pith. sign in

arxiv: 2509.24725 · v3 · pith:VVMEEJK5new · submitted 2025-09-29 · 💻 cs.LG · cs.AI

Q-Net: Queue Length Estimation via Kalman-based Neural Networks

Ting Gao , Elvin Isufi , Winnie Daamen , Erik-Sander Smits , Serge Hoogendoorn This is my paper

Pith reviewed 2026-05-21 22:05 UTC · model grok-4.3

classification 💻 cs.LG cs.AI
keywords queue length estimationkalman filterneural networkstraffic data fusionsignalized intersectionsfloating car datastate-space models
0
0 comments X

The pith

Q-Net fuses loop-detector counts and floating-car speeds inside a state-space Kalman model to estimate queue lengths at signalized intersections.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper sets out to show that a state-space formulation can integrate two privacy-preserving but mismatched traffic data streams—aggregated loop counts near stop lines and segment-wise average speeds from floating cars—into reliable queue length estimates. It does this by embedding an AI-augmented Kalman filter that learns time-varying gains while keeping the underlying state-evolution and measurement equations physically interpretable. A sympathetic reader would care because accurate, real-time queue estimates matter for signal timing, congestion management, and avoiding the cost of cameras or radar. The design also groups floating-car measurements into fixed-size local blocks so the number of trainable parameters stays independent of road length, improving spatial transferability across different intersections.

Core claim

Q-Net follows the standard Kalman predict-update cycle inside a state-space model that explicitly accounts for violations of traffic conservation. An AI component augments the filter by learning the time-varying Kalman gains directly from the two heterogeneous data sources. This structure preserves physical meaning in both the state transition and the measurement equations, supports real-time execution, and produces queue estimates that track formation and dissipation more accurately than baseline methods while reducing the impact of floating-car data delays.

What carries the argument

The AI-augmented Kalman filter that learns time-varying gains from loop-detector counts and aggregated floating-car speeds while retaining explicit state-evolution and measurement equations.

If this is right

  • Queue formation and dissipation can be tracked in real time at signalized intersections without additional sensing hardware.
  • Delays caused by the lower temporal resolution of floating-car data are reduced.
  • The same trained model can be transferred to intersections of different lengths because the number of parameters does not grow with section size.
  • Physical interpretability of state and measurement models is retained while still benefiting from data-driven gain adaptation.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same structure could be tested on other traffic estimation tasks that combine fixed and mobile sensors, such as travel-time prediction on arterials.
  • If the learned gains generalize across cities, agencies could deploy the method with minimal local calibration.
  • Extending the state vector to include turning movements would be a direct next step to handle more complex intersections.

Load-bearing premise

The state-space equations plus the learned gains are flexible enough to capture real queue dynamics even when traffic conservation does not hold exactly.

What would settle it

Compare Q-Net queue estimates against independent high-resolution ground truth, such as video-based vehicle trajectories, during periods of rapid queue growth and clearance; large systematic errors would refute the claim.

Figures

Figures reproduced from arXiv: 2509.24725 by Elvin Isufi, Erik-Sander Smits, Serge Hoogendoorn, Ting Gao, Winnie Daamen.

