Driver Drowsiness Prediction Using Driving Time and Inactivity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for predicting driver drowsiness often rely on physical behavior detection, which is prone to external disturbances and lacks precision, typically triggering alerts only after symptoms appear, rather than before they become dangerous.
Innovation Solution
A system using a remote server and on-board device that tracks driving time since last stop, vehicle heading and speed variations, and critical hours to predict drowsiness, customizable to individual sensitivity, with alerts triggered before symptoms occur, and refined through user feedback.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If physical behavior detection methods (facial recognition, steering wheel pulses) are used to detect driver drowsiness, then the system can identify drowsy symptoms, but the system becomes complex and prone to false alarms from external disturbances
Solution Approach 1:
The patent extracts the core predictive factors (driving time, critical hours, driver inactivity) from the complex physical behavior detection systems. Instead of using cameras and steering wheel sensors, the invention uses only vehicle operation data that is already available, eliminating the need for complex detection hardware while maintaining predictive capability.
Solution Approach 2:
The patent creates a simplified model that copies the essential predictive information from complex physical detection systems. By using driving time and inactivity indices as proxies for actual drowsiness state, the system achieves reliable prediction without the complexity of facial recognition or steering wheel pulse detection.
2Loss of time
If physical behavior detection systems are implemented, then drowsiness symptoms can be detected, but the intervention occurs only after symptoms appear rather than before
Solution Approach 1:
The patent performs preliminary action by calculating the driving time before alert (TDAS) in advance based on driving time since last stop and critical hours index. The system predicts when drowsiness will occur and issues warnings before symptoms appear, rather than detecting symptoms after they have already manifested.
Solution Approach 2:
The patent implements feedback by continuously monitoring driving time and inactivity, comparing actual driving duration against the predicted TDAS, and adjusting warnings based on driver responses. This closed-loop feedback enables timely intervention while maintaining high prediction accuracy through continuous refinement.
3Adaptability or versatility
If general drowsiness detection methods are used, then the system can alert drivers, but the system lacks customization to individual driver sensitivity and precision
Solution Approach 1:
The patent makes the system dynamic by allowing TDAS to be customized for each driver based on their individual sensitivity to falling asleep. The driving time before alert is not a fixed value but adapts to each driver's characteristics, enabling both individualization and high precision through personalized thresholds.
4Object-affected harmful factors
If systems intervene after drowsiness symptoms appear, then detection is simpler, but the driver and others are already in dangerous situations
Solution Approach 1:
The patent applies preliminary anti-action by predicting drowsiness onset and issuing warnings before the dangerous state occurs. By calculating TDAS in advance and alerting drivers proactively, the system prevents the harmful effect of drowsy driving before it can manifest, rather than reacting after danger has already arisen.
Data Source
Figure 1
Figure 2
AI summary
The present invention relates to a system and method for predicting a state of drowsiness of a vehicle driver, comprising a remote server (2) and an on-board device (1) on-board same, which are capable of communicating with each other. Said method is characterized in that the on-board device (1) comprises: a means capable of counting the length of time of driving the vehicle since the last stop of the latter; a means capable of determining the time so as to establish an index of the critical timetables; a means capable of periodically determining, over a given time period, at least the consecutive variations in the direction of travel and/or speed of the vehicle so as to establish a driver inactivity index; and a means capable of determining a length of driving time before an alert, which is based on at least said inactivity index and on the critical timetable index.