Capturing Drivers’ Lane Changing Behaviors on Operational Level by Data Driven Methods

With the development of autonomous vehicles, advanced driver-assistance systems, and vehicular social networks, the requirement for vehicle trajectory prediction in lane changing is higher. Here, we propose a neural-network-based operational level lane-changing model using data-driven methods. First, we determine the inputs of the lane-changing model by analyzing the influence factors of the lanechanging behavioral model under the framework of the social force theory. The main influencing factors include the temporal destination, the historical trajectories of the lane-changing vehicle, and the relative distance between the surrounding vehicles in our lane-changing model. Our lane-changing model is built by deep neural networks (DNNs). Then, we determine the suitable network structure and other parameters by empirical data for the DNN. Finally, tests on empirical lane-changing trajectory data sets show that the operational level lane-changing model built by DNN is promising.

Abstract licence: CC BY-NC-ND

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.