REAL-TIME OPTIMIZATION OF TRAFFIC SIGNALING TIME USING CNN
Keywords:
Convolution neural network, machine learning, surveillance video, traffic signal time, traffic volumeAbstract
Transportation is one of the major elements of developing urban areas. In urban areas, traffic congestion problem occurs due to an increase in traffic volume. Traffic volume is dynamically changing in intersections and not consistent in all legs for all-day. But traffic signals are given by fixed time or manually by traffic signal operators. For the dynamic change in traffic volume, fixed time traffic signals lead to increasing traffic congestion problems in intersections. This problem can be reduced by designing an efficient traffic signaling time in intersections. In this proposed work, a model is developed by using conventional algorithms for designing the traffic signaling time and distributing the green signaling time based on real-time traffic volume present in the legs of the intersection. The real-time traffic volume is generated from the surveillance cameras by applying machine learning, and Convolution Neural Network (CNN) techniques. CNN algorithms are classified the vehicles with an accuracy of 62.92 percent in mixed traffic condition when compared to the results of Traffic Data Extractor (TDE) software, and the signaling time is optimized for each leg of the intersection based on the traffic volume present in the legs.
References
Akoum, A.H. (2017). Automatic traffic using image processing. J. Software Eng. Appli., 10(9):765-776. doi: 10.4236/jsea. 2017.109042.
Lee, J. and Park, G.L. (2018). Big data processing in charging infrastructures for smart transportation systems. ARPN J. Eng. Appl. Sci., 13:1,770-1,774.
Ma, X., Dai, Z., He, Z., Ma, J., Wang, Y., and Wang, Y. (2017). Learning traffic as images: A deep convolutional neural network for large-scale transportation network speed prediction. Sensors (Switzerland), 17(4):818. https://doi. org/10.3390/s17040818.
Redmon, J. and Farhadi, A. (2018). YOLOv3: An Incremental Improvement. CoRR, 1804.02767. http://arxiv.org/abs/ 1804.02767.
Sadoun, B. (2003). An efficient simulation methodology for the design of traffic lights at intersections in urban areas. Simulation, 79:243–251. https://doi.org/10.1177/003754 9703038878.
Senthilkumar, K., Ellappan, V., and Arun, A.R. (2017). Traffic analysis and control using image processing. IOP Conference Series: Materials Science and Engineering, 263p. https://doi.org/10.1088/1757-899X/263/4/042047.
Xu, M., An, K., Vu, L.H., Ye, Z., Feng, J., and Chen, E. (2019). Optimizing multi-agent based urban traffic signal control system. J. Intell. Trans. Sys.: Technol. Plann., Operat., 23(4):357-369.
Yao, Z., Jiang, Y., Zhao, B., Luo, X., and Peng, B. (2020). A dynamic optimization method for adaptive signal control in a connected vehicle environment. J. Intell. Trans. Sys.: Technol. Plann. Operat., 24(2):184-200. https://doi.org/10.1080/15472450.2019.1643723
Zheng, F., Li, J., van Zuylen, H., Liu, X., and Yang, H. (2018). Urban travel time reliability at different traffic conditions. J. Intell. Trans. Sys.: Technol. Plann. Operat., 22(2):106-120. https://doi.org/10.1080/15472450.2017.1412829.
Zheng, J., Ma, X., Wu, Y.J., and Wang, Y. (2013). Measuring signalized intersection performance in real-time with traffic sensors. J J. Intell. Trans. Sys.: Technol. Plann. Operat., 17(4):304-316, https://doi.org/10.1080/15472 450.2013.771105.
