Vehicle Sensor Anomaly Detection via Dynamic Correlation Groups
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous and semi-autonomous vehicles face challenges in detecting anomalies in sensor data, which can lead to incorrect environmental perception and vehicle function decisions due to faulty sensors or malicious attacks, necessitating improved methods for anomaly detection.
Innovation Solution
The system employs sensor correlation groups based on vehicle dynamics, processing signals to identify anomalies using mean absolute error calculations, depth-first search methods, and trained thresholds, generating notification data to detect faulty sensors and protect against misbehaving data from other vehicles.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If sensor data is processed without anomaly detection, then processing speed is maintained, but reliability of vehicle decisions deteriorates due to faulty sensors or malicious attacks
Solution Approach 1:
The patent segments sensor data processing into multiple independent correlation groups (e.g., position group, speed group, acceleration group, heading group, yaw rate group), where each group processes specific sensor signals independently. This modular segmentation allows anomaly detection to be applied selectively to each group without processing all sensor data uniformly, improving reliability while managing system complexity through divided responsibility.
Solution Approach 2:
The patent introduces sensor correlation groups as intermediary structures between raw sensor data and vehicle decision-making systems. These correlation groups act as mediators that pre-process and validate sensor signals before they reach the controlling software, filtering out anomalies through correlation analysis while preserving valid data for downstream processing.
2Measurement precision
If comprehensive anomaly detection is implemented across all sensors, then detection accuracy improves, but processing time increases
Solution Approach 1:
By dividing sensor data into distinct correlation groups (position, speed, acceleration, heading, yaw rate), the system can process each group independently with optimized algorithms. This segmentation enables parallel processing of multiple sensor subsets simultaneously, maintaining high detection accuracy across all sensors while reducing overall processing time through concurrent execution.
Solution Approach 2:
The patent applies anomaly detection selectively to specific sensor correlation groups based on their criticality and anomaly susceptibility. Rather than uniformly applying complex detection algorithms to all sensors, the system focuses computational resources on high-priority groups where anomalies would have the greatest impact on vehicle safety, achieving sufficient detection accuracy with reduced processing overhead.
Data Source
AI summary
Systems and method are provided for detecting an anomaly of a sensor of a vehicle. In one embodiment, a method includes: storing a plurality of sensor correlation groups based on vehicle dynamics; processing a subset of signals based on the sensor correlation groups to determine when an anomaly exists; processing the subset of signals based on the sensor correlation group to determine which sensor of the sensor correlation group is anomalous; and generating notification data based on the sensor of the correlation group that is anomalous.