Onboard disease prediction and rehabilitation monitoring on secure edge-cloud integrated privacy preserving healthcare system

Abstract

Edge-based privacy preserving cryptosystem is identified as the upcoming amenities of cloud-based secure remote healthcare monitoring systems. Usually, the cloud-based healthcare system will directly collect the remote patient data through a sensor layer and provide the continuous monitoring and diagnosis through various prediction processes made by the decision support system. These sensing and processing of real-time patient's medical data without compromising its privacy and security become daunting issues in the traditional healthcare services. Therefore, the proposed research incorporates the security mechanism in the patient-centric edge-cloud-based healthcare system architecture. More precisely, an edge level privacy preserving additive homomorphic encryption is proposed for secure data processing and filtering non-sensitive data in the edge layer. In addition, response time and network capacity usage are minimized in the proposed healthcare system due to effective filtering and offloading mechanisms adapted in the edge level. Next, an adaptive weighted probabilistic classifier model is proposed in the cloud layer for onboard disease prediction and rehabilitation of remote patients. It will improve the disease prediction time and prediction accuracy while comparing to traditional classifier models. Finally, security and performance analysis of the proposed Secure Edge-Cloud-based Healthcare System (SECHS) was demonstrated with respect to empirical evaluation of Parkinson disease dataset.