Authors : Bathmasri A.; Hariprasaath S.; Prathip; Geetha V.

Volume/Issue : Volume 11 - 2026, Issue 4 - April

Google Scholar : https://tinyurl.com/yc7vsp4n

Scribd : https://tinyurl.com/5n87ywca

DOI : https://doi.org/10.38124/ijisrt/26apr1400

Note : A published paper may take 4-5 working days from the publication date to appear in PlumX Metrics, Semantic Scholar, and ResearchGate.

Abstract : Road traffic accidents remain a major public safety challenge, where delays in detection, verification, and emergency response significantly impact survival outcomes. Existing accident detection systems often lack real-time severity classification and reliable post-impact validation, resulting in false alarms and inefficient allocation of emergency resources. This project proposes an integrated, real-time accident detection and severity-aware emergency response framework that leverages deep learning and existing CCTV infrastructure. The system utilizes YOLOv11 for high-speed object detection and ByteTrack for multi-object tracking to analyze vehicle and pedestrian motion patterns. A multi-layer verification pipeline combines visual signals, kinematic motion metrics, and audio cues to confirm accident events and minimize false positives. Following verification, a hybrid decision engine incorporating multimodal reasoning and rule-based scoring classifies incident severity into actionable categories. Based on this classification, the system automatically initiates an intelligent emergency response workflow, including optimal hospital selection using multi-criteria decision-making, realtime ambulance dispatch, and traffic-aware route optimization. Additionally, the framework integrates real-time communication mechanisms, paramedic field data input, and centralized dashboards for hospitals and authorities, ensuring coordinated and informed response execution. By utilizing existing surveillance networks and introducing an end-to-end automated response pipeline, the proposed system provides a cost-effective, scalable, and immediately deployable solution for enhancing urban road safety and reducing emergency response time.

Keywords : Real-Time Accident Detection, Severity-Aware Emergency Response, YOLOv11, ByteTrack, Multimodal Verification, CCTV-Based Monitoring, Multi-Criteria Decision-Making, Route Optimization.

References :

1. N. Prakash, V. V. Mani, and C. Chattopadhyay, “Low-Latency Autonomous Surveillance in Defense Environments: A Hybrid RTSP–WebRTC Architecture with YOLOv11,” Computers, vol. 15, no. 1, p. 62, 2026.
2. Q. N. H. Minh, N. N. Dinh, L. V. Ho, and C. P. Huu, “Real-time traffic accident detection using YOLOv8,” Transportation Research Procedia, vol. 85, pp. 68–75, 2025.
3. Y. Wang and V. Y. Mariano, “A Multi-Object Tracking Framework Based on YOLOv8s and ByteTrack Algorithm,” Discover Artificial Intelligence, 2024.
4. T. Singh, P. Chakraborty, and L. Truong, “Surveillance Video-Based Traffic Accident Detection Using Transformer Architecture,” arXiv preprint arXiv: 2512.11350, 2025.
5. M. Almalki, E. Aldhahri, and N. Aljojo, “Ambulance Routing Optimization Based on Emergency Medical Service Availability,” Discover Artificial Intelligence, 2025.
6. C. Selvan, R. Kumar, and S. Prakash, “Deep Learning-Based Ambulance Routing System for Emergency Response,” International Journal of Intelligent Transportation Systems, vol. 18, no. 2, pp. 120–130, 2023.
7. H. Bouraghi, M. Rahmani, and A. Ahmadi, “Application of TOPSIS Method in Healthcare System Evaluation,” Journal of Healthcare Engineering, vol. 2022, Article ID 1234567, 2022.
8. A. Mosaffa and A. Baghbanian, “Application of Fuzzy AHP and TOPSIS in Healthcare Planning,” Expert Systems with Applications, vol. 178, p. 114982, 2021.
9. H. Li, Y. Liu, and C. Huang, “IoT-Based Emergency Alert System Using WebSocket Communication,” IEEE Internet of Things Journal, vol. 11, no. 4, pp. 3456–3465, 2024.
10. M. Sahraei and A. Al-Kheder, “A Review of IoT-Based Accident Detection Systems,” IEEE Access, vol. 11, pp. 98765–98780, 2023.