Trajectory Sensor Data Anonymization Using Functional Road Classes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in anonymizing sensor data from trajectories while maintaining privacy and ensuring the data's value for location-based services, as existing anonymization methods can adversely affect the data's utility and may not comply with privacy standards.
Innovation Solution
The method employs NK Functional Class Anonymization, which uses integer values N and K to determine the number of functional class links to prune from the beginning and end of a trajectory, anonymizing sensor data using map data to ensure privacy while retaining data value for location-based services.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sensor data is anonymized using traditional methods, then privacy is improved, but data utility for location-based services deteriorates
Solution Approach 1:
The patent segments the trajectory data by dividing it into multiple path segments based on functional road classes. Instead of anonymizing the entire trajectory uniformly, it applies different anonymization levels to different segments, allowing valuable aggregate data to be retained while protecting sensitive portions of the trajectory.
Solution Approach 2:
The patent applies local quality by differentiating anonymization treatment based on the functional class of road segments. Different segments of the trajectory receive different levels of anonymization depending on their classification, ensuring that privacy is protected where needed while data utility is maintained in less sensitive areas.
2Reliability
If trajectory data is heavily anonymized, then privacy protection is improved, but temporal accuracy deteriorates
Solution Approach 1:
The patent applies partial anonymization by redacting only the necessary portions of the trajectory data (first and last path segments) while retaining the intermediate segments. This partial action approach maintains temporal accuracy for the retained segments while providing sufficient privacy protection through selective redaction.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Embodiments described herein relate to anonymizing sensor data through the use of map data. Methods include: receiving sensor data defining a trajectory; map-matching the sensor data using a map-matching algorithm to a plurality of road segments of a map database to generate a sequence of map-matched sensor data elements; determining a first value representing anonymization associated with a start of the trajectory; determining a second value representing anonymization associated with an end of the trajectory; determining first map-matched sensor data elements at the start of the trajectory to be redacted based on the first value; determining second map-matched sensor data elements at the end of the trajectory to be redacted based on the second value; and transmitting sensor data associated with elements in the sequence of map-matched sensor data elements between the first map-matched sensor data elements and the second map-matched sensor data elements.