Improved Multiresolution Neural Network for Mobility-Aware Security and Content Caching for Internet of Vehicles

With the emerging communication and computation technologies, the Internet of Vehicles (IoV) has become a new paradigm that enables vehicles to communicate with roadside units (RSUs) as well as with other vehicles to collect or exchange information. However, the inherent characteristics of IoV, such as the high mobility of the vehicle and the limited storage capacity of the edge nodes, cause numerous difficulties in developing a caching scheme. Therefore, in this article, a novel improved multinomial recurrent neural network (MRNN) classifier and Caesar combined key-based elliptic curve cryptography (2CK-ECC) algorithm are proposed to predict the vehicular mobility, security, and content caching for the IoV. Initially, the vehicles are checked for registration. Afterward, vehicle login and vehicle authentication take place. Then, mobility prediction is carried out for the authenticated vehicles using the MRNN Classifier. After that, vehicle clustering is done via the novel variance-balanced iterative reducing and clustering using hierarchies (VBIRCHs) technique. Next, the relay vehicles (RVs) are selected using the novel Harris self-avoiding hawks optimization (HSAHO). Furthermore, the data contents are divided into chunks and saved in the cache memory. From the cache memory, data is transferred to the RSU. Finally, a novel 2CK-ECC algorithm is investigated for secure data transmission. The experimental outcomes demonstrate that the proposed technique outperforms the existing baseline approaches.