Calibration of the Fundamental Diagram Based on Loop and Probe Data

Clairais, Aurélien; Duret, Aurélien; Faouzi, Nour-Eddin El · 2016 · Crossref

DOI: 10.3141/2560-03

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the calibration of the Fundamental Diagram (FD), a critical component of traffic flow theory that relates traffic flow and density. Accurate FD parameters are essential for understanding traffic properties and calibrating dynamic traffic simulation models. The authors identify limitations in existing methods, which often rely solely on loop detector data (suffering from mixed traffic states and inability to capture propagation) or trajectory data (which is expensive and difficult to process). To overcome these issues, the study proposes a new method that combines Eulerian loop data with Lagrangian probe data to automatically estimate FD parameters. The methodology relies on the Lighthill-Whitham-Richards (LWR) kinematic wave theory. The authors treat traffic counts from two successive loop detectors as boundary conditions for a homogeneous road link. Using the method of characteristics, they propagate traffic states between these loops to estimate intermediate conditions based on a triangular FD model. The method then simulates probe vehicle trajectories within this estimated traffic field. The optimal FD parameters—specifically wave speed ($w$) and jam density ($k_x$)—are determined by minimizing the discrepancy between the simulated probe positions and actual observed probe travel times. A weighted goodness-of-fit function is used, assigning higher weights to probe vehicles moving at low speeds to better capture traffic dynamics. The method was validated through two simulation studies using synthetic data generated by a mesoscopic LWR model. First, an error-free dataset confirmed the method's accuracy, recovering the exact ground-truth parameters ($u=25$ m/s, $w=5$ m/s, $k_x=0.14$ veh/m). Second, the method was tested on realistic, aggregated, and noisy data to assess robustness. The results indicated that loop data aggregation periods should not exceed 180 seconds to maintain calibration efficiency and accuracy. The method remained stable with probe penetration rates as low as 3%, demonstrating viability for low-penetration scenarios. Interestingly, lower probe data frequencies (e.g., 1/80 s⁻¹) yielded higher calibration accuracy than higher frequencies, as vehicles traversing longer time intervals are more likely to cross multiple shockwaves, providing richer information for parameter estimation. The significance of this work lies in providing a robust, automated framework for calibrating macroscopic traffic models using widely available loop and probe data. By integrating these complementary data sources, the method overcomes the spatial and temporal limitations of single-source approaches. The findings suggest that the proposed algorithm is suitable for operational implementation, potentially enabling online and automatic calibration of fundamental diagram parameters in real-time traffic management systems.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• exposure measurement