Figure 1
Figure 1. Figure 1: Layout of a signalized intersection, including the definitions of a lane, section, and ground truth queue [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Queue estimation from net vehicle accumulation under unobserved traffic flow conditions for section [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: At time t, the Q-Net filtering framework takes three inputs: the previous a posteriori state estimate xˆt−1|t−1, cumulative vehicle counts At, Dt, and aFCD speed measurements yt, and produces one output: the current a posteriori state estimate ˆxt|t . Predict: At and Dt are used to derive the control input (queue change) ut, which predicts the a priori queue length xt|t−1 and a priori speed measurements ˆy… view at source ↗
Figure 4
Figure 4. Figure 4: KalmanNet architecture for dynamic Kalman gain estimation. The top panel illustrates the overall [PITH_FULL_IMAGE:figures/full_fig_p010_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Local measurement grouping strategy for transferable Kalman gain estimation. aFCD speed measurements [PITH_FULL_IMAGE:figures/full_fig_p012_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: Satellite view of N1-IN. The downstream area is divided into the study section (highlighted in dark red) and a grade-separated interchange. This interchange links to an open bridge, which serves as a critical entry point to the Rotterdam urban area. Therefore, the unobserved traffic flow in this region is significant. 1st split 2nd split 6th split ... ... ... 12 weekdays 4 weekend days Train Test Validatio… view at source ↗
Figure 7
Figure 7. Figure 7: Six-fold data split: for each split, samples from both weekdays and weekends were randomly selected to [PITH_FULL_IMAGE:figures/full_fig_p014_7.png] view at source ↗
Figure 8
Figure 8. Figure 8: Box plots of absolute estimation errors for baseline models (OSD, ISC) and Q-Net for section [PITH_FULL_IMAGE:figures/full_fig_p015_8.png] view at source ↗
Figure 9
Figure 9. Figure 9: The aFCD visualization and queue length estimation comparison among baselines (OSD, ISC, and Tree [PITH_FULL_IMAGE:figures/full_fig_p016_9.png] view at source ↗
Figure 10
Figure 10. Figure 10: Satellite view of N3-IN and N3-OUT. In both cases, the downstream area divides into an underground study section and an above-ground branch, which merge at the downstream intersection [PITH_FULL_IMAGE:figures/full_fig_p017_10.png] view at source ↗
Figure 11
Figure 11. Figure 11: Box plots of absolute estimation errors across all test datasets using different training sections. [PITH_FULL_IMAGE:figures/full_fig_p017_11.png] view at source ↗
Figure 12
Figure 12. Figure 12: Queue length estimation for section N3-IN and N3-OUT using a model trained on N1-IN without fine￾tuning (random seed 42). location-specific traffic dynamics. Performances on N3-IN and N3-OUT are close regardless of which of these two sections is used for training, attributable to their spatial proximity and shared infrastructure characteristics. Interestingly, the model trained on N1-IN generalizes better… view at source ↗
Figure 13
Figure 13. Figure 13: Comparison of queue length estimation for section [PITH_FULL_IMAGE:figures/full_fig_p019_13.png] view at source ↗
Figure 14
Figure 14. Figure 14: Queue length estimation results for different model variants on 2023-11-22 for section [PITH_FULL_IMAGE:figures/full_fig_p020_14.png] view at source ↗
read the original abstract

Estimating queue lengths at signalized intersections is a long-standing challenge in traffic management. Partial observability of vehicle flows complicates this task despite the availability of two privacy-preserving data sources: (i) aggregated vehicle counts from loop detectors near stop lines, and (ii) aggregated floating car data (aFCD) that provide segment-wise average speed measurements. However, how to integrate these sources with differing spatial and temporal resolutions for queue length estimation is rather unclear. Addressing this question, we present Q-Net: a queue estimation framework built upon a state-space formulation. This design addresses key challenges in queue modeling, such as violations of traffic conservation assumptions. Q-Net follows the Kalman predict-update structure and maintains physical interpretability in both the state evolution and measurement models. Q-Net uses an AI-augmented Kalman filter to learn time-varying gain dynamics from data. The framework supports real-time implementation and improves spatial transferability by grouping aFCD measurements into fixed-size local groups, making the number of learnable parameters independent of section length. Evaluations on urban main roads in Rotterdam, the Netherlands, show that Q-Net outperforms baseline methods, tracks queue formation and dissipation accurately, and mitigates aFCD-induced delays. By combining data efficiency, interpretability, real-time applicability, and spatial transferability, Q-Net makes accurate queue length estimation possible without costly sensing infrastructure like cameras or radar.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

3 major / 3 minor

Summary. The manuscript introduces Q-Net, a queue length estimation framework for signalized intersections that fuses loop-detector counts with aggregated floating-car data (aFCD) speeds inside a state-space model. It augments the standard Kalman predict-update cycle with a neural network that learns only the time-varying gains, while keeping the state-evolution and measurement equations physically interpretable. The design is claimed to mitigate violations of traffic conservation, support real-time operation, and achieve spatial transferability by grouping aFCD measurements into fixed-size local blocks whose parameter count is independent of section length. Experiments on urban roads in Rotterdam are reported to show that Q-Net outperforms baseline methods and tracks queue formation and dissipation accurately.

