Scholarlink

<p>The Vehicular Ad Hoc Network (VANET) has gained increasing popularity due to its seamless services and flexibility. The VANET allows communication among the vehicles by vehicle-to‐vehicle (V2V) and vehicle‐to‐infrastructure (V2I) protocols. Reliable network-wide communication depends heavily on the implementation of secure clustering. Nevertheless, the ever-changing nature of VANET network topologies makes traffic prediction more complicated. To address this complexity, this study introduces the Chronological Lyrebird Optimization Algorithm (CrLOA) for Cluster Head (CH) selection and a CrLOA-trained Deep Long Short-Term Memory (DLSTM) model, referred to as CrLOA_DLSTM, for traffic prediction. Registration, authentication, V2I communication, V2V communication, CH selection, and traffic monitoring are the major processes in this model. For CH selection, the proposed CrLOA utilizes fitness parameters like distance, energy, trustworthiness, throughput, and delay. The DLSTM model, trained using the CrLOA algorithm, is employed to perform traffic prediction. Moreover, the energy, delay, Packet Delivery Rate (PDR), and trust are considered to estimate the CH selection, and the corresponding values of 0.904&#xa0;J, 0.702ms, 90.7%, and 74.97 are obtained. In addition, the precision and recall metrics are considered to validate the CrLOA_DLSTM–based traffic prediction, which yields the finest outcome of 90.20% and 89.20%.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A novel authentication and communication protocol with chronological lyrebird optimization enabled cluster head selection and traffic prediction using deep learning in VANETs

  • 期刊论文

  • 发表日期：

    2026年1月7日

查看全文
  • Vijayakumari L,
  • Gurumoorthy K B

摘要

The Vehicular Ad Hoc Network (VANET) has gained increasing popularity due to its seamless services and flexibility. The VANET allows communication among the vehicles by vehicle-to‐vehicle (V2V) and vehicle‐to‐infrastructure (V2I) protocols. Reliable network-wide communication depends heavily on the implementation of secure clustering. Nevertheless, the ever-changing nature of VANET network topologies makes traffic prediction more complicated. To address this complexity, this study introduces the Chronological Lyrebird Optimization Algorithm (CrLOA) for Cluster Head (CH) selection and a CrLOA-trained Deep Long Short-Term Memory (DLSTM) model, referred to as CrLOA_DLSTM, for traffic prediction. Registration, authentication, V2I communication, V2V communication, CH selection, and traffic monitoring are the major processes in this model. For CH selection, the proposed CrLOA utilizes fitness parameters like distance, energy, trustworthiness, throughput, and delay. The DLSTM model, trained using the CrLOA algorithm, is employed to perform traffic prediction. Moreover, the energy, delay, Packet Delivery Rate (PDR), and trust are considered to estimate the CH selection, and the corresponding values of 0.904 J, 0.702ms, 90.7%, and 74.97 are obtained. In addition, the precision and recall metrics are considered to validate the CrLOA_DLSTM–based traffic prediction, which yields the finest outcome of 90.20% and 89.20%.