Secure Authorization and Traffic Congestion Prediction in VANET Using R3-DWT and EDICNN with Blockchain

A Secure Transportation System (STS) improves traffic safety and enables vehicles to pass through non-congested routes. However, none of the prevailing works concentrated on protecting the data used for authorization in STS. Therefore, the secure authorization and Traffic Congestion Prediction (TCP) model is proposed. Initially, the user registers into blockchain and a key is generated using Koblitz Torus Curve Cryptography (KTCC). Then, the network is initialized and vehicle information is sensed, and then using KTCC, the data are secured. From the sensed data, the hash code is generated using Entropy Davies–Meyer Streebog (EDM-Streebog). Meanwhile, the vehicle’s number-plates are detected using You Only Look Once-Version 8 (YOLOV8). Now, in the number-plate image, the watermarking is done to protect the hash code through R-squared Regression Discrete Wavelet Transform (R3-DWT). Next, the hash code is retrieved in the blockchain and the verification is processed. For the verified hash, the secured data are decrypted and TCP is done. To train the TCP model, the high traffic video data are collected and pre-processed. Then, the day/night frame is identified. Then, the image conversion of the night frame is done and along with the day frame, the vehicle detection in the frame is done by YOLOV8. From the vehicle detected image, the vehicle tracking is done and then using Elliott Deep Identity Convolutional Neural Network (EDICNN), the TCP is made. Thus, the congestion details are notified to users for secure transportation. The proposed work thus predicted the traffic congestion effectively with an accuracy of 98.0213%, and recall of 98.3024%.