Significance. If the empirical claims are substantiated with detailed metrics, the work would offer a practical advance in traffic-state estimation: an interpretable, data-efficient alternative to camera- or radar-based systems that can be deployed with existing privacy-preserving sensors. The restriction of the neural component to gain learning, together with the fixed-size grouping mechanism, addresses two common obstacles to real-world adoption—loss of physical meaning and poor spatial generalization—while preserving the real-time Kalman structure.

major comments (3)
  1. [Abstract and §5] Abstract and §5 (Evaluation): the abstract states that evaluations on Rotterdam roads show outperformance and accurate tracking, yet the manuscript provides no quantitative metrics (RMSE, MAE, error bars), baseline details, ablation results, or statistical significance tests. Without these, the central empirical claim cannot be assessed.
  2. [§3] §3 (Model Formulation): the assertion that the AI-augmented structure addresses violations of traffic conservation assumptions rests on the design choice alone; no explicit verification (e.g., residual analysis of flow conservation or comparison against a standard Kalman filter on the same data) is shown. This is load-bearing for the interpretability claim.
  3. [§4] §4 (Data and Implementation): the description of how the two heterogeneous data sources are synchronized and how the fixed-size grouping is performed lacks sufficient detail on temporal alignment, missing-data handling, and the exact neural-network architecture used for the gains. These choices directly affect reproducibility and the claimed spatial transferability.
minor comments (3)
  1. [§2] Notation: define the state vector x_k and the measurement vector z_k explicitly, including units, at the first appearance in §2 or §3.
  2. [Figures] Figures: ensure all queue-length time-series plots include the ground-truth reference, multiple runs or confidence bands, and clear axis labels with units.
  3. [References] References: add citations to recent Kalman-filter traffic papers that also learn gains or parameters to better situate the novelty.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for the constructive and detailed review. The comments highlight important areas for strengthening the empirical support, interpretability claims, and reproducibility of Q-Net. We address each major comment below and will incorporate the suggested improvements in the revised manuscript.

read point-by-point responses

  1. Referee: [Abstract and §5] Abstract and §5 (Evaluation): the abstract states that evaluations on Rotterdam roads show outperformance and accurate tracking, yet the manuscript provides no quantitative metrics (RMSE, MAE, error bars), baseline details, ablation results, or statistical significance tests. Without these, the central empirical claim cannot be assessed.

    Authors: We agree that the current presentation of results is insufficient for assessing the central claims. In the revision we will expand both the abstract and §5 to report RMSE, MAE, error bars, full baseline specifications, ablation studies, and statistical significance tests on the Rotterdam dataset. revision: yes

  2. Referee: [§3] §3 (Model Formulation): the assertion that the AI-augmented structure addresses violations of traffic conservation assumptions rests on the design choice alone; no explicit verification (e.g., residual analysis of flow conservation or comparison against a standard Kalman filter on the same data) is shown. This is load-bearing for the interpretability claim.

    Authors: The referee is correct that design alone does not substantiate the claim. We will add to §3 residual analysis of flow conservation and side-by-side comparisons against a standard Kalman filter on the same data to provide explicit verification of how the learned gains mitigate conservation violations. revision: yes

  3. Referee: [§4] §4 (Data and Implementation): the description of how the two heterogeneous data sources are synchronized and how the fixed-size grouping is performed lacks sufficient detail on temporal alignment, missing-data handling, and the exact neural-network architecture used for the gains. These choices directly affect reproducibility and the claimed spatial transferability.

