Recognition method of construction conflict based on driver’s eye movement

•Correlation coefficients between blink moments, gaze change angles, pupil diameters and construction conflicts are calculated and analyzed. Key points of the conflict have been defined through analyzing the relationship between peak points of pupil diameter, mean pupil diameter, blink moments, and driver’s mental load.•Key points and the GSA are used to explore the construction conflict fast recognition method.•It is verified that recognizing construction conflicts using pupil diameters and blink moments is feasible. And the higher simulated construction conflict recognition speed and more accurate positions of proposed CCFRM are verified. Drivers eye movement data in simulated construction conflicts at different speeds were collected and analyzed to find the relationship between the drivers’ eye movement and the construction conflict. On the basis of the relationship between the drivers’ eye movement and the construction conflict, the peak point of wavelet processed pupil diameter, the first point on the left side of the peak point and the first blink point after the peak point are selected as key points for locating construction conflict periods. On the basis of the key points and the GSA, a construction conflict recognition method so called the CCFRM is proposed. And the construction conflict recognition speed and location accuracy of the CCFRM are verified. The good performance of the CCFRM verified the feasibility of proposed key points in construction conflict recognition.