GNSS Receiver Interference Detection Using Neural Network Classification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing GNSS receiver interference detection methods require multiple rule-based algorithms, leading to complexity, maintenance challenges, and inefficiency in handling new interference types.

Innovation Solution

Utilizing neural networks, particularly CNNs and ANNs, to classify RF interference environments, enabling accurate and efficient detection and mitigation of interference without the need for new algorithm development.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If multiple rule-based algorithms (spectral analysis and statistical analysis) are used for interference detection, then detection coverage for different interference types is improved, but device complexity and difficulty of maintenance increase

Engineering Contradiction:
Improvedetection coverageVSAvoidalgorithm complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent combines multiple rule-based algorithms (spectral analysis and statistical analysis) into a single neural network model. The neural network integrates the functionality of both algorithms, processing RF signal data through a unified architecture that performs both spectral and statistical analysis simultaneously, eliminating the need to manage multiple separate algorithms and their concurrent operation.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The neural network is designed as a universal detector that can handle multiple types of interference (in-band and out-of-band) with a single model. Unlike the prior approach requiring separate algorithms for different interference types, this universal neural network adapts to various interference scenarios through its learned parameters, providing multi-functional detection capability.

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

2Measurement precision

If new interference types are detected by developing new rule-based algorithms, then detection accuracy for new interference is improved, but ease of manufacture and maintenance deteriorate

Engineering Contradiction:
Improvedetection accuracyVSAvoidalgorithm development effort
Core Design Contradiction:
Measurement precisionVSEase of manufacture

Solution Approach 1:

The patent transforms the approach from developing new algorithms to changing model parameters. When new interference types need to be detected, instead of creating new rule-based algorithms, the neural network's parameters (weights and biases) are retrained or adjusted based on new training data representing the interference type. This parameter adjustment approach maintains high detection accuracy while dramatically reducing development effort.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent uses training data that copies the characteristics of various interference types to teach the neural network. By providing labeled examples of different interference patterns during training, the network learns to recognize and detect these patterns without requiring explicit rule-based programming for each type, thereby improving ease of manufacture.

Inventive Principle:
Principle #26Copying

3Reliability

If spectral analysis and statistical analysis algorithms run concurrently, then interference detection reliability is improved, but device complexity and operational difficulty increase

Engineering Contradiction:
Improvedetection reliabilityVSAvoidoperational simplicity
Core Design Contradiction:
ReliabilityVSEase of operation

Solution Approach 1:

The patent merges spectral analysis and statistical analysis into a single neural network processing pipeline. The network receives RF signal data and internally performs both types of analysis through its architecture, producing a unified detection output. This eliminates the operational complexity of managing concurrent algorithms, including coordinating their execution and resolving conflicting detection outcomes.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The neural network autonomously performs interference detection without requiring external coordination between multiple algorithms. The model self-adjusts its processing based on the input data characteristics, automatically determining which detection approaches are most appropriate for the current signal conditions, thereby improving ease of operation.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP3904912B1GNSS-receiver interference detection using deep learning
Publication Date: 2026.03.18 NOVATEL INC
  • EP3904912B1 patent drawingFigure 1~3
  • EP3904912B1 patent drawingFigure 2
  • EP3904912B1 patent drawingFigure 4

AI summary

Systems and methods are described for classification of interference for GNSS receivers. One or more neural networks are utilized to classify RF signal data received by a GNSS receiver. The classification associates the RF signal data with an RF environment. Appropriate interference mitigation techniques can be implemented by the receiver based on the classification.