Navigation Signal Spoofing Detection Using LSTM Time Series Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current navigation systems lack effective methods to detect spoofing attacks, which can redirect vehicles or devices to incorrect locations, and existing non-machine learning based solutions require hardware modifications, making them impractical for commercial use.
Innovation Solution
Implementing a sequential model, such as Long Short-Term Memory (LSTM) or Recurrent Neural Network (RNN), to timestamp and analyze navigation signals, capturing long-term dependencies and differentiating between genuine and spoofed signals by processing time series data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If non-machine learning based spoofing detection methods (signal quality monitoring, signal power measuring, arrival time measurement) are implemented, then spoofing detection capability is improved, but hardware complexity increases due to requirement of additional antennas or hardware modifications
Solution Approach 1:
The patent replaces hardware-based detection methods (additional antennas, signal processing hardware) with a software-based machine learning approach. The LSTM neural network processes existing navigation signal data to detect spoofing, eliminating the need for physical hardware modifications while maintaining detection capability.
Solution Approach 2:
The patent uses timestamped copies of navigation signal data for training and detection. By creating and analyzing temporal copies of the signal data through sequencing and timestamping, the system achieves detection without requiring additional physical sensing hardware.
2Ease of operation
If simple ML models or DNN-based models are used to identify spoofed signals, then ease of operation is improved, but measurement precision deteriorates because they do not capture the dependency of signals for different time steps
Solution Approach 1:
The patent employs an LSTM (Long Short-Term Memory) neural network that dynamically processes sequential navigation signal data. The LSTM architecture captures temporal dependencies and dynamic patterns in the signal over time, improving detection accuracy while maintaining ease of operation through automated processing.
Solution Approach 2:
The patent adds the time dimension to the analysis by sequencing navigation signal data and using timestamp information. This transforms static signal analysis into dynamic temporal pattern recognition, enabling the model to capture dependencies across different time steps and improve spoofing detection accuracy.
Data Source
AI summary
Embodiments disclosed herein relate to monitoring navigation signals, and more particularly to verifying navigation signals, which can comprise of detecting spoofing and glitches in navigation signals by analyzing sequential time series data in the navigation signal. embodiments herein is to disclose methods and systems for verifying navigation signals, wherein navigation signals received by a receiver is timestamped, using a sequential model (such as, but not limited to, Long Short-Term Memory (LSTM), Recurrent Neural Network (RNN), and so on) for verifying patterns in the timestamped time series navigational signals, and determining whether the navigation signal is spoofed using the verified/unverified patterns.