Parameters Estimation of a Microscopic Traffic Flow Sub-Model Within a Multiscale Approach Using Experimental Data

Abstract

Future traffic contexts will likely involve the coexistence of human-driven vehicles and connected and automated vehicles (CAVs). To assess the impact of CAVs, especially in large-scale applications, intermediate hybrid multi-scale models can be used. These models are easily adaptable to traffic control strategies by employing disaggregated modeling in regions where such strategies are implemented and macroscopic modeling in other regions indirectly affected by the controlled infrastructure. This paper focuses on a model previously established in the literature, the H - CA&CTM (Hybrid Cellular Automata -CA- Cell Transmission Model-CTM), with an emphasis on the micro model that can be implemented in the hybrid traffic flow model. The research has two primary aims: 1) Investigate the calibration of the CA model with respect to various cell lengths using two distinct approaches: simulating all vehicles together in a closed ring layout and simulating each vehicle using data obtained from its respective follower; 2) Utilize vehicle trajectory data for the calibration procedure, enabling a comprehensive comparison of methods. Two detailed approaches were considered: 1. Measured Leader – Simulated Follower interaction approach. 2. Simulated Leader – Simulated Follower interaction approach. The major finding of the paper is that the calibrated parameters obtained using the Simulated Leader approach display greater regularity across different cell lengths.

Keywords

• Cellular automata
• experimental data
• calibration
• CCAM