Probability Model and Warning System for Improper Driving Behavior in Vehicle Ad Hoc Networks

Abstract

Intelligent transportation systems (ITS) have improved road safety and traffic management; however, improper driving behavior remains a main cause of accidents. Real-time detection and warning systems are crucial to proactively address and moderate these hazards. This research recommends a novel probability approach as well as a warning system for detecting improper driving behavior in Vehicle Ad Hoc Networks (VANETs). The system incorporates real-time data from onboard sensors, vehicle-to-vehicle (V2V), and vehicle-to-infrastructure (V2I) communication to observe driving behavior, like erratic lane changes, sudden acceleration, and harsh braking. It utilizes the Kalman Filtering (KF) technique and Interquartile Range (IQR) to remove noise and irrelevant data from the sensors and communication channels. This system establishes an Intelligent White Shark Optimized Support Vector Machine (IWSO-SVM) approach to detect improper driving behavior in VANETs. The IWSO-SVM method probabilistically assesses the probability of unsafe actions. When the probability of improper driver behavior exceeds a defined threshold, the system triggers an immediate warning to the driver and other nearby vehicles. The system in a real-world VANET integrates feedback loops from different sources to continuously improve the system's performance. The efficiency of the model is established through simulations, showcasing its ability to improve traffic flow, reduce accidents, and advance safer driving practices within the VANET environment. This classification offers a promising solution for real-time traffic safety monitoring, leveraging the ability of VANETs and advanced probability models to mitigate the risks of improper driving behavior.