π Academic References
[1] Highway Capacity Manual (HCM), 6th Ed.
Transportation Research Board. Highway Capacity Manual: A Guide for Multimodal Mobility Analysis, 6th Edition. National Academies of Sciences, Washington, D.C., 2016.
LOS AβF classification, Capacity = 1800 veh/hr/lane (urban), V/C ratio thresholds, PHF calculations. PCE (level terrain): Car=1.0, Motorcycle=0.33, Bus=2.0, Truck=2.0, Heavy Truck=3.0.
[2] Highway Capacity Manual (HCM), 7th Ed.
National Academies of Sciences, Engineering, and Medicine. Highway Capacity Manual, 7th Edition: A Guide for Multimodal Mobility Analysis. Transportation Research Board, Washington, D.C., 2022.
Latest HCM edition with updated capacity, LOS criteria, and PCE methodology.
[3] Greenshields, B.D. et al. (1935)
Greenshields, B.D., Shannon, J.S., & Weidl, E.R. "The Density Method of Measuring Traffic Capacity." Highway Research Board Proceedings, Vol. 15, pp. 448β477, 1935.
Foundational traffic flow theory: speedβdensityβflow relationship (Greenshields model). ~1,500+ citations.
[4] Roess, Prassas & McShane (2004)
Roess, R.P., Prassas, E.S., & McShane, W.R. Traffic Engineering, 3rd Edition. Pearson Prentice Hall, 2004.
Standard traffic engineering textbook. Peak Hour Factor (PHF), saturation flow rate, capacity analysis methodology.
[5] Yuksel, E. et al. (2022)
Yuksel, E., Bertini, R.L., Li, X., Staes, B., & Ozkul, S. "Data-Driven Computation of State-Dependent Passenger Car Equivalency for Multiple Truck Lengths." Transportation Research Record, 2022.
PCE methodology using real-world data from PORTAL. State-dependent PCE values vs. HCM-6 static values.
[6] Hurtado-Beltran, A. et al. (2022)
Hurtado-Beltran, A. et al. "Impact of Capacity Definition on the HCM-6 Passenger Car Equivalent Values." J. of Transportation Engineering, Part A: Systems, Vol. 148, No. 7, 2022.
Comparison of HCM-6 PCE values under different capacity assumptions and data aggregation levels.
[7] Omar, Kar & Chunchu (2020)
Omar, B.S., Kar, P., & Chunchu, M. "Passenger Car Equivalent Estimation for Rural Highways: Methodological Review." Transportation Research Procedia, Vol. 48, pp. 801β816, 2020.
Comprehensive review of PCE estimation methods for heterogeneous traffic conditions.
π€ Computer Vision & Detection
[8] Liu et al. (2016) β SSD
Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.Y., & Berg, A.C. "SSD: Single Shot Multibox Detector." European Conf. on Computer Vision (ECCV), pp. 21β37, 2016.
Single Shot MultiBox Detector architecture. ~20,000+ citations. Foundation of COCO-SSD model used in this application.
[9] Lin et al. (2014) β COCO Dataset
Lin, T.Y., Maire, M., Belongie, S., Hays, J., Perona, P., Ramanan, D., DollΓ‘r, P., & Zitnick, C.L. "Microsoft COCO: Common Objects in Context." European Conf. on Computer Vision (ECCV), pp. 740β755, 2014.
COCO dataset for object detection, segmentation, and captioning. 80 object categories. ~18,000+ citations.
[10] Howard et al. (2017) β MobileNet
Howard, A.G. et al. "MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications." arXiv:1704.04861, 2017.
MobileNet architecture for efficient on-device inference. Used as backbone in COCO-SSD (lite_mobilenet_v2).
[11] TensorFlow.js & COCO-SSD
Smilkov, D. et al. "TensorFlow.js: Machine Learning for the Web and Beyond." arXiv:1712.01815, 2017. β TensorFlow.js COCO-SSD model (v2.2.3): GitHub
Browser-based ML inference engine. COCO-SSD provides 80-class real-time object detection via WebGL.
π Multi-Object Tracking
[12] Bewley et al. (2016) β SORT
Bewley, A., Ge, Z., Ott, L., Ramos, F., & Upcroft, B. "Simple Online and Realtime Tracking." IEEE International Conf. on Image Processing (ICIP), pp. 3464β3468, 2016.
Simple Online and Realtime Tracking using Kalman filter + Hungarian algorithm. ~3,000+ citations.
[13] Zhang et al. (2022) β ByteTrack
Zhang, Y., Sun, P., Jiang, Y., Yu, D., Yuan, Z., Luo, P., Liu, W., & Wang, X. "ByteTrack: Multi-Object Tracking by Associating Every Detection Box." European Conf. on Computer Vision (ECCV), 2022.
Associates every detection box, not just high-score ones. 80.3 MOTA, 77.3 IDF1 on MOT17. ~2,057 citations.
[14] Cao et al. (2023) β OC-SORT
Cao, X., Wang, X., Xiao, J., & Yuan, Y. "Observation-Centric SORT: Rethinking SORT for Robust Multi-Object Tracking." CVPR 2023.
Observation-centric re-update for robust tracking in crowded/non-linear motion scenes. 700+ FPS on CPU.
[15] Aharon, Orfaig & Bobrovsky (2022) β BoT-SORT
Aharon, N., Orfaig, R., & Bobrovsky, B.Z. "BoT-SORT: Robust Associations Multi-Pedestrian Tracking." arXiv:2206.14651, 2022.
Combines motion + appearance + camera-motion compensation. 80.5 MOTA on MOT17. ~761 citations.
[16] Du et al. (2022) β StrongSORT
Du, Y., Song, Z., Yang, B., & Zhao, S. "StrongSORT: Make DeepSORT Great Again." IEEE Transactions on Multimedia, 2022.
Enhanced appearance model + Kalman filter improvements. High accuracy tracking with appearance re-identification.
π§ Open Source Libraries
Leaflet.js + OpenStreetMap
Interactive map with Nominatim geocoding β leafletjs.com
Chart.js
Doughnut & line charts for vehicle distribution and flow trends β chartjs.org
Google Fonts β Cairo & Roboto
Bilingual typography for EN/AR support.
Methodology
AI detection via COCO-SSD (TensorFlow.js) + multi-object tracking (ByteTrack/SORT/OC-SORT/StrongSORT/BoT-SORT). Vehicle classification by neural network confidence + frame-area percentage. 15-minute interval aggregation per HCM guidelines.
