Scholarlink

<p>Smart cities are increasingly adopting IoT infrastructure to enhance smart manufacturing and industrial automation. However, existing systems face critical security and operational challenges: (1) latency in real-time data processing compromises emergency response and service reliability; (2) centralized architectures create single points of failure, making systems vulnerable to Distributed Denial-of-Service (DDoS) attacks and data tampering; (3) weak encryption exposes sensitive IoT data to eavesdropping and unauthorized access; and (4) conventional validation mechanisms lack adaptability to detect sophisticated transaction fraud or cyberattacks. To address these gaps, this paper proposes an Iot-based blockchain framework integrated with deep learning for secure smart city ecosystems. The framework collects raw Iot data and pre-processes it using Adaptive Data Cleaning (ADC) with a Denoising Autoencoder (DAE) to eliminate noise. An Improved Grey Wolf Optimizer (IGWO) enhances feature extraction via Levy Flight (to diversify exploration) and adaptive position updates (accelerating convergence by dynamically weighting leadership roles). Processed data is secured in a blockchain network, where transactions undergo decentralized validation. A one-dimensional Convolutional Neural Network (IDCNN), optimized via grid search, detects anomalies (e.g., tampering, fraud) during validation. The IDCNN employs batch normalization and dropout layers to stabilize training. Malicious transactions trigger immediate termination, ensuring robust security. Tested on real-world IoT datasets, the framework achieves 98.7% accuracy and 0.03% error, surpassing benchmarks like SVM (83%), ANN (91%), and Naive Bayes (70%). Key metrics like precision (97.5%), specificity (97.7%), and F1-score (94.2%) highlight its reliability. By integrating blockchain’s immutability, metaheuristic optimization, and deep learning, the framework mitigates latency, centralization risks, and cyber threats. This work advances smart city security through decentralized trust, real-time validation, and adaptive attack detection, offering a scalable solution for resilient urban ecosystems.&#xa0;</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

IoT-based blockchain secure data for smart cities infrastructures utilizing improved grey wolf optimisation with deep learning

  • 期刊论文

  • 发表日期：

    2026年1月27日

查看全文
  • Isaac Osei Asante

摘要

Smart cities are increasingly adopting IoT infrastructure to enhance smart manufacturing and industrial automation. However, existing systems face critical security and operational challenges: (1) latency in real-time data processing compromises emergency response and service reliability; (2) centralized architectures create single points of failure, making systems vulnerable to Distributed Denial-of-Service (DDoS) attacks and data tampering; (3) weak encryption exposes sensitive IoT data to eavesdropping and unauthorized access; and (4) conventional validation mechanisms lack adaptability to detect sophisticated transaction fraud or cyberattacks. To address these gaps, this paper proposes an Iot-based blockchain framework integrated with deep learning for secure smart city ecosystems. The framework collects raw Iot data and pre-processes it using Adaptive Data Cleaning (ADC) with a Denoising Autoencoder (DAE) to eliminate noise. An Improved Grey Wolf Optimizer (IGWO) enhances feature extraction via Levy Flight (to diversify exploration) and adaptive position updates (accelerating convergence by dynamically weighting leadership roles). Processed data is secured in a blockchain network, where transactions undergo decentralized validation. A one-dimensional Convolutional Neural Network (IDCNN), optimized via grid search, detects anomalies (e.g., tampering, fraud) during validation. The IDCNN employs batch normalization and dropout layers to stabilize training. Malicious transactions trigger immediate termination, ensuring robust security. Tested on real-world IoT datasets, the framework achieves 98.7% accuracy and 0.03% error, surpassing benchmarks like SVM (83%), ANN (91%), and Naive Bayes (70%). Key metrics like precision (97.5%), specificity (97.7%), and F1-score (94.2%) highlight its reliability. By integrating blockchain’s immutability, metaheuristic optimization, and deep learning, the framework mitigates latency, centralization risks, and cyber threats. This work advances smart city security through decentralized trust, real-time validation, and adaptive attack detection, offering a scalable solution for resilient urban ecosystems.