Candidate
Refga Elgadi
Title
An Intelligent Multi-stage Channel Acquisition Model for CR-WBANs: A Context Aware Approach
Supervisors
Otman Basir and Allaa R. Hilal (Adjunct)
Abstract
Cognitive Radio (CR) came as a solution to mitigate challenges that WBANs suffer from. CR is an intelligence-based technology that senses, observes, and learns from its operating environment to access licensed bands in the spectrum, when they are not being utilized by primary users. Deploying a CR technology in WBANs applications, enhances spectrum scalability, increases system robustness, and decreases latency. Accordingly, CR-WBANs help in building a more efficient and reliable ubiquitous healthcare system than conventional WBANs do. However, CR-WBANs are still evolving, and many challenges need to be investigated, in particular, is how to acquire a channel and prioritize data streams among multiple CR-users (i.e., multiple patients) based on the severity of their health status, in a manner to decrease network latency and increase network scalability. To address this challenge, this work proposes a novel intelligent channel acquisition model for multiple CR-WBANs within ubiquitous healthcare system, whereby contextual data, namely, channel properties, intra-node characteristics, and patients’ profile information, is integrated in channel acquisition decision process. The proposed work is a multi-stage fusion system that is composed of local and global decisions units. A fuzzy logic system is utilized to make decisions in the local unit, which are sensing channel availability and assessing the severity of patient's health status. Moreover, a neural network is employed as a global sensing decision center, whereby local sensing decisions, channel properties, and intra-node characteristics are augmented in the decision process. Furthermore, a cluster-based heuristic algorithm is formulated, in the global decision unit, to prioritize data transmission among CR-users based on the criticality of their health conditions (i.e., acute, urgent, and normal). Patients' local health assessments and patients' avatars (e.g., age, medical history, etc.) are exploited in the prioritization process. The proposed model has improved spectrum sensing accuracy and channel acquisition probability, for all CR-users in the network, under the consideration of the severity of their health status. Thus, network latency has reduced and network scalability has increased, and so more lives can be saved. The proposed work has gone through extensive experimental simulations to evaluate its performance. The results have shown that the channel acquisition model is robust, scalable, accurate, and reliable in acquiring a channel and prioritizing data streams among patients based on their health conditions.