State of the Art and Future Directions of Deep LearningDriven Collision Avoidance Mechanisms in Vehicular AdHoc Networks
Keywords:
Artificial Intelligence (AI), Autonomous Vehicles, Reinforcement Learning, Edge ComputingAbstract
Evolution of Intelligent Transportation Systems (ITS) has led to the transformation of modern vehicular safety strategies, particularly due to incorporation of Artificial Intelligence (AI) in collision avoidance technologies. These AI-enabled systems are designed in order to minimize the risk of accidents. To achieve this, real-time data is exchanged
through Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure (V2I) communication frameworks. This study explores the Technological Components of Autonomous Vehicular Systems. Core components of Intelligent Transportation Systems (ITS), Methodology to be adopted in AI-based collision avoidance Systems and Use case of AI-based collision avoidance System thereby emphasizing the role of connected vehicle communication in enhancing awareness and minimizing response delays. It also discusses the contribution of various AI methodologies, as in predictive modeling and reinforcement learning, which help systems learn from previous events and dynamically adjust to new traffic conditions. Few challenges identified are sensor inaccuracies, latency in data transmission, cybersecurity risks, and privacy concerns which must be addressed for ITS systems to achieve broader adoption. The future scope proposed in this area is incorporating emerging technologies such as 5G connectivity, for enabling ultra-fast communication speeds; edge computing, which facilitates for local and immediate data processing; and federated learning, which enables AI model training across distributed devices without the need for centralized data storage.
References
[1] Zhang, Y., Lin, Y., Wu, D.: Ai-driven collision prediction in connected vehicle environments using deep learning. IEEE Transactions on Intelligent Transportation Systems 24(1), 89–100 (2023) https://doi.org/10.1109/TITS.2023.3240123
[2] Ahmed, M., Kumar, S., Bhattacharya, P.: A review of v2v and v2i communication systems for intelligent transportation networks. Journal of Transportation Technologies 14(2), 125–137 (2024) https://doi.org/10.4236/jtts.2024.142008
[3] Liu, S., Cong, S., Yang, L.: Survey of research on autonomous driving testing with large models. Communications in Transportation Research 5, 100179 (2025) https://doi.org/10.1016/j.commtr.2025.100179
[4] Li, F., Zhao, W.: Real-time sensor fusion for autonomous vehicle navigation: A bayesian framework. Sensors 23(4), 1122 (2023) https://doi.org/10.3390/s23041122
[5] Selvaraj, D.C., Vitale, C., Panayiotou, T., Kolios, P., Chiasserini, C.F., Ellinas, G.: Edge-assisted ml-aided uncertainty-aware vehicle collision avoidance at urban intersections. IEEE Transactions on Intelligent Vehicles 11(5), 573–589 (2026) https://doi.org/10.1109/TIV.2023.3296190
[6] Chen, L., Han, Q.: Edge intelligence for real-time collision avoidance in autonomous driving. Future Internet 17(1), 1–18 (2025) https://doi.org/10.3390/fi17010002
[7] Elallid, B., Abouaomar, A., Benamar, N., Kobbane, A.: Vehicles control: Collision avoidance using federated deep reinforcement learning. In: 2023 IEEE Global Communications Conference (GLOBECOM), pp. 4369–4374 (2023). Corpus ID: 260681158
[8] Wang, X., Liu, T.: Deep reinforcement learning for collision-free path planning in urban scenarios. Neural Computing and Applications 35, 4567–4580 (2023) https://doi.org/10.1007/s00521-022-07536-2
[9] Singh, A., Chopra, N.: Vehicle safety and ai: An analysis of decision-making algorithms in high-speed scenarios. Autonomous Systems Journal 8(2), 74–85 (2024) https://doi.org/10.1016/asj.2024.05.004
[10] El-Sayed, H., Rana, M.: Design and implementation of v2x-enabled autonomous driving frameworks. IEEE Access 11, 104332–104345 (2023) https://doi.org/10.1109/ACCESS.2023.3287845
[11] Mahajan, K., Srivastava, R.: Ai-based cooperative driving systems using graph neural networks. International Journal of Vehicular Technology 2024, 1–11 (2024) https://doi.org/10.1155/2024/5378943
[12] Azfar, T., Huang, K., Tracy, A., Misiewicz, S., Liu, C., Ke, R.: Traffic co-simulation framework empowered by infrastructure camera sensing and reinforcement learning. Journal of Intelligent Transportation Systems, 1–19 (2025) https://doi.org/10.1080/15472450.2025.2559410