Driver Sleepiness Prediction Using Stimulus and Non-Acclimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems fail to accurately predict a driver's sleepiness level in advance, leading to potential drowsiness during driving, especially when transitioning from automatic to manual driving, as external stimuli may take time to be effective.
Innovation Solution
A sleepiness level prediction device and method that acquires vehicle and driver information, calculates the stimulus amount and non-acclimation degree, and uses a Markov model to predict future sleepiness levels based on current levels and transition probabilities, enabling timely interventions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If external stimuli (temperature or olfactory) are given to suppress driver sleepiness, then the driver's awake state is improved, but it takes time until the suppression effect is exhibited
Solution Approach 1:
The system performs preliminary action by predicting future sleepiness levels before the driver actually becomes sleepy. It calculates transition probabilities and estimates sleepiness at future time points, allowing the system to issue warnings or apply stimuli in advance, thereby eliminating the time delay associated with reactive stimuli application.
Solution Approach 2:
The system continuously monitors current sleepiness levels and uses this feedback to update transition probabilities and refine future sleepiness predictions. This closed-loop feedback mechanism allows the system to adapt to the driver's actual state and improve prediction accuracy over time, ensuring timely and effective intervention.
2Measurement precision
If the system monitors driver eye status to detect sleepiness, then current sleepiness level is detected, but it cannot predict future sleepiness level in advance
Solution Approach 1:
The system performs preliminary calculation of future sleepiness levels by computing transition probabilities based on current state and historical data. It estimates sleepiness at future time points before the actual state occurs, enabling advance warning and preventive measures rather than just detecting current state.
Solution Approach 2:
The system dynamically models sleepiness evolution using transition probabilities that capture the stochastic nature of sleepiness progression. It updates predictions based on current monitoring data and adjusts future estimates accordingly, transforming static detection into dynamic prediction.
Data Source
AI summary
A sleepiness level prediction device includes a vehicle information acquisition unit, a driver information acquisition unit, and a control unit. The vehicle information acquisition unit is configured to acquire information relating to a vehicle. The driver information acquisition unit is configured to acquire information relating to a driver. The control unit is configured to calculate a stimulus amount to be given to the driver based on a current driving load of the driver and a driving load after a prescribed time, based on the information relating to the vehicle, calculate a non-acclimation degree indicative of a level at which the driver is not acclimated to driving based on a change in the stimulus amount, and predict a sleepiness level of the driver after the prescribed time based on a current sleepiness level of the driver estimated based on the information relating to the driver and the non-acclimation degree.