GPS Spoofing Detection Using Synthetic Aviation Data at the Edge
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for detecting and responding to GPS spoofing events are inaccurate, unreliable, resource-intensive, and suffer from high latency due to insufficient training data and remote communication challenges, particularly in polar and ocean regions.
Innovation Solution
A cloud-based device trains a generative machine learning model using aviation specification data to generate synthetic aviation data, which is used to train GPS spoofing detection models that can be deployed to an edge-based device like an electronic flight bag for real-time detection and response.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing GPS spoofing detection systems are deployed, then detection capability is provided, but accuracy and reliability are insufficient due to lack of training data
Solution Approach 1:
The patent creates synthetic copies of aviation operation data through generative adversarial networks. The generator network produces artificial GPS data, aircraft state data, and operation data that mimic real aviation scenarios, including spoofing attacks. This synthetic data copying approach solves the scarcity of real labeled spoofing data while maintaining statistical properties useful for training detection models
Solution Approach 2:
The system performs preliminary data generation and model training in advance using synthetic data. The generative model is trained beforehand to produce realistic aviation data distributions, and the spoofing detection model is pre-trained with this synthetic data before deployment. This preliminary action ensures the detection system is ready with adequate training even when real spoofing data is unavailable
2Measurement precision
If cloud-based processing is used for GPS spoofing detection, then comprehensive analysis is possible, but latency increases due to remote communication
Solution Approach 1:
The system segments the detection architecture into cloud-based training component and edge-based inference component. The generative model and detection model are trained comprehensively in the cloud using synthetic data, then the trained detection model is deployed to edge devices on aircraft. This segmentation allows comprehensive cloud processing for model development while enabling low-latency local inference for real-time spoofing detection
Solution Approach 2:
The patent introduces a synthetic data generation mechanism as an intermediary between data collection and model training. The generative adversarial network acts as a mediator that creates intermediate synthetic data representations, which then serve as training material for the detection model. This intermediary layer enables the system to bridge the gap between limited real data and comprehensive training requirements
3Reliability
If more computing resources are allocated to GPS spoofing detection, then detection performance improves, but resource consumption increases
Solution Approach 1:
The system performs computationally intensive model training in advance using synthetic data generated by the generative model. The detection model is pre-trained and optimized before deployment to edge devices. This preliminary action shifts the heavy computational burden to the training phase, allowing the deployed detection system to operate with lower resource consumption during actual spoofing detection while maintaining high reliability
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Systems, apparatuses, methods, and computer program products are provided herein. For example, a method may include access aviation specification data. In some embodiments, the method may include training a generative machine learning model using aviation specification data (504). In some embodiments, the method may include generating synthetic aviation data using the generative machine learning model (506). In some embodiments, the method may include training one or more global positioning system (GPS) spoofing detection machine learning models using the synthetic aviation data and historical aviation operations data (508). In some embodiments, the method may include deploying a first GPS spoofing detection machine learning model of the one or more GPS spoofing detection machine learning models to an edge-based device (510).