Driver gaze behavior classification for autonomous vehicle comfort
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current driver monitoring systems in autonomous vehicles cannot effectively predict driver discomfort or distraction based on gaze behavior, leading to unnecessary take-over events and an inability to adapt behavior models to individual driver preferences.
Innovation Solution
A system that monitors driver gaze behavior using a driver monitoring system and a data processor to classify driver statuses, sending instructions to vehicle systems to adjust operating parameters and update pre-defined gaze models based on actual gaze models and vehicle operation status, thereby reducing take-over events and improving driver comfort.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If driver monitoring systems use basic gaze detection, then driver safety can be monitored, but the system cannot predict driver discomfort or distraction leading to unnecessary take-over events
Solution Approach 1:
The system transforms basic gaze detection into comprehensive driver status monitoring by analyzing multiple parameters including gaze duration, gaze frequency, gaze direction, and blink patterns. This multi-parameter approach enables accurate classification of driver states (attentive, distracted, drowsy) and predicts discomfort before take-over events occur, resolving the contradiction between measurement precision and prediction reliability
Solution Approach 2:
The system implements feedback mechanisms where driver gaze behavior data is continuously collected, analyzed, and used to update driver status classifications in real-time. This feedback loop allows the system to adapt to individual driver patterns and improve prediction accuracy over time, enhancing both measurement precision and take-over event prediction reliability
2Adaptability or versatility
If a fixed behavior model is used for vehicle control, then autonomous vehicle operation can be standardized, but the system cannot adapt to individual driver preferences leading to increased take-over events
Solution Approach 1:
The behavior model transitions from a static, fixed configuration to a dynamic system that continuously adapts to individual driver preferences. The system monitors driver gaze behavior and automatically adjusts vehicle control parameters such as warning thresholds, alert timing, and operational boundaries. This dynamic adaptation enables the system to personalize the autonomous driving experience for each driver while maintaining manageable complexity through automated learning algorithms
Solution Approach 2:
The system implements self-service capabilities where the behavior model automatically learns and adapts to driver preferences without requiring manual configuration or intervention. The automated system analyzes driver responses to various driving scenarios and independently optimizes control parameters, reducing the burden on drivers while improving adaptability to individual preferences
3Reliability
If the vehicle control module maintains strict autonomous operation parameters, then safety can be ensured, but driver comfort decreases leading to more take-over events
Solution Approach 1:
The system performs preliminary classification of driver status (attentive, distracted, drowsy) before take-over events occur by analyzing gaze behavior patterns. This early detection enables the vehicle control module to proactively adjust operational parameters within safety boundaries to enhance driver comfort, such as modifying warning alert timing or adjusting driving parameters, thereby preventing take-over events while maintaining safety
Data Source
AI summary
A method of monitoring a driver in an autonomous vehicle includes monitoring, with a driver monitoring system, a driver of a vehicle, collecting, with a data processor, data from the driver monitoring system related to gaze behavior of the driver, classifying, with the data processor, the driver as one of a plurality of driver statuses based on the data from the driver monitoring system, and sending, with the data processor, instructions to at least one vehicle system based on the classification of the driver.