Optical Receiver Likelihood Tables for Lower-Scale Soft Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for likelihood generation in optical transmission systems, such as 6b4D-2A8PSK, assume ideal soft decision error correction without considering circuit implementation, leading to increased signal processing circuit scale and implementation difficulties.
Innovation Solution
A likelihood generation device that receives code-modulated signals with n-k information bits and k parity bits in an N-dimensional arrangement, using L-dimensional address spaces to determine and update likelihoods for information and parity bits, where L<N, to reduce circuit scale by dividing the signal processing into smaller tables.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If ideal soft decision error correction is assumed without considering circuit implementation, then error correction performance is improved, but circuit scale increases and implementation becomes difficult
Solution Approach 1:
The patent segments the N-dimensional likelihood generation into multiple L-dimensional tables where L < N. Each table handles a subset of the dimensions, and the results are combined to produce the final likelihood values. This segmentation reduces the complexity of individual tables while maintaining overall error correction performance through coordinated processing across multiple tables.
2Productivity
If N-dimensional modulation is used to increase transmission capacity, then spectral efficiency is improved, but the number of constellation points increases leading to larger circuit scale
Solution Approach 1:
The patent divides the N-dimensional signal space into multiple L-dimensional subspaces (where L < N). Each subspace is processed by a separate table that generates likelihood values for a subset of bits. This segmentation allows high-order modulation to be implemented without requiring a single large N-dimensional table, thereby reducing circuit scale while maintaining transmission capacity.
Solution Approach 2:
The patent transforms the problem from generating N-dimensional likelihoods directly to generating L-dimensional likelihoods in multiple separate tables, then combining them. This dimensional decomposition converts a high-dimensional complex problem into multiple lower-dimensional problems that are easier to implement in circuit form.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
A likelihood generation device according to the present invention is included in a receiving apparatus that receives a code-modulated signal including n-k information bits and k parity bits in N-dimensional arrangement, where N is a natural number greater than or equal to four, and n and k are natural numbers with k<n. The likelihood generation device includes: a temporary likelihood determination unit (51a and 51b) to determine, by using a table that includes an L-dimensional address space and stores a likelihood in each of the L-dimensional address spaces, likelihoods of an information bit and a parity bit that are transmitted by the code-modulated signal received, where L is a natural number with L<N; and a likelihood correction unit (52) to update, among the likelihood information items determined by the likelihood derivation unit, the likelihood of the information bit on a basis of a rule for generation of the parity bit.