Close

Presentation

Driver Situational Awareness Prediction During Takeover Transitions: A Multimodal Machine Learning Approach
DescriptionThis study aims to predict drivers' situational awareness (SA) during takeover transitions in Level 3 conditionally automated vehicles using a multimodal machine learning approach. We collected data from 264 takeover events involving 44 participants in a driving simulator. The data included drivers' characteristics, physiological states, eye movements, and environmental attributes. Using feature selection, grid search, hyperparameter tuning, and time window optimization, we evaluated various machine learning models including Logistic Regression, Random Forest, Linear Discriminant Analysis, Extreme Gradient Boosting, Support Vector Machine (SVM), and Neural Network. The SVM model, with a 3-second post-takeover request (TOR) and a 1-second pre-TOR time window, achieved the best performance with a macro F1 score of 0.75 and an accuracy of 0.77. This model enables quick and accurate predictions of drivers' SA, enhancing the safety and efficiency of autonomous driving takeovers. This research contributes to the design and implementation of advanced driver monitoring and support systems.
Event Type
Lecture
TimeTuesday, September 10th4:15pm - 4:35pm MST
LocationFLW Salon G
Tracks
Surface Transportation