Hybrid Precoding for DSL Crosstalk Management

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current DSL systems face instability and bit error issues due to the coexistence of linear and non-linear CPEs, where existing precoding methods either maximize the rate of one type of CPE at the cost of the other, leading to suboptimal performance and potential system instability.

Innovation Solution

A precoding method that calculates a combined precoding matrix to eliminate crosstalk between linear and non-linear CPEs, performing non-linear precoding on signals for non-linear CPEs and linear precoding on signals for linear CPEs, along with power control to optimize signal transmission across both types of CPEs.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If linear precoding is set for the CO to support linear CPE, then the linear CPE can operate, but the rate of non-linear CPE is reduced to suboptimal levels

Engineering Contradiction:
Improvecompatibility with linear CPEVSAvoidrate of non-linear CPE
Core Design Contradiction:
Adaptability or versatilityVSProductivity

Solution Approach 1:

The precoding process is segmented into two distinct stages: first, non-linear precoding is applied to eliminate crosstalk for non-linear CPEs, and then linear precoding is applied to eliminate remaining crosstalk for linear CPEs. This segmentation allows each precoding method to be optimized for its target CPE type without compromising the other's performance.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Non-linear precoding is performed as a preliminary action before linear precoding. By first eliminating the strong crosstalk components using non-linear methods, the subsequent linear precoding operates on a cleaner signal, improving overall effectiveness for both CPE types.

Inventive Principle:
Principle #10Preliminary action

2Productivity

If non-linear precoding is set for the CO to maximize rate of non-linear CPE, then the rate of non-linear CPE increases to maximum, but the rate and stability of linear CPE are greatly affected

Engineering Contradiction:
Improverate of non-linear CPEVSAvoidstability of linear CPE
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The precoding operations are segmented into non-linear and linear components executed in sequence. The non-linear precoding handles the complex crosstalk suppression for non-linear CPEs, while the linear precoding provides a stable, simplified processing stage that ensures compatibility and stability for linear CPEs.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different precoding qualities are applied locally to different CPE types: non-linear precoding with its complex computational capabilities is applied where needed for non-linear CPEs to maximize their rate, while linear precoding is applied for linear CPEs to ensure stability, with each CPE receiving the appropriate processing quality for its capabilities.

Inventive Principle:
Principle #3Local quality

3Device complexity

If simple precoding methods are used to support both linear and non-linear CPE, then implementation is simple, but bit errors increase and system stability deteriorates

Engineering Contradiction:
Improveprecoding implementation complexityVSAvoidsystem stability
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The precoding system is segmented into two processing stages with clearly defined functions. This segmentation allows the complex non-linear processing to be isolated and optimized for specific CPEs, while the linear processing stage provides a stable foundation for all CPEs, reducing overall system bit errors and improving stability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent merges non-linear and linear precoding methods into a unified two-stage processing framework. This combination leverages the strengths of both approaches: the crosstalk elimination capability of non-linear methods and the computational simplicity and stability of linear methods, achieving improved reliability without excessive complexity.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentEP3236605B1Pre-coding method, device and system
Publication Date: 2019.02.20 HUAWEI TECH CO LTD
  • EP3236605B1 patent drawingFigure 1
  • EP3236605B1 patent drawingFigure 2
  • EP3236605B1 patent drawingFigure 3~4

AI summary

Embodiments of the present invention provide a precoding method, an apparatus, and a system. The precoding method in the present invention includes: obtaining a channel matrix on each subcarrier, where the channel matrix includes channel vectors of all CPEs on the subcarrier, and the CPEs include a linear CPE supporting linear precoding and a non-linear CPE supporting non-linear precoding; obtaining a precoding matrix by means of calculation according to the channel matrix; performing, according to the channel matrix, non-linear precoding on a signal to be sent to the non-linear CPE; and performing linear precoding on to-be-sent signals according to the precoding matrix, where the to-be-sent signals include a signal to be sent to the linear CPE and the signal that is to be sent to the non-linear CPE and on which non-linear precoding has been performed. The embodiments of the present invention achieve a win-win effect of maintaining a rate of a linear CPE and improving a rate of a non-linear CPE, and can decrease a bit error rate and improve system stability.