Adaptive deep conditional random field with blockchain for secure data sharing in software-defined wireless body area networks

Abstract

Abstract Presently, blockchains are widely employed to execute the secure data transmission process among users. However, sharing the sensitive information about the patients among multiple users in the healthcare sector is difficult due to integrity and confidentiality issues. Thus, to tackle these issues in prior models, a secure data-sharing framework for Software-Defined Wireless Body Area Networks (SDWBANs) is designed. In order to ensure privacy and controlled access in SDWBANs, blockchain technology and encryption techniques are considered, which help to protect sensitive medical data. Then, the Adaptive Deep Conditional Random Field (ADCRF) is employed to perform the decision-making procedure. Further, an advanced encryption technique, Optimal Key-based Multi-Authority Attribute-Based Encryption (O-MA-ABE), is employed to ensure that authorized users can access the confidential health data. Here, the Modified Escape Search-based Piranha Foraging Optimization Algorithm (MES-PFOA) is employed to tune the Hyperparameters of ADCRF and keys of O-MA-ABE. Then, the overall performance of the developed framework is compared with classical approaches using metrics such as computational time, decryption time, and decision-making accuracy. In various validations, the developed MES-PFOA-O-MA-ABE + ADCRF-based data sharing model accomplished higher accuracy as 98.7%, precision as 98.3%, minimal encryption time as 210 (ms) and throughput as 275 (TPS) than the recent data sharing models like DTAC-TL-QM, SCCE-DS, BFL-hIoT and PPFL-ICP.