Driver State Monitoring · Multimodal Sensing
Measuring driver trust with physiological signals
Evaluating which sensors and signal features can help estimate trust in automated driving.
Research question
Which combination of physiological measurements can help estimate a driver’s trust in an automated vehicle in real time?
Approach
We designed a simulator study using eye tracking, electrocardiography (ECG), electroencephalography (EEG), and electrodermal activity (EDA). We used machine learning classifiers to compare sensor combinations and visualized which signal features contributed most to the predictions.