Blockchain and Deep Learning-Based Decentralized Anomaly Detection Framework for VANET

The automotive industry is increasing exponentially day by day providing a safer and efficient experience to the consumers. Intelligent Transportation Systems (ITS) has become more crucial with the increasing number of vehicles. Moreover, there is a growing demand for intelligent services like efficient traffic monitoring in vehicle-to-vehicle (V2V) and vehicle-to-everything (V2X) communication environments. Vehicular Ad-hoc Network (VANET) represents the demanding and vibrant environment for securing the interaction between the vehicle, RSUs, and pedestrians through the Controller area network (CAN). RF jamming Attacks can be integrated through wireless mediums. The paper presents an improvement in securing VANETs through the combination of Blockchain technology (BT) with Deep learning (DL). In DL-based model, RNN is utilized for detecting and classifying malicious activities. BT ensures secure transactions and automated trust-less interactions using smart contracts. Thus, the non-attack data classified by DL layer is communicated using BT. IPFS is used for storing the information to enhance the scalability while mitigating the maintenance cost of BT. The accuracy achieved by the RNN model during evaluation was 97%. Multiple models were evaluated over metrics like accuracy, precision, recall, F1-score, and ROC-AUC curve.