Vehicle Event Correlation Using Interior and Exterior Sensor Context
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle operation analysis systems lack the ability to effectively correlate interior and exterior events to provide comprehensive contextual data for autonomous vehicle control, vehicle maintenance, and insurance claim analysis, while also failing to account for real-world driving conditions and driver behavior.
Innovation Solution
A method and system that utilizes onboard vehicle systems with multiple sensors, including inward and outward facing cameras, to correlate interior and exterior events, generating combined event data through geometric, semantic, and temporal correlations, and optionally training event-based models for autonomous vehicle control and maintenance determination.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If multiple sensors including inward and outward facing cameras are deployed to capture comprehensive driving data, then the completeness and contextual richness of vehicle operation data is improved, but the device complexity and cost increase
Solution Approach 1:
The system segments the complex task of contextualized vehicle operation determination into distinct functional modules: exterior event detection module, interior event detection module, correlation module, and model training module. Each module processes specific aspects of the data independently before integration, reducing overall system complexity while maintaining comprehensive data capture capabilities
Solution Approach 2:
The event-based model serves multiple functions simultaneously: it determines exterior events from exterior camera data, determines interior events from interior camera data, correlates these events to generate combined event data, and can be retrained with new datasets. This multi-functionality reduces the need for separate specialized systems for each task
2Measurement precision
If comprehensive contextual data including interior and exterior events is collected and correlated, then the accuracy of autonomous vehicle control and insurance claim analysis is improved, but the processing time and computational resources increase
Solution Approach 1:
The system performs preliminary actions by continuously capturing and pre-processing sensor data streams from multiple cameras, maintaining buffered data ready for analysis. This allows the correlation module to quickly access pre-processed data when events occur, reducing real-time processing delays while maintaining comprehensive analysis capabilities
Solution Approach 2:
The patent replaces traditional mechanical or rule-based event correlation systems with machine learning-based event-based models. These models automatically learn patterns and correlations from training data, enabling faster and more accurate event correlation without complex manual rule sets, thereby reducing processing time while improving accuracy
3Reliability
If event-based models are trained with combined interior and exterior event data, then the reliability of autonomous vehicle determination systems is improved, but the training time and computational resources increase
Solution Approach 1:
The system implements partial training by allowing the event-based model to be trained incrementally with new datasets rather than requiring complete retraining. The model can be updated with subsets of data relevant to specific scenarios or conditions, reducing training time while maintaining high reliability through continuous improvement with targeted data
Solution Approach 2:
The patent enables parameter changes in the event-based model during training, allowing adjustment of learning rates, data sampling frequencies, and correlation thresholds. These parameter optimizations accelerate convergence during training while ensuring the model achieves high reliability, balancing training duration with model performance
Data Source
AI summary
A method for determining event data including: sampling a first data streamwithin a first time window at a first sensor of an onboard vehicle system coupled to a vehicle, extracting interior activity data from the first data stream; determining an interior event based on the interior activity data; sampling a second data stream within a second time window at a second sensor of the onboard vehicle system; extracting exterior activity data from the second image stream; determining an exterior event based on the exterior activity data; correlating the exterior event and the interior event to generate combined event data; automatically classifying the combined event data to generate an event label; and automatically labeling the first time window of the first data stream and the second time window of the second data stream with the combined event label to generate labeled event data.