Zhang, Wenqi (Tsinghua University), Qin, Yanjun (Tsinghua University), Tao, Xiaoming (Tsinghua University)

FATCM: Frequency-Aware Temporal Convolution Model for Driver Risk Responsiveness Detection Based on EEG

Scheduled for presentation during the Poster Session "Human Drivers in Intelligent Transportation Systems" (FrAT15), Friday, September 27, 2024, 10:30−12:30, Foyer

2024 IEEE 27th International Conference on Intelligent Transportation Systems (ITSC), September 24- 27, 2024, Edmonton, Canada

This information is tentative and subject to change. Compiled on October 3, 2024

Abstract

Traffic safety is closely related to the driver's state, which includes the driver's ability to perceive risk. Most relevant studies use human data. However, research based on facial data is subject to interference from occlusions and lighting conditions, and studies based on electroencephalogram (EEG) signals face challenges in maintaining consistent detection accuracy across subjects. This paper is dedicated to the study of drivers' risk responsiveness during the driving process and proposes a frequency-aware temporal convolutional model based on EEG data. We employed a dataset of drivers' sustained attention and conducted a series of experiments. The experimental results demonstrate that our model achieves a high accuracy in detecting drivers' risk responsiveness, reaching up to 82.44%, which is a 4.54% improvement over the best baseline. This study advances the work on investigating human factors based on physiological signals and is of assistance in preventing traffic accidents and reducing driving risks.