Detecting Critical Mismatched Driver Visual Attention During Lane Change: An Embedding Kernel Algorithm

The visual behavior before and during a lane change maneuver is critical to driving safety. This paper aims to link the mismatched driver visual attention during lane changes with crash risk. A novel Embedding Kernel (EK) algorithm is proposed to identify the critical moment at which the driver’s eye-glance pattern differs significantly between safety-critical events (SCEs) and normal driving maneuvers. The embedding approach translates the eye-glance patterns into embedding representations containing information on the timing and location of eye glances. A subsequent two-sample kernel test is applied to identify the safety-critical moment. The proposed EK algorithm was implemented on lane change events from the Second Strategic Highway Research Program Naturalistic Driving Study. The analysis revealed that the duration and number of safety-critical moments could vary depending on factors such as the driver’s age, the time of day, traffic conditions, driving speed, and lane change direction. Young drivers’ two critical moments happen (3.2 s to 3.4 s and 4.4 s to 4.6 s from the beginning of events, i.e. 5 s before lane change initiation) later than middle-aged drivers one critical moment (0.8 s to 1.2 s). The study shows that mismatched driver visual attention differs significantly between SCEs and normal driving. This finding is essential for deploying safety countermeasures such as the driver monitoring system and the advanced driver assistance system.