Prediction and Analysis of Machine Learning Models for Efficient Routing Protocol in VANET using Feature Information

Abstract Vehicular Ad-hoc Networks (VANET) as the key correspondence organizing innovation has been pulled in by the scholarly world and enterprises with surprising turn of events. With each vehicle acting as a node in an ad hoc network made up of immobile or mobile vehicles, the VANET, which connects vehicles over a wireless connection, is a developing research field that is garnering prominence. The authors of this study examined real-time vehicles and the outcomes of four routing protocols on the basis of three parameters are recorded using Network Simulator (NS-3) as network simulator and synchronized with Simulation on Urban Mobility (SUMO) as mobility simulator. A dataset is compiled using recorded results with NS-3 and SUMO. For selecting the efficient routing protocol, collection of dataset and selection of different features is done. Machine Learning (ML) models such as Random Forest (RF), Logistic Regression (LR), and k-Nearest Neighbor (k-NN) are implemented utilizing a set of relevant information regarding the relationship between sender and receiver. The effectiveness of ML models is assessed using a novel dataset and especially in comparison to that with others. The results shows that k-NN outperforms on the basis of evaluation parameters: F-score (75.5%), Accuracy (97.2%), Recall (79.9%) and Precision (75.3%) of classification learning techniques. The purpose of this research is prediction and analysis of ML Models for efficient routing protocol in VANET using different feature information that may be utilized to improve effectiveness of VANET and provided efficient routing protocol for safe, secure, reliable connection between vehicles.