Neural Network Channel Estimation for 6G Wireless Systems

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In 6G wireless communication, accurate channel estimation is hindered by the trade-off between signaling overhead and spectral efficiency, as dedicated channel estimation signals reduce overall spectral efficiency, and existing prediction methods are scenario-dependent and hard to implement effectively.

Innovation Solution

The use of neural networks, specifically convolutional and residual neural networks, to estimate channel estimation signals from non-dedicated reference signals like Synchronization Signal Blocks and Demodulation Reference Signals, reducing the need for frequent updates by leveraging temporal correlations and side-channel information, and requesting updates only when errors exceed a threshold.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If dedicated channel estimation signals (CSI-RS, DMRS) are used, then channel estimation accuracy is improved, but spectral efficiency deteriorates due to increased signaling overhead

Engineering Contradiction:
Improvechannel estimation accuracyVSAvoidspectral efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent applies multi-functionality by using non-dedicated reference signals (such as synchronization signals and broadcast channel reference signals) that serve multiple purposes: their primary function for synchronization or broadcast, and a secondary function for channel estimation. This eliminates the need for separate dedicated channel estimation signals, thereby reducing signaling overhead while maintaining channel estimation capability.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system enables self-service by allowing the network to estimate channel conditions using reference signals that are already being transmitted for other purposes. The channel estimation is performed autonomously by the receiving device using these multi-purpose signals, without requiring additional dedicated estimation signals from the network side.

Inventive Principle:
Principle #25Self-service

2Productivity

If non-dedicated reference signals are used for channel estimation, then spectral efficiency is improved by reducing signaling overhead, but channel estimation accuracy deteriorates

Engineering Contradiction:
Improvespectral efficiencyVSAvoidchannel estimation accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The patent introduces an intermediary mechanism in the form of a neural network that processes the reference signals to extract channel estimation information. The neural network acts as a mediator that transforms the non-dedicated reference signals into accurate channel estimates, bridging the gap between using fewer signals and maintaining estimation accuracy.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system applies parameter changes by using machine learning models with adjustable parameters (weights and biases) that are optimized to extract maximum channel information from the non-dedicated reference signals. The neural network parameters are trained to compensate for the reduced signal quality, enabling accurate channel estimation despite using fewer or non-dedicated signals.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If existing prediction methods are used, then some channel estimation capability is achieved, but implementation difficulty increases due to scenario-dependency

Engineering Contradiction:
Improvechannel estimation capabilityVSAvoidimplementation ease
Core Design Contradiction:
Measurement precisionVSEase of manufacture

Solution Approach 1:

The patent applies dynamics by using a neural network that can adapt its behavior based on different channel conditions and scenarios. The model dynamically adjusts its processing based on the input reference signals and can be retrained for different scenarios, providing a flexible solution that works across diverse conditions without requiring scenario-specific algorithm design.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system replaces traditional mechanical/mathematical prediction methods with a neural network-based approach. Instead of using fixed algorithmic rules that require manual tuning for different scenarios, the neural network learns optimal prediction strategies during training, substituting complex manual algorithm design with automated learning that simplifies implementation across various scenarios.

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

Data Source

PatentEP4270884A1Channel estimation using neural networks
Publication Date: 2023.11.01 NOKIA TECHNOLOGIES OY
  • EP4270884A1 patent drawingFigure 1
  • EP4270884A1 patent drawingFigure 2
  • EP4270884A1 patent drawingFigure 3~4

AI summary

This specification relates to channel estimation in wireless systems, and in particular to the use of neural networks for channel estimation. According to a first aspect of this specification, there is described apparatus comprising one or more receivers configured to receive signals comprising one or more first reference signals and one or more channel estimation signals. The apparatus is configured to perform a method comprising: generating, by a first neural network, a first estimate of a current value of a channel estimation signal from a current value of a first reference signal; generating, by a second neural network, a second estimate of the current value of the channel estimation signal from a previous value of the channel estimation signal; determining, by a third neural network, a current estimated error of the channel estimation signal from the first estimate of the current value of the channel estimation signal and the second estimate of the current value of the channel estimation signal.