Driver Risk Estimation Models by Task State and Biometric Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems struggle to accurately estimate accident risk in drivers due to the complexity of biological factors beyond drowsiness, such as reduced concentration, excitement, and excessive nervousness, making it difficult to discern causal relationships between these factors and accidents.
Innovation Solution
An operation management aid system that utilizes a processor to access and analyze biological measurement data in conjunction with task state data to generate an estimation model for each specific task state, enabling accurate estimation of accident risk through a first acquisition process and generation of an estimation model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single estimation model is used for all task states, then the system complexity is reduced, but the measurement precision of accident risk estimation deteriorates
Solution Approach 1:
The patent divides the estimation system into multiple task-state-specific models (driving model, non-driving model, transition model) instead of using a single unified model. Each model is trained on data specific to its task state, enabling precise accident risk estimation tailored to the characteristics of each state while maintaining manageable system complexity through modular architecture.
2Measurement precision
If multiple task-state-specific estimation models are generated, then the measurement precision of accident risk estimation is improved, but the device complexity increases
Solution Approach 1:
The system dynamically selects and switches between different estimation models based on the current task state detected by the task state determination unit. The model selection is not static but adapts in real-time to changing driving conditions, allowing the system to maintain high precision across varied scenarios without requiring all models to be actively managed simultaneously, thus controlling complexity.
3Ease of operation
If biological measurement data is analyzed without considering task states, then the ease of operation is improved, but the reliability of accident risk estimation deteriorates
Solution Approach 1:
The task state determination unit acts as an intermediary between raw biological measurement data and the accident risk estimation process. It first classifies the current task state (driving, non-driving, transition) and then routes the data to the appropriate task-state-specific model, thereby automatically incorporating task state context without requiring manual intervention, thus maintaining ease of operation while significantly improving estimation reliability.
Data Source
AI summary
An operation management aid system can access a first group of associated data in which biological measurement data pertaining to a biology of a driver is associated with task state data pertaining to a task state of the driver, and a second group of danger determination results indicating a degree of danger of driving by the driver, and wherein a processor executes: acquiring, the associated data pertaining to a specific task state of the driver from the first group, and acquiring a specific danger determination result group in the specific task state of the driver from the second group; and using an associated data group pertaining to the specific task state and a specific danger determination result group for the specific task state, to generate, for each of the specific task states, an estimation model that estimates an accident risk of the driver during the specific task state.