A Robust Vehicular Networking Management System for the Traffic Control Using PK-DCNN Classification and L-SSA Based Rerouting

Abstract

For implementing a myriad of applications associated with vehicles, vehicle traffic, drivers, passengers, along with pedestrians, vehicular networking acts as the most imperative enabling technology that is needed. The main concern is handling vehicular networking for evading traffic congestion (TC). A ro-bust vehicular networking management system (VNMS) is presented by the work for controlling traffic-utilizing Polynomial Kernel-Based Deep Convolutional Neural Network (PK-DCNN). Utilizing Modified Adler 32 (M-Adler32), the work has initially established a vehicle authentication process where the user identity along with vehicle data is encrypted. Secure data transmission against internal along with external attacks is offered by vehicle authentication. Then, the cloud downtime is checked. The optimal cloud server availability is searched utilizing the Polynomial Distribution based BAT algorithm (PD-BAT) algorithm if the cloud downtime is high otherwise, attribute selection is executed. Lastly, centered on attribute extraction, the PK-DCNN classification categorizes whether there is traffic or not. Rerouting is conducted utilizing Linear Scaling Based Sparrow Search Algorithm(LSSA) with a low response time if the outcome attains high traffic. Robust traffic detection is attained by the proposed work via attaining 95.27% detection accuracy, 95.94% recall as indicated by the result. A low response time of 7811ms is obtained for recommending rerouting of the vehicles and stays ro-bust as analogized to the existent top-notch methods.