Woo, Wen Hui (2024) Enhancement of throughput for dynamic device-to-device (D2D) communication in 5G networks through network coding. Master dissertation/thesis, UTAR.
Abstract
This dissertation delves into enhancing 5G network throughput for Device-to-Device (D2D) communication through relay selection and network coding. The devised relay selection methodology considers parameters such as Received Signal Strength (RSS) and the current battery status to identify the most suitable relay. This selected relay then engages in Random Linear Network Coding (RLNC) to transmit the message to the intended destination efficiently. While D2D communication was initially conceived for direct interaction without relying on a central infrastructure, interference and path loss can adversely impact performance, leading to retransmission delays. This model regards relay selection and network coding to improve throughput. Optimal relay nodes are determined based on the status of RSS and the remaining battery level. RSS effectively handles changing channel conditions, while battery level ensures reliable communication. RLNC improves data reliability by mitigating packet loss, optimising resource allocation, and enhancing transmission efficiency through encoding data into linear combinations. It combines multiple packets into smaller blocks and performs XOR operation with GF coefficients to reduce complexity and completion iii time for the transmission, improving throughput and reducing retransmission rate. A dynamic scenario is applied to fulfil near to the actual network. Next, it is required to design a proposed relay selection criterion based on the RSS with the threshold of the battery. Once the relay selection criterion is in place, the next step is to implement the RLNC in the relay nodes. After implementing RLNC, the nodes can perform encoding and decoding operations. Lastly, analyse the simulation results based on various factors such as packet loss, throughput, and end-to-end (E2E) delay to determine the ability of the proposed drelay selection criterion and RLNC implementation. Consequently, the suggested model adeptly enhances system throughput while diminishing packet loss and E2E delay under identical conditions, in contrast to alternative relay selection techniques. Furthermore, the network incorporating RLNC experiences a 26% increase in throughput when juxtaposed with those networks not utilising RLNC.
Actions (login required)
![[img]](http://eprints.utar.edu.my/style/images/fileicons/application_pdf.png)
