Moto VANET Using Machine Learning Models

This paper presents Moto VANET, a collision warning system specifically designed for two-wheelers, with the sole aim to reduce road accidents and fatalities caused by human error and poor visibility conditions. This system integrates advanced object detection model algorithms YOLOv5, Convolution Neural Networks (CNN), and Kalman filters to provide real-time obstacle detection and lane change monitoring. Moto VANET utilizes a multi-sensor setup comprising ultrasonic sensors and a camera module to achieve robust collision prediction, ensuring effective detection of vehicles, pedestrians and lane boundaries. The Kalman filter is implemented to refine tracking by predicting object motion and reduce noise, which improves system reliability. Ensures a rapid response through a three-tiered alert system based on the surroundings. The system architecture is optimized for real-time processing with low latency and low computation using segmentation of vehicles, Pedestrians, etc … This cost-effective solution is designed for retrofitting into existing vehicles, contributing to safer mobility.