-dimensional likelihood evaluation allows QPSK to maintain its spectral efficiency while gaining noise tolerance, as the system can distinguish between symbols even when phase noise causes traditional 2D detection regions to overlap.
Data Source
AI summary
In a coherent receiver of an optical communication system, a method of processing a detected symbol estimate to determine a most likely value of a corresponding transmitted data word, the transmitted data word comprising one or more data bits encoded in a transmitter using a predetermined constellation of at least two symbols. A set of two or more virtual constellation points are define in a decision region corresponding to a possible value of the data word. The detected symbol estimate is processed to find a most likely virtual constellation point given the detected symbol estimate. The most likely value of the corresponding transmitted data word is determined based on the most likely virtual constellation point.