Scholarlink

<p>The exponential expansion of the Internet of Things (IoT) has fueled the demand for secure, scalable, and privacy-preserving architectures that can support autonomous communication and intelligent decision-making at the network edge. Traditional centralized IoT systems suffer from latency, single points of failure, and data privacy challenges. To address these limitations, we propose FL-BEE (Federated Learning–Blockchain Enhanced Edge Ecosystem), a novel multi-layered framework that integrates Federated Learning, Blockchain, IPFS (interplanetary file system), and Apache Kafka to enable decentralized, secure, and intelligent collaboration among heterogeneous IoT devices. FL-BEE supports privacy-aware model training without raw data exchange, tamper-proof update tracking via blockchain, decentralized storage through IPFS, and adaptive control of model compression and update frequency using a Quantum Adaptive Firefly Optimization (QAFO) algorithm. The framework further ensures robust authentication via JWT (JSON web token) and PKI (public key infrastructure) mechanisms. Experimental results demonstrate significant improvements in classification accuracy (98.5%), latency (95 ms), energy efficiency, and blockchain performance (110 TX/sec), surpassing conventional and federated approaches.</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

FL-BEE-QAFO: federated learning blockchain enhanced IoT framework for autonomous device communication and data integrity using quantum adaptive firefly optimization

  • 期刊论文

  • 发表日期：

    2026年1月12日

查看全文
  • P. Bachan,
  • Ayalapogu Ratna Raju,
  • Naga Raju Hari Manikyam,
  • Mahesh Uday Mangaonkar,
  • Rajshree Jodha,
  • Neha Dutta,
  • Anuradha Pillai

摘要

The exponential expansion of the Internet of Things (IoT) has fueled the demand for secure, scalable, and privacy-preserving architectures that can support autonomous communication and intelligent decision-making at the network edge. Traditional centralized IoT systems suffer from latency, single points of failure, and data privacy challenges. To address these limitations, we propose FL-BEE (Federated Learning–Blockchain Enhanced Edge Ecosystem), a novel multi-layered framework that integrates Federated Learning, Blockchain, IPFS (interplanetary file system), and Apache Kafka to enable decentralized, secure, and intelligent collaboration among heterogeneous IoT devices. FL-BEE supports privacy-aware model training without raw data exchange, tamper-proof update tracking via blockchain, decentralized storage through IPFS, and adaptive control of model compression and update frequency using a Quantum Adaptive Firefly Optimization (QAFO) algorithm. The framework further ensures robust authentication via JWT (JSON web token) and PKI (public key infrastructure) mechanisms. Experimental results demonstrate significant improvements in classification accuracy (98.5%), latency (95 ms), energy efficiency, and blockchain performance (110 TX/sec), surpassing conventional and federated approaches.