Real Time Traffic Condition Based Smart Signal System

Abstract

Intheeraofcrowddeliveryvehiclerouting and traffic management in smart cities, a complex challenge appears indistinctly, affecting both developed and developing nationsworldwide.Thischallengingprobleminvolvesoptimizing multi-depot routes while addressing various hurdles: minimizing travel time, distance, fuel consumption, and carbon emissions,all while navigating dynamic traffic congestion across diverse pathways.Existingapproachesoftenfocusonisolatedaspectslike shortestpaths,carbonemissions,ortrafficprediction,leavingthe comprehensive multi-depot traffic management problem unad-dressed. In response, this research work proposes an Intelligent Multi-DepotVehicleRoutingandManagement(IM-VRM)model which provides a comprehensive and holistic solution. It employs aGraphNeuralNetwork(GNN)learning-basedroutingwith agreedyoptimizationtoestablishinitialoptimalpathways for multi-depot journeys. Subsequently, the IM-VRM model integrates traffic congestion prediction with green parameter computation,engagingtheDijkstraalgorithmtoselectthe mostadmissibleroutes. This consecutive steps-based travel route guidance process optimizes routing for heterogeneous vehicles, including both heavy-duty and light-duty types. It accounts for load-dependentfuelconsumption,velocity,andcarbonemissions. By doing so, it simplifies the complexities of multi-depot traffic routing and management. The proposed model has been rigor-ously evaluated using a real-world multi-depot traffic dataset, demonstrating its practical viability. Notably, IM-VRM model achieves a remarkable improvement in fuel savings, reduced carbonemissions,andshortertraveltimeoutperformingprevious state-of-the-art methods in both efficiency and precision.

Introduction

The study addresses the growing challenges of urban logistics in smart cities, where increasing delivery vehicles contribute to traffic congestion, fuel consumption, air pollution, and CO? emissions. Existing solutions such as crowd delivery, car-pooling, and conventional routing methods often struggle with real-time adaptability, scalability, and efficient traffic management.

To overcome these issues, the authors propose an Intelligent Multi-Depot Vehicle Routing and Management (IM-VRM) model. The framework optimizes delivery operations by minimizing travel time, travel distance, fuel consumption, and carbon emissions while dynamically adapting to changing traffic conditions. The model integrates Graph Neural Networks (GNNs) with a Greedy Search optimization algorithm to determine efficient routes across multiple depots. It also incorporates real-time traffic congestion prediction and green transportation parameters to support sustainable urban logistics.

The system, managed through an Intelligent Traffic Organizer and Management Engine (ITOME), utilizes live traffic data, historical traffic records, and city route information to generate optimal routes for heterogeneous delivery vehicles. Key factors considered include fuel consumption, carbon emissions, travel distance, travel time windows, vehicle capacities, and traffic congestion.

The proposed approach formulates vehicle routing as a multi-objective optimization problem and employs mathematical models to estimate fuel consumption, emissions, travel time, and congestion levels. GNNs capture spatial relationships among locations, while the Greedy Search algorithm selects routes based on optimized node embeddings. Additionally, a Traffic Congestion Forecasting Model continuously predicts traffic conditions using historical and real-time data.

Experimental evaluations using real-world benchmark datasets demonstrate that the IM-VRM model outperforms conventional and state-of-the-art routing methods in terms of delivery efficiency, congestion handling, fuel savings, and emission reduction. Overall, the framework provides a scalable, practical, and environmentally sustainable solution for multi-depot delivery management in smart city environments.

Conclusion

This paper proposes an innovative and comprehensive approach to tackle the challenging problem of crowd delivery vehicle routing and traffic management problem. By combin-ing GNN with greedy optimization for initial route selection and incorporating traffic congestion prediction with green parametercomputationtooptimi zerouteguidanceusing theDijkstraalgorithm, weprovidean efficientand effec-tive solution. This approach simplifies the multi-depot traffic route guidance and management problem compared to pre-viousstate-of-the-artmethods.Futureresearchwillfocus on real-time weather conditions with real-time traffic data integration and user-focused priorities to enhance practical implementation. Additionally, we will explore the scalability of the proposed approach for larger and more complex traffic networks while providing valuable insights into addressing evolving challenges in crowd delivery vehicle routing and traffic management.

