Enhance Intelligent Method for Accident Detection and Best Road Choosing to Prevent Congestion in VANET Vehicle

This research addresses the critical issue of accident detection and congestion prevention in VANET (Dedicated Vehicle Network). VANETs play a pivotal role in enhancing road safety, traffic efficiency, and providing value-added services to both drivers and passengers. To tackle this challenge, we employ a combination of clustering and machine learning techniques. Specifically, we utilize the K-means algorithm to divide the network's coverage area into clusters and gather vital information about the roads within these clusters. Subsequently, we introduce a hybrid approach, blending Support Vector Machines (SVM) with Multi-Layer Perceptron (MLP) algorithms, to assess the severity of accidents and predict congestion levels caused by these accidents. Our experimental results demonstrate the effectiveness of our approach. Notably, the K-means algorithm, used to identify road clusters, reveals superior accuracy when implemented with the MDORA protocol compared to the DSDV protocol. Furthermore, the SVM algorithm exhibits an impressive accuracy rate of 99% before hybridization. Following the hybridization process, its accuracy surges to 99.66%. These findings lay the groundwork for future endeavors aimed at optimizing route selection through the integration of advanced technologies like the Internet of Things (IoT).