Scholarlink

<p>Reliable Vehicle-to-Vehicle (V2V) and vehicle-to-infrastructure (V2I) communication are vital for enhancing the performance and safety of intelligent transportation systems (ITS). However, resource allocation faces major challenges owing to interference, limited channel State Information (CSI) feedback, and high vehicular mobility. This study proposes a hybrid framework that integrates Artificial Neural Networks (ANNs) with autoencoders to compress CSI and jointly optimize spectrum reuse and power allocation. The methodology involved training the model with simulated vehicular data and evaluating its performance under varying speeds, SNR levels, and interference conditions. Simulation results show that the proposed ANN–Autoencoder framework achieves a higher sum capacity, reduced outage probability, and improved fairness compared to baseline methods such as delayed CSI and the Kuhn–Munkres algorithm. Specifically, the model demonstrates significant improvements in SINR distribution, throughput, and minimum capacity guarantees, while balancing the trade-off between maximizing total throughput and ensuring fairness. These findings confirm that the proposed framework provides a scalable, low-complexity, and practical solution for real-time resource allocation in hybrid vehicular networks, meeting the demands of the emerging 5G/6G-enabled ITS.</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 Hybrid Approach for Vehicular Spectrum Allocation Using Artificial Neural Networks with Autoencoders and CSI Feedback

  • 期刊论文

  • 发表日期：

    2025年10月6日

查看全文
  • S. Sheela,
  • S. Jyothi,
  • B. P. Pradeep kumar

摘要

Reliable Vehicle-to-Vehicle (V2V) and vehicle-to-infrastructure (V2I) communication are vital for enhancing the performance and safety of intelligent transportation systems (ITS). However, resource allocation faces major challenges owing to interference, limited channel State Information (CSI) feedback, and high vehicular mobility. This study proposes a hybrid framework that integrates Artificial Neural Networks (ANNs) with autoencoders to compress CSI and jointly optimize spectrum reuse and power allocation. The methodology involved training the model with simulated vehicular data and evaluating its performance under varying speeds, SNR levels, and interference conditions. Simulation results show that the proposed ANN–Autoencoder framework achieves a higher sum capacity, reduced outage probability, and improved fairness compared to baseline methods such as delayed CSI and the Kuhn–Munkres algorithm. Specifically, the model demonstrates significant improvements in SINR distribution, throughput, and minimum capacity guarantees, while balancing the trade-off between maximizing total throughput and ensuring fairness. These findings confirm that the proposed framework provides a scalable, low-complexity, and practical solution for real-time resource allocation in hybrid vehicular networks, meeting the demands of the emerging 5G/6G-enabled ITS.