References

[1] Q.Wang,H.Feng,H.Feng,Y.Yu,J.Li,andE.Ning,“Theimpactsof road traffic on urban air quality in jinan based GWR and remotesensing,” Sci. Rep., vol. 11, no. 1, p. 15512, Jul. 2021. [2] J. K. Stolaroff, C. Samaras, E. R. O’Neill, A. Lubers, A. S. Mitchell,and D. Ceperley, “Energy use and life cycle greenhouse gas emissionsofdronesforcommercialpackagedelivery,”Nat.Commun.,vol.9,no.1, [3] p.409,Dec.2018. [4] L. Wu, Y. Ci, Y. Wang, and P. Chen, “Fuel consumption at theoversaturated signalized intersection considering queue effects: A casestudy in harbin, China,” Energy, vol. 192, Feb. 2020, Art. no. 116654. [5] G. J. L. Micheli and F. Mantella, “Modelling an environmentally-extended inventory routing problem with demand uncertainty and aheterogeneous fleet under carbon control policies,” Int. J. Prod. Econ.,vol. 204, pp. 316–327, Oct. 2018. [6] H. Cheng et al., “Truck platooning reshapes greenhouse gas emissionsof the integrated vehicle-road infrastructure system,” Nature Commun.,vol. 14, no. 1, p. 4495, Aug. 2023. [7] C. Nangini et al., “A global dataset of CO2 emissionsand ancillary datarelated to emissions for 343 cities,” Sci. Data, vol. 6, no. 1, pp. 1–29,Jan.2019. [8] M. Geisslinger, F. Poszler, and M. Lienkamp, “An ethical trajectoryplanning algorithm for autonomousvehicles,” Nat.Mach. Intell.,vol.5, [9] pp.137–144,Feb.2023. [10] M. Maimaiti, X. Zhao, M. Jia, Y. Ru, and S. Zhu, “How we eatdetermines what we become: Opportunities and challenges brought byfood delivery industry in a changing world in China,” Eur. J. Clin.Nutrition, vol. 72, no. 9, pp. 1282–1286, Sep. 2018. [11] A. C. de Araujo and A. Etemad, “End-to-end prediction of parceldelivery time with deep learning for smart-city applications,” IEEEInternet Things J., vol. 8, no. 23, pp. 17043–17056, Dec. 2021. [12] N.Serok,S.Havlin,andE.BlumenfeldLieberthal,“Identification,cost evaluation, and prioritization of urban traffic congestions and theirorigin,” Sci. Rep., vol. 12, no. 1, p. 13026, Jul. 2022. [13] A. M. Avila and I. Mezic´, “Data-driven analysis and forecasting ofhighway traffic dynamics,” Nature Commun., vol. 11, no. 1, p. 2090,Apr.2020. [14] A.A.JolfaeiandM.Alinaghian,“Multi-depotvehicleroutingprob-lem with roaming delivery locations considering hard time windows:Solved by a hybrid ELS-LNS algorithm,” Expert Syst. Appl., vol. 255,Dec. 2024, Art. no. 124608. [15] J. Du et al., “CrowDNet: Enabling a crowdsourced object deliverynetwork based on modern portfolio theory,” IEEE Internet Things J.,vol. 6, no. 5, pp. 9030–9041, Oct. 2019. [16] M.Chen,X.Yu,andY.Liu,“PCNN:Deepconvolutionalnet-works for short-term traffic congestion prediction,” IEEE Trans. Intell.Transp. Syst., vol. 19, no. 11, pp. 3550–3559, Nov. 2018. [17] G. N. Kouziokas, “Deep bidirectional and unidirectional LSTM neuralnetworks in traffic flow forecasting from environmental