Adaptive Nonlinear Receiver Modeling for Distortion-Tolerant Modulation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing communication systems are power hungry and spectrally inefficient, particularly due to the degradation of complex linear modulation schemes in the presence of phase noise and non-linear distortion, which increases the gap between desired and actual spectral efficiency as higher-order modulation is needed for increased throughput.

Innovation Solution

An adaptive non-linear model is implemented in communication systems, where a receiver generates estimates of transmitted symbols using a sequence estimation process that incorporates a non-linear model, adapting based on communication information to improve spectral efficiency and tolerance to non-linearity, by using partial response signals and dynamically updating the non-linear model to compensate for distortion.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If higher-order modulation is used to increase throughput, then spectral efficiency is improved, but sensitivity to non-linear distortion and phase noise increases

Engineering Contradiction:
ImprovethroughputVSAvoidsensitivity to non-linear distortion
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent transforms the received signal through mathematical operations including frequency domain conversion and iterative processing to change the signal parameters, enabling higher-order modulation to be used while maintaining robustness against non-linear distortion and phase noise

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent employs iterative processing where estimated symbols are fed back into the processing chain to refine subsequent estimates, allowing the system to compensate for non-linear distortion effects and maintain reliability at higher spectral efficiencies

Inventive Principle:
Principle #23Feedback

2Productivity

If complex linear modulation schemes are used, then spectral efficiency is improved, but performance degrades in the presence of phase noise and non-linear distortion

Engineering Contradiction:
Improvespectral efficiencyVSAvoidperformance degradation
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The patent replaces traditional time-domain equalization methods with frequency domain processing and iterative signal transformation, substituting conventional approaches with mathematical transformations that are more robust to non-linear distortion and phase noise

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The patent performs preliminary frequency domain conversion and signal transformation before symbol detection, preparing the signal in advance to be more resistant to the effects of non-linear distortion and phase noise during the detection process

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS8824611B2Adaptive non-linear model for highly-spectrally-efficient communications
Publication Date: 2014.09.02 AVAGO TECHNOLOGIES INTERNATIONAL SALES PTE LTD
  • US8824611B2 patent drawing
  • US8824611B2 patent drawing
  • US8824611B2 patent drawing

AI summary

A receiver may be operable to generate estimates of transmitted symbols using a sequence estimation process that may incorporate a non-linear model. The non-linear model may be adapted by the receiver based on particular communication information that may be indicative of non-linearity experienced by the transmitted symbols. The receiver may generate a reconstructed signal from the estimates of the transmitted symbols. The receiver may adapt the non-linear model based on values of an error signal generated from the reconstructed signal, and the values of the error signal may be generated from a portion of the generated estimates that may correspond to known symbols and/or information symbols. The values of the error signal corresponding to the known symbols may be given more weight in an adaptation algorithm, and the values of the error signal corresponding to the information symbols may be given less weight in the adaptation algorithm.