Backpressure control with estimated queue lengths for urban network traffic

Li, Li; Okoth, Victor; Jabari, Saif Eddin · 2021 · OpenAlex-citations

DOI: 10.1049/itr2.12027

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 practical limitation of Backpressure (BP) control in urban traffic networks, specifically its reliance on perfect knowledge of queue lengths. While BP control offers provable network stability and scalability, existing variants assume fully connected environments where traffic volumes are accurately known. The authors identify a research gap regarding the robustness of BP when queue length data is incomplete or imperfect, which is the current reality in partially connected vehicle environments. To bridge this gap, the study proposes a system called Backpressure with Estimated Queue Lengths (BP-EQ), which integrates BP control with a traffic state estimation module. The methodology involves a three-step process to estimate queue lengths using data from connected vehicle (CV) probes. First, the road network is discretized into cells, and a speed interpolation technique is applied to create a real-time speed map using available CV data, weighted by spatial and temporal proximity. Second, these estimated speeds are converted into traffic densities using the invertible Newell-Franklin speed-density relationship. Third, queue lengths are calculated from the estimated densities and used as inputs for the BP pressure calculations. The authors tested this approach using microscopic traffic simulations with field-calibrated data, evaluating performance across varying CV penetration rates (10%, 20%, 30%, and 100%). The results demonstrate that BP-EQ is robust to incomplete data. Even with a low connected vehicle penetration rate of 10%, BP-EQ outperformed both a real-world adaptive signal controller and an optimized fixed timing controller in terms of average delay, throughput, and maximum stopped queue lengths under high-demand scenarios. The study also validated the accuracy of the speed interpolation technique, showing that estimated speed maps closely matched ground truth data at higher penetration rates. The findings indicate that BP control does not require perfect queue length information to be effective; rather, it can leverage estimated traffic states derived from sparse probe data to achieve superior performance compared to traditional control methods. The significance of this work lies in its demonstration that BP control is viable for current, partially connected traffic environments, rather than requiring a future fully connected infrastructure. By proving that BP-EQ outperforms adaptive and fixed controllers with minimal CV penetration, the paper suggests a practical pathway for implementing distributed, stable traffic control systems. This approach reduces the dependency on expensive, dense sensor infrastructure while maintaining high network efficiency. The study highlights the potential of combining distributed control algorithms with simple estimation techniques to handle real-world data limitations, offering a scalable solution for urban traffic management.

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.

• traffic density