Analysis and modelling of drivers’ responses to safety-critical lane changes performed by heavy vehicles

Abstract

According to the National Highway Transport Safety Administration (NHTSA), around 240,000-610,000 crashes due to lane change are reported to the police annually. In addition, over 300,000 lane change crashes remain unreported to the police. These high crash numbers indicate that lane change maneuvers pose a high risk to the lane changing and surrounding vehicles. Automated driving can possibly reduce lane change crashes or mitigate their consequences by promoting safer lane changes. In this context, this thesis mainly focuses on the analysis of safety critical lane change maneuvers using Naturalistic Driving Data from the SHRP2 database. The lane changes being considered in this thesis are performed by heavy vehicles, as observed from the perspective of a following vehicle equipped with forward facing camera. A total of 89 cases were found relevant for the scope of the study and manually annotated to extract variables which could describe the interaction between the heavy vehicles changing lane and the following passenger car. The annotated variables were used to quantitatively describe the lane change maneuver and to determine the factors responsible for the start of braking of the following passenger car. The results show that lateral motion of the heavy vehicle and a combination of lateral motion and blinking of the heavy vehicle are the main factor triggering the braking of the following car’s driver. The speed difference of the following vehicle between the points when heavy vehicle starts its lateral motion and when the following vehicle starts to brake also has an impact on the braking of the following vehicle. These results have potential application for the design and testing on critical lane change maneuvers performed by an automated truck.