Derivation, calibration and verification of macroscopic model for urban traffic flow. Part 2

TL;DR

This paper presents a unified calibration method for a second-order macroscopic traffic model using stationary detector data and GPS traces, effectively capturing different traffic phases and outperforming first-order models in accuracy.

Contribution

It introduces a novel calibration procedure combining detector data and GPS traces for second-order traffic models, validated on real-world Moscow data.

Findings

The calibrated model accurately represents free flow, synchronized flow, and wide moving jams.

The second-order model outperforms the first-order LWR model in simulation accuracy.

Validation on Moscow data confirms the model's effectiveness in real traffic scenarios.

Abstract

In this paper, we propose a unified procedure for calibration of macroscopic second-order multilane traffic models. The focus is on calibrating the fundamental diagram using the combination stationary detector data and GPS traces. GPS traces are used in estimation of the deceleration wave speed. Thus calibrated model adequately represents the three phases of traffic: free flow, synchronized flow and the wide moving jam. The proposed approach was validated in simulation using stationary detection data and GPS traces from the Moscow Ring Road. Simulation showed that the proposed second-order model is more accurate than the first-order LWR model.