Decentralized federated deep Q-learning for IoMT security: leveraging MK-VQFHE and blockchain with IPFS.

The explosion of patient data, the demand for real-time insights, and the critical importance of data security drive healthcare innovation. Medical Internet of Things (IoMT) offers a promising solution, connecting medical devices, sensors, and healthcare systems to improve patient care. However, managing and securing vast amounts of complex data produced by IoMT devices remains a significant challenge. Existing approaches often fall short of providing comprehensive solutions. To address this, this research proposes a novel approach combining Multi-Key Verifiable Quaternion Fully Homomorphic Encryption (MK-VQFHE) and blockchain for decentralized Federated Learning (FL) to enhance data management capabilities. Initially, the proposed study collects IoT healthcare data from publicly available datasets after that, Multi-Key Verifiable Quaternion Fully Homomorphic Encryption for encryption is used to safeguard data security. Encrypted data is then used in a collaborative learning model enabled by blockchain technology. The proposed Federated Deep Q-learning (FDQL) model enhances privacy protection by training inputs and evaluating threats. The data is securely stored using Interplanetary File System (IPFS) technology within the blockchain network. This study introduces a Practical Byzantine Fault Tolerant (PBFT) consensus technique to verify proposed structure’s integrity. Performance metrics demonstrate proposed approach produce superior accuracy ranging from 99.2% to 99.4% across multiple datasets compared to existing approaches. Meanwhile the existing models such as BiLSTM, CNN, DNN, and ANN are attained accuracy of below 99%. The proposed research aims to enhance IoMT data security, introduce advanced encryption strategies, ensure efficient data storage, analyze privacy protection effectiveness, validate blockchain framework efficiency, and evaluate performance metrics for comprehensive insights.