GNSS Spoofing Detection with Multisignal Neural Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing rule-based algorithms for detecting navigation system spoofing face challenges in distinguishing spoofers with unclear boundaries, struggle with inconsistent detection conditions, and require frequent rule modifications to address new spoofing techniques, leading to reduced accuracy for older techniques.
Innovation Solution
A system utilizing deep learning, specifically neural networks, to analyze GNSS and non-GNSS metrics for spoofing detection, combined with a rules-based system for fail-safe operation, enabling accurate detection and classification of spoofing types.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If rule-based algorithms are used for spoofing detection, then the system structure is simple and easy to implement, but the detection accuracy deteriorates due to difficulty in distinguishing spoofers with unclear boundaries and inconsistent detection conditions
Solution Approach 1:
The patent replaces the mechanical rule-based detection system with a neural network-based system. The neural network learns complex patterns from training data and automatically identifies spoofing conditions, eliminating the need for manually crafted rules and achieving both high accuracy and adaptability to various spoofing techniques.
Solution Approach 2:
The patent transforms the detection approach by changing from fixed rule parameters to learned model parameters. The neural network adjusts its internal parameters during training to optimize detection performance, enabling it to handle unclear boundaries and inconsistent conditions that fixed rules cannot accommodate.
2Adaptability or versatility
If rules are modified to detect new spoofing techniques, then the system can adapt to new threats, but the accuracy for detecting older known spoofing techniques deteriorates
Solution Approach 1:
The neural network is trained on a comprehensive dataset that includes multiple types of spoofing techniques. This universal training enables the single model to detect both older known techniques and new emerging techniques without requiring separate rules for each technique, maintaining high accuracy across all categories.
Solution Approach 2:
The system performs preliminary training on a diverse set of spoofing examples before actual detection occurs. This pre-learning phase enables the neural network to recognize patterns from older techniques while remaining adaptable to new techniques, as the underlying model captures fundamental spoofing characteristics rather than memorizing specific rules.
3Ease of operation
If rule-based systems are used, then the system is easier to operate and maintain, but it struggles with inconsistent detection conditions and cannot handle edge cases effectively
Solution Approach 1:
The neural network system is self-learning and self-adjusting. It automatically processes input data, identifies patterns, and makes detection decisions without requiring manual rule configuration or intervention. This self-service capability ensures consistent and reliable detection across all conditions, including edge cases that would confuse rule-based systems.
Data Source
AI summary
A system and method for a navigation system (NS) including global navigation satellite system (GNSS) spoofing detection using deep learning, such as a neural network, is provided. A GNSS signal is received by a receiver and one or more metrics are obtained from the received signal. Optionally, one or more non-GNSS signal metrics may also be obtained, e.g., information from an inertial measurement unit, vision system, independent time source, etc. The various metrics are fed into a trained neural network that decides as to whether spoofing is present and, if so, what is the type of spoofing that is occurring.