Driver Condition Assessment Device Using Adaptive Likelihood Thresholds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional driver condition determination devices inaccurately assess degraded consciousness, leading to unnecessary alarms, as they rely on constant threshold values regardless of driver characteristics and environmental factors during tailgating scenarios.
Innovation Solution
A driver condition determination device that models the accelerating operation condition of each driver when tailgating a preceding vehicle using a mixed normal distribution, calculating likelihood and threshold values to accurately determine degraded consciousness based on individual driving characteristics.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a constant determination threshold value is used for proximity-based degraded consciousness determination, then the device complexity is reduced, but the measurement precision of driver consciousness state deteriorates
Solution Approach 1:
The patent applies dynamics by transitioning from a static constant threshold to a dynamic adaptive threshold that changes based on individual driver characteristics. The system learns each driver's normal tailgating behavior patterns and adjusts the determination threshold accordingly, allowing the threshold to adapt over time while maintaining system simplicity.
Solution Approach 2:
The patent changes the determination parameter from a fixed constant value to a variable threshold based on probability distributions. By modeling driver behavior parameters (distance changes, relative speed changes) and calculating likelihood ratios, the system dynamically adjusts the threshold to distinguish normal tailgating from degraded consciousness states.
2Measurement precision
If individual driver characteristics are considered in degraded consciousness determination, then the measurement precision improves, but the device complexity increases
Solution Approach 1:
The system applies self-service by automatically learning and adapting to each driver's individual characteristics without requiring manual configuration. The determination threshold is automatically adjusted based on observed driver behavior patterns, and the system performs self-calibration during normal operation.
Solution Approach 2:
The patent implements feedback mechanisms where the system continuously monitors driver behavior, compares actual operations against learned patterns, and adjusts determination thresholds based on the likelihood ratio calculations. This closed-loop feedback enables accurate individualized determination while maintaining systematic simplicity.
3Ease of operation
If a constant threshold is used for proximity determination, then the ease of operation is improved, but the reliability of alarm output deteriorates
Solution Approach 1:
The system performs preliminary action by pre-learning normal driver tailgating behavior patterns before making degraded consciousness determinations. It establishes baseline probability distributions of each driver's normal operations in advance, enabling more reliable real-time alarm decisions.
Solution Approach 2:
The patent applies dynamics by making the alarm threshold adaptive rather than static. The system dynamically adjusts determination criteria based on individual driver characteristics and current operating conditions, improving alarm reliability while maintaining ease of operation through automated adaptation.
Data Source
Figure 1
Figure 2(a)~2(b)
Figure 3
AI summary
A driver condition determination device determines whether a driver of a host vehicle has degraded consciousness. A main component for distribution of an accelerating operation condition with respect to proximity when tailgating a preceding vehicle is analyzed (S 16), all data is origin-shifted (S18), and the accelerating operation condition of the driver when the driver has consciousness is created as a normal traveling model (S20). The likelihood of the accelerating operation condition, a likelihood average, a likelihood variance, and a likelihood threshold value are calculated (S22 to S26), and it is determined whether the likelihood of a current driving operation is lower than the likelihood threshold value (S28). When the likelihood is lower than the likelihood threshold value, the driver is determined as being in degraded consciousness. Since the likelihood threshold value is calculated for determination on the basis of data when the driver has consciousness, an erroneous determination of degraded consciousness due to difference in driver operation characteristics is suppressed.