    Authors: We accept that additional implementation details are required. The revised §4 will specify the temporal alignment procedure between loop-detector counts and aFCD speeds, the strategy for missing-data imputation, and the precise neural-network architecture (layers, activations, hyperparameters, and training protocol) used to predict the time-varying gains. revision: yes

Circularity Check

0 steps flagged

No significant circularity

full rationale

The paper derives Q-Net from a standard Kalman predict-update structure applied to a physically motivated state-space model that fuses loop-detector counts and aFCD speeds. The neural augmentation is restricted to learning time-varying gains from the two data sources, which does not reduce any claimed prediction or result to the inputs by construction. No self-citations, uniqueness theorems, or ansatzes imported from prior author work appear in the derivation chain. The framework remains self-contained: the state evolution and measurement models preserve interpretability independently of the learned gains, and spatial transferability follows directly from the fixed-size grouping design rather than from any fitted parameter renamed as output.

Axiom & Free-Parameter Ledger

1 free parameters · 1 axioms · 0 invented entities

The approach relies on a state-space model that is assumed to handle traffic dynamics despite conservation violations, plus learnable neural components whose training data and stability are not detailed in the abstract.

free parameters (1)
  • neural network parameters for time-varying Kalman gains
    Learned from data to capture dynamics; number made independent of section length via grouping.
axioms (1)
  • domain assumption State-space model can represent queue dynamics and address violations of traffic conservation assumptions
    Explicitly stated as addressing key challenges in queue modeling.

pith-pipeline@v0.9.0 · 5787 in / 1283 out tokens · 33777 ms · 2026-05-21T22:05:33.089443+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

53 extracted references · 53 canonical work pages

  1. [1]

    Multivariate time-varying kalman filter approach for cycle-based maximum queue length estimation

    Wanuji Abewickrema, Mehmet Yildirimoglu, and Jiwon Kim. Multivariate time-varying kalman filter approach for cycle-based maximum queue length estimation. Transportation research part C: emerging technologies, 154: 0 104238, 2023

    work page 2023

  2. [2]

    Examining queue-jumping phenomenon in heterogeneous traffic stream at signalized intersection using uav-based data

    Afzal Ahmed, Fatma Outay, Syeda Ofaq Raza Zaidi, Muhammad Adnan, and Dong Ngoduy. Examining queue-jumping phenomenon in heterogeneous traffic stream at signalized intersection using uav-based data. Personal and Ubiquitous Computing, 25: 0 93--108, 2021

    work page 2021

  3. [3]

    Video-based traffic queue length estimation

    Antonio Albiol, Alberto Albiol, and Jose M Mossi. Video-based traffic queue length estimation. In 2011 IEEE International Conference on Computer Vision Workshops (ICCV Workshops), pages 1928--1932. IEEE, 2011

    work page 2011

  4. [4]

    Detecting queue length on urban arterials using floating car data (FCD)

    Oru c Alt nta s . Detecting queue length on urban arterials using floating car data (FCD). PhD thesis, Middle East Technical University (Turkey), 2018

    work page 2018

  5. [5]

    Real-time queue length estimation using event-based advance detector data

    Chengchuan An, Yao-Jan Wu, Jingxin Xia, and Wei Huang. Real-time queue length estimation using event-based advance detector data. Journal of Intelligent Transportation Systems, 22 0 (4): 0 277--290, 2018

    work page 2018

  6. [6]

    A survey and comparison of low-cost sensing technologies for road traffic monitoring

    Marcin Bernas, Bart omiej P aczek, Wojciech Korski, Piotr Loska, Jaros aw Smy a, and Piotr Szyma a. A survey and comparison of low-cost sensing technologies for road traffic monitoring. Sensors, 18 0 (10): 0 3243, 2018

    work page 2018

  7. [7]

    Latent-kalmannet: Learned kalman filtering for tracking from high-dimensional signals

    Itay Buchnik, Guy Revach, Damiano Steger, Ruud JG Van Sloun, Tirza Routtenberg, and Nir Shlezinger. Latent-kalmannet: Learned kalman filtering for tracking from high-dimensional signals. IEEE Transactions on Signal Processing, 72: 0 352--367, 2023

    work page 2023

  8. [8]

    Shock wave approach for estimating queue length at signalized intersections by fusing data from point and mobile sensors

