Evaluating the Impact of Heavy Vehicles on Lane-changing Decisions of Car Drivers: A Neural-network-based Methodology

From common driving experiences, car drivers are unwilling to follow heavy vehicles owing to speed and visibility obstructions, and most of them choose to change lanes to mitigate such obstructions. In literature, however, the impact of heavy vehicles on lane-changing decisions of following car drivers has rarely been investigated. In this study, a methodology for shedding light on such impact is proposed, which is based on two neural network models-the lane-changing model and the vehicle conversion model. Estimated by the large-scale trajectory data, the models both acquire high accuracy. By converting the nearest lead vehicles to heavy vehicles and detecting driving decisions of the following car drivers, the percentage of changed driving decisions under traffic conditions with the different proportion of heavy vehicles is clearly given, and the number of increased lane changes caused by heavy vehicles is quantitatively provided, as well.