GPS Spoofing Detection With Causal Shapley Explanations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine learning methods for detecting GPS spoofing attacks lack interpretability and overlook causal relationships among features, leading to uncertainty in classification and ineffective countermeasures.
Innovation Solution
Utilize a deep learning model with asymmetric Shapley values (ASVs) to calculate feature contributions and incorporate causal structures, providing a non-uniform distribution and Shapley attributions for improved GPS spoofing attack detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine learning methods are used to detect GPS spoofing attacks, then detection capability is improved, but interpretability deteriorates
Solution Approach 1:
The patent introduces Shapley values as an intermediary mechanism between the deep learning model and the user. Shapley values quantify the contribution of each input feature to the model's prediction, providing interpretable explanations for why certain GPS signals are classified as spoofed while maintaining the high detection capability of neural networks.
2Loss of information
If traditional Shapley values are used to explain model predictions, then interpretability is improved, but accuracy deteriorates due to overlooking causal relationships
Solution Approach 1:
The patent applies asymmetric Shapley values that incorporate causal relationships among features. Instead of treating all feature permutations equally, the method assigns different weights based on causal directions, ensuring that features with stronger causal influence on spoofing detection contribute more to the explanation, thereby improving both accuracy and interpretability.
3Measurement precision
If causal structures are incorporated into feature analysis, then detection accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent performs preliminary causal structure learning from the training data before deploying the detection model. By pre-computing the causal relationships and embedding them into the Shapley value calculation framework, the method avoids repeated complex causal inference during real-time detection, thus improving accuracy while managing computational complexity.
Data Source
AI summary
A method for detecting GPS spoofing attacks includes providing a trained deep learning (DL) model based on neural networks, feeding GPS signals into the trained DL model, using asymmetric Shapley values (ASVs) to calculate feature contributions, using the ASVs to assign a non-uniform distribution over an ordering of features, obtaining causal structures among the features, applying the ASVs to causal Shapley additive explanation to obtain Shapley attributions, incorporating the Shapley attributions and the causal structures, and detecting GPS spoofing attacks by running the trained DL model and using the causal structures, non-uniform distribution, feature contributions, and Shapley attributions.