    Qing Cai, Zhongyu Wang, Lingyu Zheng, Bing Wu, and Yinhai Wang. Shock wave approach for estimating queue length at signalized intersections by fusing data from point and mobile sensors. Transportation Research Record, 2422 0 (1): 0 79--87, 2014

    work page 2014

  9. [9]

    Driving behavior at signalized intersections operating under disordered traffic conditions

    Ritvik Chauhan, Ashish Dhamaniya, and Shriniwas Arkatkar. Driving behavior at signalized intersections operating under disordered traffic conditions. Transportation research record, 2675 0 (12): 0 1356--1378, 2021

    work page 2021

  10. [10]

    Cycle-by-cycle queue length estimation for signalized intersections using sampled trajectory data

    Yang Cheng, Xiao Qin, Jing Jin, Bin Ran, and Jason Anderson. Cycle-by-cycle queue length estimation for signalized intersections using sampled trajectory data. Transportation research record, 2257 0 (1): 0 87--94, 2011

    work page 2011

  11. [11]

    An exploratory shockwave approach to estimating queue length using probe trajectories

    Yang Cheng, Xiao Qin, Jing Jin, and Bin Ran. An exploratory shockwave approach to estimating queue length using probe trajectories. Journal of intelligent transportation systems, 16 0 (1): 0 12--23, 2012

    work page 2012

  12. [12]

    Improving noise assessment at intersections by modeling traffic dynamics

    Estelle Chevallier, Arnaud Can, M Nadji, and Ludovic Leclercq. Improving noise assessment at intersections by modeling traffic dynamics. Transportation Research Part D: Transport and Environment, 14 0 (2): 0 100--110, 2009

    work page 2009

  13. [13]

    Split-kalmannet: A robust model-based deep learning approach for state estimation

    Geon Choi, Jeonghun Park, Nir Shlezinger, Yonina C Eldar, and Namyoon Lee. Split-kalmannet: A robust model-based deep learning approach for state estimation. IEEE transactions on vehicular technology, 72 0 (9): 0 12326--12331, 2023

    work page 2023

  14. [14]

    Real-time queue-end detection on freeways with floating car data: Practice-ready algorithm

    Tu-Uyen J Dinh, Romain Billot, Eric Pillet, and Nour-Eddin El Faouzi. Real-time queue-end detection on freeways with floating car data: Practice-ready algorithm. Transportation Research Record, 2470 0 (1): 0 46--56, 2014

    work page 2014

  15. [15]

    Traffic detector handbook: Third edition---volume ii

    Federal Highway Administration . Traffic detector handbook: Third edition---volume ii. Technical Report FHWA-HRT-06-139, U.S. Department of Transportation, Federal Highway Administration, Washington, DC, 2006. URL https://www.fhwa.dot.gov/publications/research/operations/its/06139/chapt5b.cfm. Accessed: 2025-09-18

    work page 2006

  16. [16]

    Identification and analysis of queue spillovers in city street networks

    Nikolas Geroliminis and Alexander Skabardonis. Identification and analysis of queue spillovers in city street networks. IEEE Transactions on Intelligent Transportation Systems, 12 0 (4): 0 1107--1115, 2011

    work page 2011

  17. [17]

    Multi-sensor traffic data fusion for congestion detection and tracking

    J Gitahi, M Hahn, M Storz, C Bernhard, M Feldges, and R Nordentoft. Multi-sensor traffic data fusion for congestion detection and tracking. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 43: 0 173--180, 2020

    work page 2020

  18. [18]

    A queue length estimation and prediction model for long freeway off-ramps

    Seiran Heshami and Lina Kattan. A queue length estimation and prediction model for long freeway off-ramps. Journal of Intelligent Transportation Systems, 25 0 (1): 0 122--134, 2021

    work page 2021

  19. [19]

    Lessons learned from field operational test of integrated network management in amsterdam

    SP Hoogendoorn, J Van Kooten, and R Adams. Lessons learned from field operational test of integrated network management in amsterdam. Transportation Research Record, 2554 0 (1): 0 111--119, 2016

    work page 2016

  20. [20]

