Data-Driven Approach for Modeling the Nonlane-Based Mixed Traffic Conditions

The diverse nature of vehicle categories and the resultant lane discipline in mixed (heterogeneous) traffic cause complex spatial interactions. As a result, driving behavior process conditions is meaningfully different, where both longitudinal lateral movements vehicles continuously occur. Under prevailing homogeneous developed countries, partially discrete, following outboard lane-change models can model behavior. However, established car-following cannot be directly used shaping conditions. Such also warrant use high-quality microlevel vehicular trajectory data. Accordingly, realizing this need, data for different flow were developed. to extract parameters required modeling vehicles’ positions using machine learning algorithms. Three supervised algorithms (k-NN, random forest, regression tree) deep are selected which influence identified Spearman correlation analysis. Furthermore, simulation runs performed python language. performance evaluated at microscopic macroscopic levels relevant indicators. results show that k-NN tend replicate better-mixed than forest trees.