factors,” inAdvancesinMobility-as-a-ServiceSystems,E.G.Nathanail,G.Adamos,andI.Karakikes,Eds.,Cham,Switzerland:Springer,2021,pp.171–180. [18] W. Joe and H. C. Lau, “Deep reinforcement learning approach to solvedynamic vehicle routing problem with stochastic customers,” in Proc.Int. Conf. Automated Planning Scheduling, vol. 30, 2020, pp. 394–402. [19] B. Wang and R. Su, “A distributed platoon control framework forconnectedautomated vehiclesinanurban trafficnetwork,” IEEETrans.Control Netw. Syst., vol. 9, no. 4, pp. 1717–1730, Dec. 2022. [20] X.-H. Liu, M.-Y. Shan, R.-L. Zhang, and L.-H. Zhang, “Green vehicleroutingoptimizationbasedoncarbonemissionandmultiobjectivehybridquantum immune algorithm,” Math. Problems Eng., vol. 2018, pp. 1–9,Jun.2018. [21] L.Abbatecola,M.P.Fanti,A.M.Mangini,andW.Ukovich,“Adecisionsupportapproachforpostaldeliveryandwastecollectionservices,”IEEETrans. Autom. Sci. Eng., vol. 13, no. 4, pp. 1458–1470, Oct. 2016. [22] G.Perboli,M.Rosano,andQ.Wei,“Asimulation-optimizationapproachfor the management of the on-demand parcel delivery in sharing econ-omy,”IEEETrans.Intell.Transp.Syst.,vol.23,no.8,pp.10570–10582,Aug.2022. [23] C.Chenetal.,“Crowddeliver:Planningcity-widepackagedeliverypathsleveragingthecrowdoftaxis,”IEEETrans.Intell.Transp.Syst.,vol.18,no. 6, pp. 1478–1496, Jun. 2017. [24] N. Kafle, B. Zou, and J. Lin, “Design and modeling of a crowdsource-enabled system for urban parcel relay and delivery,” Transp. Res. B,Methodol., vol. 99, pp. 62–82, May 2017. [25] H. Hong, X. Li, D. He, Y. Zhang, and M. Wang, “Crowdsourcingincentives for multi-hop urban parcel delivery network,” IEEE Access,vol. 7, pp. 26268–26277, 2019. [26] D. Saxena and A. K. Singh, “A self-healing and fault-tolerant cloud-based digital twin processing management model,” IEEE Trans. Ind.Informat., early access, Feb. 26, 2025, doi: 10.1109/TII.2025.3540498. [27] B. Fleischmann, S. Gnutzmann, and E. Sandvoß, “Dynamic vehiclerouting based on onlinetrafficinformation,” Transp.Sci.,vol.38,no.4, [28] pp.420–433,Nov. 2004. [29] F.An,M.Barth,J.Norbeck,andM.Ross,“Developmentofcomprehen-sivemodalemissionsmodel:Operatingunderhot-stabilizedconditions,”Transp.Res. Record, J. Transp.Res.Board,vol.1587,no.1,pp. 52–62,Jan.1997. [30] H. Li,J.Yuan,T.Lv,andX.Chang,“Thetwo-echelontime-constrainedvehicle routing problem in linehaul-delivery systems considering car-bondioxideemissions,”Transp.Res.D,Transp.Environ.,vol.49,pp.231–245,Dec.2016. [31] Y. Zheng. (Aug. 2011). T-drive Trajectory Data Sample. [Online].Available: https://www.microsoft.com/en-us/research/publication/t-drive-trajectory-data-sample/ [32] M. Hou, F. Xia, H. Gao, X.Chen, and H. Chen,“Urban region profilingwith spatio-temporal graph neural networks,” IEEE Trans. Computat.Social Syst., vol. 9, no. 6, pp. 1736–1747, Dec. 2022. .

Copyright

Copyright © 2026 Gavini Venkateswari, Yetra Pujitha Gayathri, Meka Naga Phani, Nandam Hurmi Nandalal. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.