Scholarlink

<p>Routing in Mobile Ad Hoc Networks (MANETs) is complex due to their decentralized topology and dynamic environments. Conventional routing protocols frequently have trouble maintaining reliable end-to-end communication, leading to issues such as high packet loss, unstable routes, and increased delays. The research presented a novel deep-learning model with hybrid optimization to maintain route stability through stable node prediction and optimal route discovery in a MANET environment. Here Dynamic Sparse Recurrent-Convolutional Neural Network (DSR-CNN), an advanced variant of the R-CNN family is employed to detect the stable node in a sparse and dynamic environment. The weight parameters of the DSR-CNN are optimized by the Enhanced Fick’s Law Optimization Algorithm for the improved node prediction maintaining reliability and connectivity. Moreover, the most stable routes are constructed through the Giant Trevally Optimizer, optimizing route exploration and recovery. The model is designed and simulated in a Python environment with the required network settings. The utilization of the proposed research in the MANET communication shows 98% prediction accuracy, 2.48&#xa0;ms delay, 25Mbps throughput, and 99% of packet delivery ratio which are more efficient than the prevailing technique’s results showing the highest route stability. These findings highlight the potential of DSR-CNN to significantly improve the reliability and performance of MANETs, setting a new benchmark for future routing protocols.</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

Neural network-driven scenario prediction for adaptive routing in MANETs using expanding ring search and random early detection

  • 期刊论文

  • 发表日期：

    2025年3月27日

查看全文
  • M. A. Gunavathie,
  • Ujwal Ramesh Shirode,
  • Nichenametla Rajesh,
  • V. Sudha

摘要

Routing in Mobile Ad Hoc Networks (MANETs) is complex due to their decentralized topology and dynamic environments. Conventional routing protocols frequently have trouble maintaining reliable end-to-end communication, leading to issues such as high packet loss, unstable routes, and increased delays. The research presented a novel deep-learning model with hybrid optimization to maintain route stability through stable node prediction and optimal route discovery in a MANET environment. Here Dynamic Sparse Recurrent-Convolutional Neural Network (DSR-CNN), an advanced variant of the R-CNN family is employed to detect the stable node in a sparse and dynamic environment. The weight parameters of the DSR-CNN are optimized by the Enhanced Fick’s Law Optimization Algorithm for the improved node prediction maintaining reliability and connectivity. Moreover, the most stable routes are constructed through the Giant Trevally Optimizer, optimizing route exploration and recovery. The model is designed and simulated in a Python environment with the required network settings. The utilization of the proposed research in the MANET communication shows 98% prediction accuracy, 2.48 ms delay, 25Mbps throughput, and 99% of packet delivery ratio which are more efficient than the prevailing technique’s results showing the highest route stability. These findings highlight the potential of DSR-CNN to significantly improve the reliability and performance of MANETs, setting a new benchmark for future routing protocols.