Deep Learning Weather Classification From Telecom Signal Attenuation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Weather sensing technologies, such as weather stations, are costly and complex to implement, and there is a need for more sustainable techniques to monitor weather conditions, particularly for hyperlocal weather forecasting.

Innovation Solution

A deep learning model is trained using spectrogram images generated from telecommunication signals, specifically millimeter-wave signals, to detect and classify weather conditions like rain, snow, and no precipitation, by leveraging existing telecommunication infrastructure without the need for dedicated weather sensors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If weather station sensors are used to measure and quantify weather data, then weather information can be obtained, but the cost and complexity of implementation increase

Engineering Contradiction:
Improveweather data measurementVSAvoidweather station implementation
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies universality by enabling telecommunication signals to serve dual purposes: maintaining communication functionality while simultaneously acting as weather sensing probes. The millimeter-wave signals used for telecommunication also detect precipitation by measuring signal attenuation, eliminating the need for dedicated weather sensing infrastructure.

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

Solution Approach 2:

The telecommunication system serves itself by using its own transmitted signals to gather weather data. The infrastructure automatically measures signal attenuation caused by weather conditions without requiring external weather stations, making the system self-sufficient for both communication and weather monitoring functions.

Inventive Principle:
Principle #25Self-service

2Measurement precision

If weather station sensors are deployed, then weather data can be collected, but the cost of implementation increases

Engineering Contradiction:
Improveweather condition detectionVSAvoidimplementation cost
Core Design Contradiction:
Measurement precisionVSEase of manufacture

Solution Approach 1:

The patent enables existing telecommunication infrastructure to perform dual functions of communication and weather sensing. By analyzing signal attenuation of millimeter-wave telecommunication signals, the system extracts weather information without requiring separate weather station deployments, significantly reducing implementation costs.

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

Solution Approach 2:

The telecommunication system automatically generates weather data through its own signal transmissions. The infrastructure measures how its transmitted signals are attenuated by weather conditions, eliminating the need for external funding or deployment of dedicated weather monitoring equipment.

Inventive Principle:
Principle #25Self-service

3Ease of manufacture

If telecommunication signals are used for weather sensing, then cost is reduced, but the complexity of signal processing increases

Engineering Contradiction:
Improveimplementation costVSAvoidsignal processing complexity
Core Design Contradiction:
Ease of manufactureVSDevice complexity

Solution Approach 1:

The patent introduces spectrogram analysis as an intermediary tool that translates complex telecommunication signal attenuation patterns into visual representations. This intermediary step converts raw signal data into interpretable images showing weather conditions, simplifying the processing complexity while maintaining accurate weather detection capabilities.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces traditional mechanical weather sensing methods with signal processing and machine learning approaches. By using deep learning models trained on spectrogram images, the system automatically classifies weather conditions without requiring complex manual signal analysis or mechanical weather measurement devices.

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

4Measurement precision

If deep learning models are trained on spectrogram images, then weather classification accuracy improves, but computational requirements increase

Engineering Contradiction:
Improveweather condition classification accuracyVSAvoidcomputational energy consumption
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent applies preliminary action by pre-training deep learning models on spectrogram images during system deployment. Once trained, the model is stored and can perform rapid weather classification without requiring real-time retraining. This preliminary computational effort enables fast, low-energy inference operations during actual weather monitoring.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent uses spectrogram images as simplified copies or representations of the actual telecommunication signals. These visual representations capture the essential weather-related information in a format that is more efficient for machine learning processing, reducing computational complexity while maintaining classification accuracy.

Inventive Principle:
Principle #26Copying

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

The deep learning model accurately classifies weather conditions in real-time, providing hyperlocal weather data with high accuracy and reduced costs, using existing telecommunication signals and spectrogram analysis.

Implementation Method 1

receiving a telecommunication signal that is attenuated in multiple different weather conditions

Methodology Applied
Scientific EffectSignal attenuation: Absorption (EM radiation)

Data Source

PatentEP4641262A1Deep learning model for detecting and classifying weather conditions
Publication Date: 2025.10.29 NOKIA SOLUTIONS & NETWORKS OY
  • EP4641262A1 patent drawingFigure 1
  • EP4641262A1 patent drawingFigure 2
  • EP4641262A1 patent drawingFigure 3

AI summary

Disclosed is a method comprising receiving a telecommunication signal (211) that is attenuated in multiple different weather conditions; labeling the telecommunication signal (211) with the multiple different weather conditions; generating a set of spectrogram images (221-229) based on the telecommunication signal labeled with the multiple different weather conditions; and training a deep learning model (240) for detecting and classifying the multiple different weather conditions based on the set of spectrogram images (221-229).