Turbo Equalization with Parallel BCJR Segmentation for Optical Throughput

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

High-speed optical fiber transmission systems face limitations in throughput due to nonlinear effects, polarization mode dispersion, and differential coding, which current Turbo equalization methods struggle to adapt to, especially at speeds greater than 100 Gbit/s.

Innovation Solution

Implementing a Turbo equalization compensation method that involves segmenting data blocks into overlapping segments for concurrent forward and backward recursive processing, reducing storage needs and improving system throughput by using an OP-BCJR unit and LDPC convolutional code decoding units, with channel estimation to determine transition probability distribution parameters.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If a common sliding window BCJR with serial structure is used, then the Turbo equalizer can implement Turbo equalization compensation, but the system throughput is limited

Engineering Contradiction:
Improvesystem throughputVSAvoidBCJR module structure
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent divides the data block into multiple parallel processing windows, allowing concurrent processing of multiple data segments simultaneously. This segmentation transforms the serial BCJR structure into a parallel architecture, directly increasing system throughput while maintaining the Turbo equalization compensation functionality

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces parallelism as a new dimension by processing multiple windows simultaneously rather than sequentially. This dimensional change from serial to parallel processing architecture enables the system to achieve higher throughput without proportionally increasing device complexity

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Reliability

If the entire LDPC code word is stored in the BCJR module, then complete Turbo equalization can be performed, but the BCJR module requires numerous storage resources

Engineering Contradiction:
ImproveTurbo equalization compensationVSAvoidstorage resources
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent segments the LDPC code word into multiple windows that are processed in parallel. Each window contains only the portion of data needed for that specific processing stage, eliminating the need to store the entire code word in the BCJR module. This reduces storage resource requirements while maintaining the reliability of Turbo equalization compensation

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary segmentation of the LDPC code word into manageable windows before processing. This allows the system to load and process only the necessary data portions incrementally, rather than requiring all storage resources to be allocated at once, thereby reducing the overall storage burden on the BCJR module

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3001570B1Turbo equalization method and turbo equalization system
Publication Date: 2018.06.06 HUAWEI TECH CO LTD
  • EP3001570B1 patent drawingFigure 1~3
  • EP3001570B1 patent drawingFigure 4~6
  • EP3001570B1 patent drawingFigure 7

AI summary

Embodiments of the present invention relate to a method for implementing Turbo equalization compensation, an equalizer, and a system. The method for implementing Turbo equalization compensation includes: dividing a first data block into n data segments, where D bits in two adjacent data segments in the n data segments overlap, performing recursive processing on each data segment in the n data segments, and merging the n data segments on which the recursive processing has been performed, to obtain a second data block; and performing iterative decoding on the second data block, to output a third data block, where data lengths of the first data block, the second data block, and the third data block are all 1/T of a code length of a LDPC convolutional code. The embodiments of the present invention are applied to a receive end of a high-speed optical fiber transmission system. By performing, in an OP-BCJR unit, segmentation processing and forward and backward recursive operations on a received data block, and performing, in an LDPC convolutional code decoding unit, Turbo iterative processing on data obtained from the OP-BCJR unit, a system throughput can be effectively improved.