Truncated-ARQ Aided Adaptive Network Coding for Cooperative Two-Way Relaying Networks: Cross-Layer Design and Analysis

• Main Authors: Yanping Yang, Wei Chen, Ou Li, Qingwen Liu, Lajos Hanzo
• Format: Article
• Language: English
• Published: IEEE 2016-01-01
• Series: IEEE Access
• Subjects: Cooperative communication; cross-layer design; two-way relaying; network-coded modulation; automatic repeat request
• Online Access: https://ieeexplore.ieee.org/document/7556288/

Network coding (NC) constitutes a promising technique of improving the throughput of relay-aided networks. In this context, we propose a cross-layer design for both amplify-and-forward and decode-and-forward two-way relaying based on the NC technique invoked for improving the achievable throughput under specific quality of service requirements, such as the maximum affordable delay and error rate. We intrinsically amalgamate adaptive analog network coding (ANC) and network coded modulation (NCM) with truncated Automatic Repeat reQuest (ARQ) operating at the different open system interconnection layers. At the data-link layer, we design a pair of improved NC-based ARQ strategies based on the stop-and-wait and the selective-repeat ARQ protocols. At the physical layer, adaptive ANC/NCM are invoked based on our approximate packet error ratio. We demonstrate that the adaptive ANC design can be readily amalgamated with the proposed protocols. However, adaptive NC-QAM suffers from an SNR-loss, when the transmit rates of the pair of downlink channels spanning from the relay to the pair of destinations are different. Therefore, we develop a novel transmission strategy for jointly selecting the optimal constellation sizes for both of the relay-to-destination links that have to be adapted to both pair of channel conditions. Finally, we analyze the attainable throughput, demonstrating that our truncated ARQ-aided adaptive ANC/NCM schemes attain considerable throughput gains over the schemes dispensing with ARQ, while our proposed scheme is capable of supporting bidirectional NC scenarios.