Author: HyukJoon Kwon

Publisher:

Published: 2010

Total Pages:

ISBN-13:

The design of efficient algorithms using cooperation has gained much attention as an emerging technique for the next-generation wireless system. Such a cooperative system allows wireless devices to communicate with each other over relaying. With astute cooperative algorithms, wireless systems are expected to increase sum-rate performance and to support reliable communication. In general, wireless systems can be classified into centralized cellular infrastructures and decentralized ad-hoc multi-hop networks. Cellular networks require high quality channel information to increase sum-rate performance. However, due to finite-rate feedback channels, the base station cannot obtain uncorrupted channel information from mobile stations (MSs), thereby preventing the improvement in the sum-rate performance. On the other hand, multi-hop networks also require high level credit information about neighbor nodes to support reliable communication. Otherwise, traffic is likely to stop at some selfish nodes while being relayed to the destination. The first part of this thesis is motivated with the challenging issue: saving the number of feedback bits while maintaining sum-rate performance. To achieve the objective, this work exploits the cooperation between MSs, known as conferencing, in addition to feedback channels. It has been theoretically shown that cooperating encoders increase the capacity region in multiple-access channels. Similarly, the increase of achievable rate region in broadcast channels with cooperating decoders has been also revealed. In practical systems, the feedback rate is finite as well as cooperation is imperfect. Therefore, it is essential to exploit both cooperation and feedback effectively. Moreover, when multiple MSs are considered, multi-user diversity can be also exploited as yet another independent resource. Using these resources, i.e., feedback, cooperation and user-selection, available in broadcast channels, this thesis introduces the enhancement of the sum-rate performance through rigorous investigation of the relation among the resources. Moreover, this work derives the requirement for the number of feedback bits that achieve the multiplexing gain. Simulation results are presented to evaluate the sum-rate and to verify the derivation. The second part of this thesis focuses on multi-hop networks where each node operates independently without any centralized base stations. This multi-hop network can use cooperation among nodes to increase the total throughput with respect to a single-hop network. However, each node is autonomous and selfish in nature, and thus spontaneous cooperation among nodes is challenged. To accommodate this otherwise selfish nature of multi-hop networks, this thesis proposes a cooperative relay strategy under an energy-limited condition with a game-theoretic perspective. The main focuses are 1) to motivate each node to be cooperative, 2) to decide optimally the amount of cooperation, 3) to analyze equilibrium for the proposed scheme, and thus 4) to maximize the overall throughput. The proposed scheme formulates a Stackelberg game where two nodes sequentially bid their willingness weights to cooperate for their own benefit. Accordingly, all the nodes are encouraged to be cooperative only if a sender is cooperative and alternatively to be non-cooperative only if a sender is non-cooperative. This selective strategy changes the reputations of nodes depending on the amount of their bidding at each game and motivates them to maintain a good reputation so that all their respective packets can be treated well by other relays. Thus, every node forwards other packets with higher probability, thereby achieving a higher overall payoff.