    High time-resolution queue profile estimation at signalized intersections based on extended kalman filtering

    Simon Hu, Qishen Zhou, Junyi Li, Yibing Wang, Claudio Roncoli, Lihui Zhang, and Lewis Lehe. High time-resolution queue profile estimation at signalized intersections based on extended kalman filtering. IEEE Transactions on Intelligent Transportation Systems, 23 0 (11): 0 21274--21290, 2022

    work page 2022

  21. [21]

    Subcycle waveform modeling of traffic intersections using recurrent attention networks

    Yashaswi Karnati, Rahul Sengupta, Anand Rangarajan, and Sanjay Ranka. Subcycle waveform modeling of traffic intersections using recurrent attention networks. IEEE Transactions on Intelligent Transportation Systems, 23 0 (3): 0 2538--2548, 2021

    work page 2021

  22. [22]

    An advanced deep learning approach to real-time estimation of lane-based queue lengths at a signalized junction

    Seunghyeon Lee, Kun Xie, Dong Ngoduy, and Mehdi Keyvan-Ekbatani. An advanced deep learning approach to real-time estimation of lane-based queue lengths at a signalized junction. Transportation research part C: emerging technologies, 109: 0 117--136, 2019

    work page 2019

  23. [23]

    Backpressure control with estimated queue lengths for urban network traffic

    Li Li, Victor Okoth, and Saif Eddin Jabari. Backpressure control with estimated queue lengths for urban network traffic. IET Intelligent Transport Systems, 15 0 (2): 0 320--330, 2021

    work page 2021

  24. [24]

    Kalman filter and its application

    Qiang Li, Ranyang Li, Kaifan Ji, and Wei Dai. Kalman filter and its application. In 2015 8th international conference on intelligent networks and intelligent systems (ICINIS), pages 74--77. IEEE, 2015

    work page 2015

  25. [25]

    On kinematic waves i

    Michael James Lighthill and G Be Whitham. On kinematic waves i. flood movement in long rivers. ii a theory of traffic flow on long crowded roads. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, 229 0 (1178): 0 281--345, 1955

    work page 1955

  26. [26]

    Real-time queue length estimation for congested signalized intersections

    Henry X Liu, Xinkai Wu, Wenteng Ma, and Heng Hu. Real-time queue length estimation for congested signalized intersections. Transportation research part C: emerging technologies, 17 0 (4): 0 412--427, 2009

    work page 2009

  27. [27]

    A multisource data approach for estimating vehicle queue length at metered on-ramps

    Xiaoling Luo, Xiaobo Ma, Matthew Munden, Yao-Jan Wu, and Yangsheng Jiang. A multisource data approach for estimating vehicle queue length at metered on-ramps. Journal of Transportation Engineering, Part A: Systems, 148 0 (2): 0 04021117, 2022

    work page 2022

  28. [28]

    A model of air pollution from road traffic, based on the characteristics of interrupted flow and junction control: Part i—model description

    Athanasios Matzoros and Dirck Van Vliet. A model of air pollution from road traffic, based on the characteristics of interrupted flow and junction control: Part i—model description. Transportation Research Part A: Policy and Practice, 26 0 (4): 0 315--330, 1992

    work page 1992

  29. [29]

    Floating car data (fcd)

    Mobiliteitsscan . Floating car data (fcd). URL https://mobiliteitsscan-info.nl/analyse/modulegroepen/modulegroep-externe-bronnen/floating-car-data/. Accessed: 2025-05-21

    work page 2025

  30. [30]

    u rgen M \

    J \"u rgen M \"U CK. Using detectors near the stop-line to estimate traffic flows. Traffic engineering & control, 43 0 (11): 0 429--434, 2002

    work page 2002

  31. [31]

    i am sick and tired of this congestion

    Haidar Nadrian, Mohammad Hossein Taghdisi, Kowsar Pouyesh, Maryam Khazaee-Pool, and Towhid Babazadeh. “i am sick and tired of this congestion”: Perceptions of sanandaj inhabitants on the family mental health impacts of urban traffic jam. Journal of Transport & Health, 14: 0 100587, 2019

    work page 2019

  32. [32]

