Spoofing Detection for Autonomous Vehicle Sensors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles are vulnerable to spoofing attempts, where sensors provide inaccurate information, leading to undesirable vehicle operations, as existing systems fail to effectively detect and prevent such attacks.
Innovation Solution
A method and apparatus that acquire and annotate sensor data from various sensors, compare it with traffic reference information, and identify inconsistent data to detect spoofing attempts, ignoring the affected sensor information and updating the traffic reference data using machine learning for future prevention.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If sensor data is acquired and processed for autonomous vehicle operation, then the vehicle can perform driving functions, but the system becomes vulnerable to spoofing attempts that provide inaccurate information
Solution Approach 1:
The system continuously compares sensor data with traffic reference information obtained from multiple sources (cloud server, other vehicles, roadside machines) to detect inconsistencies that indicate spoofing attempts. This feedback mechanism allows the vehicle to identify and ignore compromised sensor information while maintaining autonomous driving functionality.
Solution Approach 2:
The patent introduces traffic reference information as an intermediary layer between sensor data and vehicle control decisions. By comparing sensor information against this reference data from external sources, the system can detect spoofing attempts without directly compromising the autonomous driving function.
2Measurement precision
If the system compares sensor information with traffic reference information to detect spoofing, then detection accuracy improves, but processing time and computational load increase
Solution Approach 1:
The system pre-obtains and stores traffic reference information from cloud servers, other vehicles, and roadside machines before needing to detect spoofing. This preliminary preparation of reference data enables rapid comparison with sensor information during critical moments without significant processing delays.
Solution Approach 2:
The patent creates multiple copies of traffic reference information from different sources (cloud server, neighboring vehicles, roadside infrastructure) to enable parallel comparison with sensor data. This copying approach allows the system to quickly identify inconsistencies without relying on a single data source.
3Reliability
If the vehicle ignores sensor information associated with spoofing attempts, then safety improves, but the quantity of usable sensor data decreases
Solution Approach 1:
The system extracts and removes only the specific portions of sensor information that are associated with spoofing attempts, while retaining and utilizing the remaining valid sensor data. This selective extraction approach maintains vehicle safety by ignoring compromised information without unnecessarily discarding useful sensor inputs.
Solution Approach 2:
The patent applies different quality assessments to different portions of sensor data based on their consistency with traffic reference information. Rather than uniformly discarding all sensor data when spoofing is detected, the system identifies and ignores only the locally compromised portions while maintaining trust in consistent data sources.
Data Source
AI summary
A method and an apparatus for detecting a spoofing attempt associated with an autonomous vehicle are provided. The method includes acquiring, via interface circuitry of the apparatus for spoofing prevention, one or more sensor data from one or more sensors. The one or more sensor data is annotated to obtain sensor information. The sensor information extracts traffic information that the one or more sensor data carries. Abnormal sensor data that fails to capture surrounding traffic information is discarded. Furthermore, a spoofing attempt is determined based on a determination that at least one inconsistent sensor data is identified. The at least one inconsistent sensor data provides different traffic information compared to other sensor data of the one or more sensor data generated by the one or more sensors. The vehicle is therefore informed to ignore a portion of the sensor information associated with the spoofing attempt when the spoofing attempt is identified.