Scholarlink

<p>Autonomous vehicles require precise real-time localization and secure V2X communication, yet GPS fails in urban canyons and centralized networks invite attacks. This paper proposes an adaptive IoT-blockchain framework integrating four components: Adaptive Hybrid Filtering Technique (AHFT) for sensor data cleaning; Spatio-Temporal Fusion Neural Network (STFNN) for spatial-temporal localization (MAE 1.8&#xa0;m, RMSE 0.9&#xa0;m in GPS-denied areas); Dynamic Consensus Load Balancing Protocol (DCLBP) for scalable blockchain (21ms latency at 500 veh/km²); and Smart Contract-Based Trust Verification System (SCTV) for 99.7% secure messaging. Evaluated on Hyperledger Fabric testbed with nuScenes/Waymo datasets (1000 vehicles), it achieves 95.2% accuracy, 78% latency reduction vs. baselines, and 98.7% packet delivery. Results outperform SOTA by 62% in localization and 3.2x throughput, with robustness to 30% failures.</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

Real-Time Vehicle Localization and Secure Communication Using Deep Learning with an Adaptive IOT Blockchain Framework

  • 期刊论文

  • 发表日期：

    2026年5月4日

查看全文
  • M. Muniraju,
  • S. Vikas Reddy

摘要

Autonomous vehicles require precise real-time localization and secure V2X communication, yet GPS fails in urban canyons and centralized networks invite attacks. This paper proposes an adaptive IoT-blockchain framework integrating four components: Adaptive Hybrid Filtering Technique (AHFT) for sensor data cleaning; Spatio-Temporal Fusion Neural Network (STFNN) for spatial-temporal localization (MAE 1.8 m, RMSE 0.9 m in GPS-denied areas); Dynamic Consensus Load Balancing Protocol (DCLBP) for scalable blockchain (21ms latency at 500 veh/km²); and Smart Contract-Based Trust Verification System (SCTV) for 99.7% secure messaging. Evaluated on Hyperledger Fabric testbed with nuScenes/Waymo datasets (1000 vehicles), it achieves 95.2% accuracy, 78% latency reduction vs. baselines, and 98.7% packet delivery. Results outperform SOTA by 62% in localization and 3.2x throughput, with robustness to 30% failures.