    Optimal control for reducing congestion and improving safety in freeway systems

    Cecilia Pasquale, Simona Sacone, Silvia Siri, and Markos Papageorgiou. Optimal control for reducing congestion and improving safety in freeway systems. IEEE Transactions on Intelligent Transportation Systems, 19 0 (11): 0 3613--3625, 2018

    work page 2018

  33. [33]

    Recursive constrained state estimation using modified extended kalman filter

    Jagdeesan Prakash, Biao Huang, and Sirish L Shah. Recursive constrained state estimation using modified extended kalman filter. Computers & chemical engineering, 65: 0 9--17, 2014

    work page 2014

  34. [34]

    Kalmannet: Neural network aided kalman filtering for partially known dynamics

    Guy Revach, Nir Shlezinger, Xiaoyong Ni, Adria Lopez Escoriza, Ruud JG Van Sloun, and Yonina C Eldar. Kalmannet: Neural network aided kalman filtering for partially known dynamics. IEEE Transactions on Signal Processing, 70: 0 1532--1547, 2022

    work page 2022

  35. [35]

    Rtsnet: Learning to smooth in partially known state-space models

    Guy Revach, Xiaoyong Ni, Nir Shlezinger, Ruud JG Van Sloun, and Yonina C Eldar. Rtsnet: Learning to smooth in partially known state-space models. IEEE Transactions on Signal Processing, 71: 0 4441--4456, 2023

    work page 2023

  36. [36]

    Mobile phones as traffic probes: practices, prospects and issues

    Geoff Rose. Mobile phones as traffic probes: practices, prospects and issues. Transport Reviews, 26 0 (3): 0 275--291, 2006

    work page 2006

  37. [37]

    Tqam: Temporal attention for cycle-wise queue length estimation using high-resolution loop detector data

    Rahul Sengupta, Yashaswi Karnati, Anand Rangarajan, and Sanjay Ranka. Tqam: Temporal attention for cycle-wise queue length estimation using high-resolution loop detector data. In 2021 IEEE International Intelligent Transportation Systems Conference (ITSC), pages 3313--3320. IEEE, 2021

    work page 2021

  38. [38]

    A hybrid model for missing traffic flow data imputation based on clustering and attention mechanism optimizing lstm and adaboost

    Qiang Shang, Yingping Tang, and Longjiao Yin. A hybrid model for missing traffic flow data imputation based on clustering and attention mechanism optimizing lstm and adaboost. Scientific Reports, 14 0 (1): 0 26473, 2024

    work page 2024

  39. [39]

    Input--output and hybrid techniques for real-time prediction of delay and maximum queue length at signalized intersections

    Anuj Sharma, Darcy M Bullock, and James A Bonneson. Input--output and hybrid techniques for real-time prediction of delay and maximum queue length at signalized intersections. Transportation Research Record, 2035 0 (1): 0 69--80, 2007

    work page 2035

  40. [40]

    Kalman filtering with state constraints: a survey of linear and nonlinear algorithms

    Dan Simon. Kalman filtering with state constraints: a survey of linear and nonlinear algorithms. IET control theory & applications, 4 0 (8): 0 1303--1318, 2010

    work page 2010

  41. [41]

    Does traffic congestion slow the economy? Journal of Planning Literature, 26 0 (4): 0 391--404, 2011

    Matthias Sweet. Does traffic congestion slow the economy? Journal of Planning Literature, 26 0 (4): 0 391--404, 2011

    work page 2011

  42. [42]

    Robust queue length estimation for ramp metering in a connected vehicle environment

    Yu Tang, Kaan Ozbay, and Li Jin. Robust queue length estimation for ramp metering in a connected vehicle environment. In 2023 IEEE 26th International Conference on Intelligent Transportation Systems (ITSC), pages 4402--4407. IEEE, 2023

    work page 2023

  43. [43]

