Virtual Vehicle Entity for Connected Vehicle Security
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern vehicles with Controller Area Networks (CAN) lack standardized security measures to protect against malicious operations injected via external networks, posing safety risks due to the potential for disruptions in critical subsystems.
Innovation Solution
Implementing a virtual vehicle entity (VVE) that operates as a proxy for the CAN, utilizing distributed deep learning platforms to evaluate operations before they reach the vehicle, thereby detecting and mitigating malicious attacks by either forwarding or discarding them.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a proxy system with deep learning evaluation is implemented to detect malicious operations, then security against malicious injections is improved, but device complexity increases
Solution Approach 1:
The patent introduces a proxy system that acts as an intermediary between external networks and the vehicle's CAN bus. This proxy intercepts, evaluates, and filters operations before they reach the vehicle systems, using deep learning models to distinguish malicious from legitimate operations. The intermediary approach enhances security without directly modifying the vehicle's existing ECU architecture.
Solution Approach 2:
The system segments the security evaluation function into a separate proxy component distinct from the vehicle's core control systems. By isolating the deep learning evaluation logic in an external or semi-external proxy, the system achieves enhanced security while maintaining modularity and minimizing impact on the vehicle's existing architecture.
2Measurement precision
If deep learning evaluation is performed on all incoming operations, then detection precision of malicious attacks is improved, but processing time increases
Solution Approach 1:
The proxy system applies deep learning evaluation selectively rather than uniformly to all operations. By prioritizing evaluation of suspicious or high-risk operations and using heuristic filters for routine legitimate operations, the system achieves high detection precision for malicious attacks while minimizing average processing time through differentiated handling.
Solution Approach 2:
The system performs preliminary filtering and pre-evaluation of incoming operations using lightweight heuristics and rules before applying full deep learning analysis. This preliminary action identifies and processes obviously legitimate operations quickly, reserving computationally intensive deep learning evaluation for operations that require more thorough scrutiny, thus reducing overall processing time.
Data Source
AI summary
A network device implements a method to protect a vehicle from insertion of malicious operations. The method includes establishing a communication session with a requestor as a proxy for the vehicle, receiving status information from the vehicle, querying a deep learning platform with the status information and message from the requestor, and dropping the message from the requestor in response to the deep learning platform indicating the message is malicious.