Presentation
Situation Awareness Classification Using Multi-modal Sensing in Autonomous Driving
Event Type
Lecture
In-Person
Cognitive Engineering & Decision Making
TimeTuesday, October 5th10:48am - 11:06am EDT
LocationGrand Salon V
DescriptionIn Society of Automotive Engineers (SAE) Level 3 automated vehicle systems, it is critical for drivers to maintain sufficient situation awareness (SA) to deal with situations that automated vehicle systems are not designed to handle. This study demonstrates a multi-modal sensing approach to measure drivers’ SA without interrupting the on-going task. The electroencephalogram (EEG) and eye tracking (ET) were recorded from 30 participants during driving tasks. Their non-driving related secondary tasks varied in terms of the location of visual attention before a take-over request (TOR). In the post-TOR segment, participants were expected to gather information from driving environment in preparation for a takeover and were asked a series of questions about this knowledge. Results suggest that our multi-modal sensing model can predict SA with accuracy of 81%. Future integration of this sensing approach can inform vehicle of the driver's SA and ultimately improve safety in SAE Level 3 automated driving.
Authors
Ph.D. Student
Ph.D. Student
Assistant Professor