Abstract In order to enhance the realism and diversity of traffic flow modelling, this chapter presents a data-driven traffic behavior model based on neural networks in a real-virtual interaction traffic simulation system. First, we extract individual personalized real trajectories from each vehicle, then use neural networks to develop specific traffic models from the trajectories of each vehicle. In contrast to traditional, manually-defined traffic models, we aim to develop a data-driven model to describe the relationship between the traffic states faced by a driver and the driver’s resultant actions. In this model, a driver’s behavior is influenced by the current traffic states of the leading vehicle and the following vehicle. This is a regression problem for which the inputs of the model are the traffic states of the leading and following vehicles. The output is the action of the current vehicle. Finally, this chapter presents a real-virtual interaction system. In detail, real trajectories are introduced directly into the simulation process to maximize the characteristics of real traffic flow. In comparison to existing simulation methods, traffic flows simulated by this method can depict irregular vehicle driving behavior.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Vehicle Behaviors Simulation Technology Based on Neural Network


    Contributors:
    Yang, Xin (author) / Li, Shuai (author) / Yin, Baocai (author) / Zhang, Qiang (author) / Tan, Guozhen (author) / Zhou, Dongsheng (author) / Wei, Xiaopeng (author)


    Publication date :

    2018-10-11


    Size :

    18 pages





    Type of media :

    Article/Chapter (Book)


    Type of material :

    Electronic Resource


    Language :

    English




    Evaluation of Airline Pilots' Safety Behaviors Based on Fuzzy Neural Network

    Zhang, Guozheng / Chen, Dongfeng / Feng, Yu | IEEE | 2013



    Vehicle localization based on neural network

    LI DALONG / FROEHLICH DOMINIK / MUHAREMOVIC IBRO | European Patent Office | 2021

    Free access