A smart roadside unit with integrated vision and 24 GHz IOT radar for real-time behavior prediction and V2X communication
DOI:
https://doi.org/10.54939/1859-1043.j.mst.IITE.2025.35-44Keywords:
Smart RSU; V2X communication; 24 GHz radar; AI camera; Sensor fusion; Traffic behavior prediction; Intelligent transportation systems; Smart city.Abstract
This paper proposes a Smart Roadside Unit (RSU) architecture designed for intelligent transportation systems in high-density urban environments with heterogeneous traffic. The system integrates V2X communication, 24 GHz radar, and AI-based vision to enable autonomous, context-aware decision-making for traffic regulation and incident response. Leveraging a machine learning model trained on localized traffic behavior, the RSU not only detects violations but also predicts potentially hazardous actions specific to each area. Real-time sensor fusion between radar and camera enhances the accuracy of object detection, tracking, and behavioral analysis under various environmental conditions. The proposed system improves traffic flow management, congestion mitigation, and proactive control capabilities. The solution is well-suited for deployment in smart cities, particularly in the context of increasing urbanization and the growing demand for data-driven mobility infrastructure.
References
[1]. Hesham A. Rakha, Kyoungho Ahn, Jianhe Du, and Mohamed Farag. "Quantifying the Impact of Cellular Vehicle-to-Everything (C-V2X) on Transportation System Efficiency, Energy and Environment", IEEE Transactions on Intelligent Transportation Systems, (2020).
[2]. Tom Rebbeck, Janette Stewart, Hugues-Antoine, and Andrew. "Final report for 5GAA on cellular V2X socio-economic benefits", Ref. 2011027-492, (2017).
[3]. Lejun Jiang, Tamás G. Molnár, and Gábor Orosz. "On the deployment of V2X roadside units for traffic prediction", Transportation Research Part C, 129, 103238, (2021).
[4]. Ali R. Abdellah, Ahmed Abdelmoaty, Abdelhamied A. Ateya, Ammar Muthanna, and Andrey Koucheryavy. "Accurate V2X traffic prediction with deep learning architectures", IEEE Access, PMCID: PMC11962783, (2018).
A. Ghosal and M. Conti. "Security issues and challenges in V2X: A survey", Computer Networks, 169, 107093, (2020). doi:10.1016/j.comnet.2019.107093
[5]. Y. Zhou, S. Liu, M. Xiao, S. Deng, and X. Wang. "An efficient V2I authentication scheme for VANETs", Mobile Information Systems, 2018, 1–11, (2018). doi:10.1155/2018/4070283
[6]. Shaoheng Fang, Rui Ye, Wenhao Wang, Zuhong Liu, Yuxiao Wang, Yafei Wang, Siheng Chen, and Yanfeng Wang. "Federated Learning for Scene Flow Estimation on Roadside Units", arXiv preprint, arXiv:2401.12862, (2024).
[7]. Widhi Yahya, Ying-Dar Lin, Faysal Marzuk, and Piotr Chołda. "Offloading in V2X with road side units: Deep reinforcement learning", Computer Communications, 51, 100862, (2025).
[8]. NVIDIA Corporation. "Jetson Orin NX Module Datasheet", (2023). [Online]. Available: https://developer.nvidia.com/embedded/jetson-orin
[9]. S. Liu, Y. Zhao, and H. Wang. "Edge AI computing for intelligent transportation systems: A review and case study", IEEE Transactions on Intelligent Transportation Systems, 24, 5, 4835–4848, (2023). doi:10.1109/TITS.2022.3149783
A. Ding, L. Xie, and H. Zhang. "Obstacle detection method based on RSU and vehicle camera fusion using YOLOv5", Sensors, 23, 10, 4920, (2023). doi:10.3390/s23104920
[10]. S. Gupta et al. "DeepStream: A system for real-time AI-based video analytics", NVIDIA Developer Blog, (2022). [Online]. Available: https://developer.nvidia.com/blog/deepstream-real-time-video-analytics/
[11]. R. Sundar and V. Ramachandran. "Edge AI acceleration in intelligent transportation: A case study using NVIDIA Jetson platform", Proc. IEEE Int. Conf. Intelligent Transportation Engineering (ICITE), Singapore, pp. 125–130, (2023). doi:10.1109/ICITE56765.2023.10234029
[12]. Infineon Technologies AG. "BGT24LTR22 XENSIV™ 24 GHz Radar Sensor Datasheet", Rev. 1.0, (2020). [Online]. Available: Infineon BGT24LTR22 DataSheet
[13]. Rohling. "Radar CFAR thresholding in clutter and multiple target situations", IEEE Transactions on Aerospace and Electronic Systems, 19, 4, 608–621, (1983). doi:10.1109/TAES.1983.309393
[14]. Scharf. "Statistical Signal Processing: Detection, Estimation, and Time Series Analysis", Addison-Wesley, Boston, MA, ch. 6, (1991).
[15]. S. Wang, J. Zhang, and L. Sun. "A constant false alarm rate algorithm based on the difference of box filter for radar pulse detection", IET Radar, Sonar & Navigation, 18, 2, 150–157, (2024). doi:10.1049/rsn2.12573
[16]. R. M. Elnadree and A. El-Sisi. "Flowchart of SVM algorithm", in Performance Investigation of Features Extraction and Classification Approaches for Sentiment Analysis Systems, (2021). [Online]. Available: researchgate.net/figure/Flowchart-of-SVM-algorithm_fig3_354303930
[17]. D. Mukherjee, S. Chakraborty, and S. Ghosh. "Flowchart of support vector machine (SVM)", in Power System State Forecasting Using Machine Learning Techniques, (2021). [Online]. Available: researchgate.net/figure/Flowchart-of-support-vector-machine-SVM_fig4_353341688
[18]. J. M. Moguerza and A. Muñoz. "Support Vector Machines with Applications", arXiv preprint, arXiv:math/0612817, (2006).
[19]. B. Benjdira et al. "Car detection using unmanned aerial vehicles: Comparison between Faster R-CNN and YOLOv3", arXiv preprint, arXiv:1812.10968, (2018).
[20]. Kumaran, D. P. Dogra, and P. P. Roy. "Anomaly detection in road traffic using visual surveillance: A survey", arXiv preprint, arXiv:1901.08292, (2019).
