Vectoring Processor Full-Duplex Cross-Talk Cancellation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current full-duplex communication systems face challenges in implementing effective cross-talk cancellation and compensation, particularly due to dependencies between pre-coder, post-coder, and NEXT-ECHO cancellation matrices, which can be problematic in large-scale systems, and require precise tone data processing to maintain performance.

Innovation Solution

A vectoring processor is configured to pre-compensate downstream tone data for DS-FEXT and post-compensate upstream tone data for US-FEXT and NEXT using delayed pre-compensated downstream tone data, decoupling the DS-FEXT precoding matrix from the US-FEXT postcoding and NEXT cancellation matrices, allowing independent updates and reducing memory requirements.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If the NEXT-ECHO cancellation matrix depends on the precoder and postcoder matrices, then the cross-talk cancellation can be achieved, but the NEXT-ECHO cancellation matrix must be updated every time the pre- or postcoder matrix is updated, increasing system complexity

Engineering Contradiction:
Improvecross-talk cancellation effectivenessVSAvoidmatrix update complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent segments the cross-talk cancellation function into separate components: the precoder matrix handles DS-FEXT compensation, the postcoder matrix handles US-FEXT compensation, and the NEXT-ECHO cancellation matrix operates independently. This segmentation allows each matrix to be updated independently without requiring coordinated updates of all matrices, thereby reducing system complexity while maintaining cancellation effectiveness.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary compensation of DS-FEXT at the transmitter side using the precoder matrix before signals are transmitted. This preliminary action removes the dependency chain where NEXT-ECHO cancellation would need to account for precoder changes, as the precoder's effect on FEXT is already handled separately. This allows the NEXT-ECHO cancellation matrix to be updated independently.

Inventive Principle:
Principle #10Preliminary action

2Reliability

If vectoring processor performs both pre-compensation of DS-FEXT and post-compensation of US-FEXT and NEXT, then signal quality is improved, but the processing complexity and memory requirements increase

Engineering Contradiction:
Improvesignal qualityVSAvoidprocessing complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent extracts the DS-FEXT pre-compensation function into a separate precoding stage that operates independently from the post-compensation stage. By taking out the DS-FEXT compensation from the main post-processing chain, the vectoring processor can handle US-FEXT and NEXT cancellation with reduced complexity, as the DS-FEXT component is already removed from the signal path before post-processing.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent performs preliminary DS-FEXT compensation at the transmitter using the precoder matrix before signals are sent over the transmission lines. This preliminary action reduces the burden on the vectoring processor at the receiver side, as it no longer needs to compensate for DS-FEXT in the post-processing stage, thereby reducing overall processing complexity while maintaining signal quality.

Inventive Principle:
Principle #10Preliminary action

3Reliability

If precise tone data processing is performed for full-duplex communication, then communication performance is maintained, but memory requirements and processing load increase

Engineering Contradiction:
Improvecommunication performanceVSAvoidmemory requirements
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

The patent segments the tone data processing into distinct functional blocks: precoding for DS-FEXT compensation, postcoding for US-FEXT compensation, and separate NEXT-ECHO cancellation. Each segment processes only the specific interference type it is designed for, reducing the overall memory requirements compared to a unified processing approach that would need to store and process all interference components simultaneously.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3503419B1Full-duplex vector processing
Publication Date: 2021.06.30 NOKIA SOLUTIONS & NETWORKS OY
  • EP3503419B1 patent drawingFigure 1
  • EP3503419B1 patent drawingFigure 2
  • EP3503419B1 patent drawingFigure 3

AI summary

A vectoring processor (150) for mitigating cross-talk (121, 122, 123) is disclosed. The vectoring processor is configured to interface with a plurality of communication units (110) supporting full-duplex bi-directional communication over respective transmission lines, and to to pre-compensate downstream tone data (140) received from the communication units for downstream far end cross-talk, DS-FEXT (121) thereby obtaining pre-compensated downstream tone data (141), and to supply the pre-compensated downstream tone data to the communication units (110) for further transmission over the respective transmission lines, and to receive from the communication units upstream tone data (143, 144) and associated delayed pre-compensated downstream tone data (142) previously supplied by the vectoring processor, and to post-compensate the upstream tone data, with the delayed pre-compensated downstream tone data, for near end cross-talk, NEXT (123), and for upstream far end cross-talk, US-FEXT (122), thereby obtaining post-compensated upstream tone data (145), and to supply the post-compensated upstream tone data to the communication units for further demodulation.