Data Privacy System Using MPC and TEE for Sensor Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current data privacy laws, such as GDPR and CCPA, restrict the use of data collected by cameras and sensors due to the presence of personally identifiable information (PII), making it challenging to utilize this data for purposes like autonomous driving without violating privacy regulations.
Innovation Solution
A data privacy system that employs a Multi-Party Computation (MPC) framework and a Trusted Execution Environment (TEE) to separate and protect personal data from non-personal data, ensuring compliance with privacy laws by masking personal data and allowing its secure use for data processing and analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If sensor data containing personal data is collected and processed, then data utility for autonomous driving improvement is enhanced, but compliance with privacy laws deteriorates
Solution Approach 1:
The system segments sensor data into personal data and non-personal data using multiple backend computers in an MPC framework. Each backend computer processes specific portions of data, allowing utility extraction from non-personal data while personal data remains segmented and protected through cryptographic sharing, thus maintaining both data utility and privacy compliance
Solution Approach 2:
The MPC framework introduces intermediary backend computers that act as mediators between data collection and processing. These intermediaries hold cryptographic shares of personal data without being able to reconstruct it individually, enabling data processing operations to proceed while maintaining privacy law compliance through the intermediary's distributed nature
2Reliability
If personal data is protected through MPC and TEE, then security against data breaches is improved, but data processing capability deteriorates
Solution Approach 1:
The system merges MPC framework with TEE technology to combine the security benefits of cryptographic sharing with the trusted execution capabilities of hardware-enforced enclaves. This merging allows data processing to occur within trusted boundaries while personal data remains protected, maintaining both security and processing capability
Solution Approach 2:
The system performs preliminary actions by pre-establishing trusted execution environments and cryptographic protocols before data processing begins. Data is encrypted and distributed through MPC shares beforehand, allowing secure processing to proceed without real-time security overhead, thus maintaining processing capability while ensuring security
3Reliability
If personal data is separated and masked, then privacy protection is improved, but data quality for analysis deteriorates
Solution Approach 1:
The system applies different quality levels to different portions of data: personal data receives maximum privacy protection through masking and cryptographic sharing, while non-personal data maintains full quality for analysis. This local differentiation ensures privacy protection where needed while preserving data quality for analytical purposes
Solution Approach 2:
The system creates cryptographic copies (shares) of personal data that can be processed without revealing the original information. These copies maintain the structural properties needed for analysis while being mathematically designed to prevent reconstruction of personal identifiers, thus preserving data quality for analysis while ensuring privacy protection
Data Source
AI summary
A backend computer and methods of using the backend computer are described. The method may comprise: receiving, at a first backend computer, sensor data associated with a vehicle; determining a labeling of the sensor data, comprising: determining personal data and determining non-personal data that is separated from the personal data, wherein each of the personal and non-personal data comprise labeled data, wherein the personal data comprises information relating to at least one identified or identifiable natural person; and performing via the personal data and the non-personal data that is separated from the personal data, at the first backend computer, data processing associated with collecting sensor data associated with the vehicle.