A Novel and Secure Machine Learning-Based Hyperledger Blockchain for IoT Healthcare
Sidra Aslam, Saba Aslam, Taotao Wang, Daquan Feng, Shengli Zhang · IEEE Internet of Things Journal · 2025
Data privacy protection and secure sharing are the main issues faced by smart healthcare IoT systems. In medical uses, patient health information is frequently kept in the cloud, which limits the user’s ability to entirely control their data. Additionally, standard encryption keys do not sufficiently mitigate the risks posed by malicious entities like compromised cloud service providers. To address these issues, blockchain technology, combined with Internet of Medical Things (IoMT) can securely safeguard patient medical records through a peer-to-peer, secure, and collective ledger. Therefore, we propose a novel IoT-driven architecture that leverages blockchain technology to protect patient medical files from tampering and unauthorized access. This architecture integrates patient medical files with blockchain and is enhanced by a combination of Bidirectional Long Short-Term Memory (BiLSTM) networks and Convolutional Neural Networks (CNN). Utilizing blockchain for the transmission of encrypted data significantly strengthens data security and minimizes the risk of data breaches. The process of generating encryption and decryption keys through a coupled CNN and BiLSTM ensures the robustness and uniqueness of these keys. Additionally, the selection of the best key is performed using the Gradient Descent Optimization Algorithm (GDOA), which demonstrates the effectiveness and efficiency of the encryption and decryption process. We also compare the implementation of our model with existing technologies, assessing its performance based on various metrics, including restoration efficiency, response time, record time, key generation time, encryption time, decryption time, turnaround time, and overall running time. Our proposed method is confirmed to be more effective than current techniques in terms of these performance metrics.