Road surface anomalies are a serious issue for safe and effective traffic flow; they are becoming more common due to factors like poor materials for construction, heavy traffic, climate change, and high traffic numbers. The vehicular ad hoc network (VANET), a developing research area, provides the network's vehicles with essential data. Finding and fixing these defects is essential to the mechanical safety of drivers, passengers, and automobiles. Based on the integration of artificial intelligence into automobiles, autonomous cars may find it useful to use a combination of sensors and DNN techniques to sense their environment and identify tracks and obstacles for safe and easy navigation. One of the primary problems for autonomous vehicles is avoiding dangerous situations on the road caused by serious road defects. The Intelligent Transportation System (ITS) developed the concept of a vehicle network known as a "vehicular ad hoc network (VANET)" for assuring security and safety in a traffic flow in order to handle accident issues and relay emergency information. Using Edge AI and VANET, a novel method has been put proposed allowing autonomous automobiles to recognise defects on the road automatically and notify oncoming traffic of them. The methods "Residual Convolution Neural Network (ResNet-18) and Visual Geometry Group (VGG-16)" are used to automatically detect and classify the roads with irregularities, like a bump, crack or pothole, and plain roads with no anomaly using data from several online sources. The results show that the models that were utilized performed better at identifying road abnormalities than other methods.
Automated Road anomaly detector in VANET by using Deep Learning
2023-12-13
673230 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Edge AI-Based Automated Detection and Classification of Road Anomalies in VANET Using Deep Learning
DOAJ | 2021
|Smart Traffic Monitoring and Alert System Using VANET and Deep Learning
Springer Verlag | 2021
|