Driver Attention Scoring via Multi-Source Data Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle systems cannot effectively determine what captures a driver's attention based on driving conditions, leading to inadequate detection of attention and inappropriate responses to road features or objects.
Innovation Solution
A method and system that process object data, upcoming behavior data, and gaze data to compute an attention score, using weighting factors based on environmental conditions and driving scenarios, to determine the driver's attention and generate notifications or control signals accordingly.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If general gaze direction detection is used to determine driver attentiveness, then the system can detect basic attention status, but it cannot determine what specific object captures the driver's attention
Solution Approach 1:
The patent combines multiple data sources including gaze detection data, object detection data from vehicle sensors, and behavior plan data into a unified attention determination system. By merging these diverse data streams and processing them together through a integrated framework, the system achieves precise identification of what captures driver attention without requiring each individual sensor to be overly complex
Solution Approach 2:
The patent introduces an intermediary processing layer that correlates gaze direction with detected objects and upcoming vehicle behaviors. This intermediary layer processes the raw gaze data and matches it with contextual information from object detection and behavior planning systems, enabling precise attention determination while maintaining modular system architecture
2Measurement precision
If the system uses multiple data sources (object data, behavior data, gaze data) to improve attention determination accuracy, then the detection precision improves, but the processing complexity increases
Solution Approach 1:
The patent segments the attention determination process into distinct functional modules: gaze data processing, object data processing, behavior data processing, and integration/correlation layer. Each module handles specific data types independently, then combines results through systematic correlation. This segmentation allows complex multi-source data processing to be managed through modular, organized steps rather than monolithic processing
Solution Approach 2:
The patent applies different processing approaches to different data types based on their specific characteristics. Gaze data receives temporal filtering and smoothing appropriate for eye tracking, object data undergoes spatial classification and filtering, behavior data receives temporal alignment processing. Each data source is processed with quality-appropriate methods before integration, optimizing overall system performance
3Loss of information
If the system provides detailed attention information about specific objects, then the usefulness of the information improves, but the response time may be delayed due to additional processing
Solution Approach 1:
The patent performs preliminary processing of object detection data and behavior plan data in advance, maintaining pre-processed information about detected objects and upcoming vehicle actions. When gaze data becomes available, the system can quickly correlate with pre-prepared object and behavior information rather than processing everything from scratch, reducing latency while providing detailed attention information
Solution Approach 2:
The patent implements continuous processing pipelines where object detection, behavior planning, and gaze tracking operate in parallel and continuously feed into the attention determination system. This continuous action ensures that when gaze data is processed, the corresponding object and behavior data is already prepared and available for immediate correlation, minimizing processing delays
Data Source
AI summary
Methods and systems are provided for processing attention data. In one embodiment, a method includes: receiving, by a processor, object data associated with at least one object of an exterior environment of the vehicle; receiving upcoming behavior data determined from a planned route of the vehicle; receiving gaze data sensed from an occupant of the vehicle; processing, by the processor, the object data, the upcoming behavior data, and the gaze data to determine an attention score associated with an attention of the occupant of the vehicle; and selectively generating, by the processor, signals to at least one of notify the occupant and control the vehicle based on the attention score.