    Efficient video-based vehicle queue length estimation using computer vision and deep learning for an urban traffic scenario

    Muhammad Umair, Muhammad Umar Farooq, Rana Hammad Raza, Qian Chen, and Baher Abdulhai. Efficient video-based vehicle queue length estimation using computer vision and deep learning for an urban traffic scenario. Processes, 9 0 (10): 0 1786, 2021

    work page 2021

  44. [44]

    Validation of floating car data for transport policy analysis and transport models

    Han van der Loop, Marco Kouwenhoven, and Peter van Bekkum. Validation of floating car data for transport policy analysis and transport models. Technical report, KiM Netherlands Institute for Transport Policy Analysis, 2017. URL https://www.kimnet.nl/publicaties/papers/2017/10/04/validation-of-floating-car-data-for-transport-policy-analysis-and-transport-m...

    work page 2017

  45. [45]

    Estimating the vehicle accumulation: Data-fusion of loop-detector flow and floating car speed data

    PB van Erp, Victor L Knoop, and Serge P Hoogendoorn. Estimating the vehicle accumulation: Data-fusion of loop-detector flow and floating car speed data. In Proceedings of the 97th TRB Annual Meeting, Washington, DC, USA, pages 7--11, 2018

    work page 2018

  46. [46]

    Real-time estimation of vehicle-count within signalized links

    Georgios Vigos, Markos Papageorgiou, and Yibing Wang. Real-time estimation of vehicle-count within signalized links. Transportation Research Part C: Emerging Technologies, 16 0 (1): 0 18--35, 2008

    work page 2008

  47. [47]

    Kalman filtering method for real-time queue length estimation in a connected vehicle environment

    Yi Wang, Zhihong Yao, Yang Cheng, Yangsheng Jiang, and Bin Ran. Kalman filtering method for real-time queue length estimation in a connected vehicle environment. Transportation Research Record, 2675 0 (10): 0 578--589, 2021

    work page 2021

  48. [48]

    Shockwave-based queue estimation approach for undersaturated and oversaturated signalized intersections using multi-source detection data

    Zhongyu Wang, Qing Cai, Bing Wu, Lingyu Zheng, and Yinhai Wang. Shockwave-based queue estimation approach for undersaturated and oversaturated signalized intersections using multi-source detection data. Journal of Intelligent Transportation Systems, 21 0 (3): 0 167--178, 2017

    work page 2017

  49. [49]

    Traffic signal settings

    Fo Vo Webster. Traffic signal settings. Technical report, 1958

    work page 1958

  50. [50]

    Queue length estimation for a metered on-ramp using mesoscopic simulation

    Guangchuan Yang, Zong Tian, Daobin Wang, and Hao Xu. Queue length estimation for a metered on-ramp using mesoscopic simulation. Transportation Letters, 11 0 (10): 0 570--579, 2019

    work page 2019

  51. [51]

    A kalman filter-based queue length estimation method with low-penetration mobile sensor data at signalized intersections

    Juyuan Yin, Jian Sun, and Keshuang Tang. A kalman filter-based queue length estimation method with low-penetration mobile sensor data at signalized intersections. Transportation Research Record, 2672 0 (45): 0 253--264, 2018

    work page 2018

  52. [52]

    Various methods for queue length and traffic volume estimation using probe vehicle trajectories

    Yan Zhao, Jianfeng Zheng, Wai Wong, Xingmin Wang, Yuan Meng, and Henry X Liu. Various methods for queue length and traffic volume estimation using probe vehicle trajectories. Transportation Research Part C: Emerging Technologies, 107: 0 70--91, 2019

    work page 2019

  53. [53]

    Maximum likelihood estimation of probe vehicle penetration rates and queue length distributions from probe vehicle data

    Yan Zhao, Wai Wong, Jianfeng Zheng, and Henry X Liu. Maximum likelihood estimation of probe vehicle penetration rates and queue length distributions from probe vehicle data. IEEE Transactions on Intelligent Transportation Systems, 23 0 (7): 0 7628--7636, 2021

    work page 2021