SAFER Seminar: Use of Driver Gaze Information for Detecting Risky Lane Changes
Welcome to a SAFER Seminar with presentation by SAFER International Scientific Advisory Board member, Prof. Kazuya Takeda, on current and past research activities!
Prof. Kazuya Takeda is Professor at Graduate School of Information Science, Green Mobility Collaborative Research Center, Nagoya University, Japan.
In this presentation, we utilize driver gaze and vehicle operation behavior for detecting risky lane changes. We break down driver gaze and vehicle operation behavior into several discrete acts, e.g., “looking in right-view mirror,” “braking,” etc., and jointly model the sequences of discrete acts of driver gaze and vehicle operation behavior using multi-stream discrete hidden Markov models. To evaluate the integrated model of driver gaze and vehicle operation behavior, we conduct a risky lane change detection experiment using driving data recorded on real expressways from eleven drivers, which includes a total of 988 lane change scenes. Experimental results show that the integrated model achieves better risky lane change direction performance than the gaze-only and vehicle operation-only models and obtains 88% and 84% true positive rates for the lane changes to the left and right, respectively, with a false positive rate less than 10%.
Speaker: Prof. Kazuya Takeda, Professor, Graduate School of Information Science, Green Mobility Collaborative Research Center, Nagoya University
Host: Anna Nilsson-Ehle, Director, SAFER