Vehicle Detection and Classification for Traffic Management and Autonomous Systems Using YOLOv8

The AI-based Intelligent Traffic Management System pro-vides an innovative way to more comprehensively solve the increasing problem of urban traffic congestion through the integration of computer vision with dynamically controlled signals. By employing a YOLOv8-based detection model, the proposed traffic management system will accurately detect both vehicles and emergency vehicles and provide real-time traffic density estimates and adaptive decision-making. In contrast to traditional, fixed-time traffic systems, the proposed traffic management system will adaptively change the amount of time signals are turned on or off based upon actual current traffic patterns to improve the efficiency of the use of the roadways, as well as reduce the wait time at signals. Additionally, the proposed implementation of an emergency vehicle clearance system in the Intelligent Traffic Management System allows immediate clearance of ambulances from traffic signals and locations; thus, it will improve emergency response time in critical life-threatening situations. Furthermore, the development of a web-based traffic monitoring, visualization, and analytic dashboard for the Intelligent Traffic Management System allows for continuous, real-time traffic data monitoring and analysis and for practical real-world implementation of the Intelligent Traffic Management System for current smart city development. Overall, through the combination of intelli-gent detection, automated control, and real-time feedback, the Intelligent Traffic Management System represents a scalable, efficient, and reliable method to improve the management and control of urban traffic.

[1]   Q. Wang, H. Feng, H. Feng, Y. Yu, J. Li, and E. Ning, “The impacts of road traffic on urban air quality in Jinan based GWR and remote sensing,” Scientific Reports, vol. 11, no. 1, 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 emissions of drones for commercial package delivery,” Nature Communications, vol. 9, no. 1, 2018. [3]   L. Wu, Y. Ci, Y. Wang, and P. Chen, “Fuel consumption at oversaturated signalized intersections considering queue effects,” Energy, vol. 192, 2020. [4]     G. J. L. Micheli and F. Mantella, “Environmentally-extended inventory routing problem under carbon control policies,” International Journal of Production Economics, vol. 204, 2018. [5]   H. Cheng et al., “Truck platooning reshapes greenhouse gas emissions,” Nature Communications, vol. 14, 2023. [6]   C. Nangini et al., “A global dataset of CO emissions for 343 cities,” Scientific Data, vol. 6, 2019. [7]    M. Geisslinger, F. Poszler, and M. Lienkamp, “Ethical trajectory planning algorithm for autonomous vehicles,” Na-ture Machine Intelligence, vol. 5, 2023. [8]  M. Chen, X. Yu, and Y. Liu, “PCNN: Deep convolutional networks for traffic congestion prediction,” IEEE Transactions on Intelligent Transportation Systems, vol. 19, no. 11, 2018. [9]   G. N. Kouziokas, “LSTM neural networks in traffic flow forecasting,” in Advances in Mobility-as-a-Service Systems, Springer, 2021. [10]   W. Joe and H. C. Lau, “Deep reinforcement learning for dynamic vehicle routing,” ICAPS, 2020.