Design of congestion control mechanism for multi-source data fusion infrastructure in telemedicine

Recent advances in wireless communications and integrated circuits have enabled design, development and implementation of wireless body area networks (WBANs). This class of network is paving techniques to develop innovative remote healthcare monitoring systems. The pervasive healthcare systems provide continuous monitoring for the chronically ill and elderly people to survive them in their independent lives. Despite having significant improvements, there are still many considerable issues and challenges that might influence quality of service (QoS) of WBANs. Therefore, the main objective of this research is to identify the existing gaps in providing a better QoS in the network. Although many problems can be addressed in this area, only one of them are studied in this thesis to address the issue of congestion. The congestion avoidance, congestion detection and control protocol are proposed to ensure acceptable data flows are maintained during the network lifetime. In this approach, the nodes play an important role in avoidance congestion, detecting congestion and also prioritizing the data received by them. That essentially helps to curb down problems caused due to congestion occurrence in healthcare applications. The proposed protocol contains of two phases. At the first phase, local data of sensor nodes is sent to avoid congestion by Active Queue Management (AQM) based on two methods to congestion Avoidance and supply QoS by proactively-dropping packets. With these system, congestion is capable of controlled, and network performance for example delay and packet loss. Second phase is based on three systems. The first system, sensor nodes are sent to congestion detection that based on two approaches that operates separately and respectively. The first approach of congestion detection based on node rate and mean inter arrival time, the second approach of congestion detection based on virtual queue state, node rate, packet service time and their automation to analyse the network traffic. According to the system output, congestion level is assessed. As for the second system, in case of congestion, the parent nodes dynamically compute and assign the new transmission rate for each of its children. In this system,physiological signs are discriminated to ensure enhancing QoS by transmitting highly important data. Finally, the third system by implicit congestion notification method to send notification messages over the network. The results show that proposed protocol is better than Congestion Control Scheme based on Fuzzy Logic (CCSFL) and A Prioritization Based Congestion Control Protocol in WSNs (PBCCP) protocols with different QoS. The percentage of excellence for the proposed protocol in total of throughput reaches 51% and 52% compared to CCSFL and PBCCP protocols, respectively. On the other hand, the percentage of excellence in the proposed protocol was 66% and 75% compared to CCSFL and PBCCP protocols in terms of the packet loss. In addition, the proposed protocol decreases the energy consumption based on traffic load by about 59% and 66% compared to CCSFL and PBCCP protocols, respectively. Furthermore, the percentage of excellence in proposed protocol based on End-To-End delay was 68% and 81% compared to CCSFL and PBCCP protocols. While, the percentage of excellence in proposed protocol was 94% and 98% compared to CCSFL and PBCCP protocols in terms of jitter. The proposed protocol achieved better results even as the number of the nodes increased.

Preview Text FK 2018 62 IR.pdf Download (1MB) | Preview

Actions (login required)

View Item