Exploiting user interaction in interference-aware wireless networks
Date of Issue2014
School of Electrical and Electronic Engineering
Due to the shared nature of the wireless media, the wireless users in a network nat- urally interact with each other and thus a ect the transmissions of each other. The aim of this thesis is to illustrate the crucial role of user interaction in designing ef- cient interference-aware communication schemes in wireless networks, from both user-centric and network-centric points of view. Speci cally, this thesis investigates various key issues on exploring wireless user interaction, such as tradeo analysis, co- existence criterion, algorithm complexity, and system capacity characterization, for wireless networks consisting of homogeneous and heterogeneous users, respectively. This thesis also demonstrates the usefulness of two mathematical tools, stochastic ge- ometry and dynamic programming, and shows their e ectiveness in designing e cient communication schemes in interference-aware wireless networks. First, in an ad hoc network consisting of homogeneous wireless users, to guide the transmit interaction between users, this thesis proposes a transmission scheme with signal-to-interference-ratio (SIR) threshold based scheduling. This thesis considers a stochastic ad hoc network, where the locations of the transmitters and receivers as well as the channel fading between any transmitter and receiver are randomly distributed. By using tools from stochastic geometry, this thesis focuses on analyzing the e ects of user interaction on the ad hoc network capacity. Speci cally, this thesis adopts a homogeneous Poisson Point Process (PPP) to model the locations of transmitters that have the intention to transmit. Due to the SIR-threshold based scheduling, it is shown that the resulting point process formed by the transmitters that are allowed to transmit is a non-PPP in general. This thesis tackles the challenges on analyzing the non-PPP based network, by proposing a new method to approximately characterizing the non-PPP network capacity. Such method is useful for analyzing performance of wireless networks with interacted transmitters. Moreover, this thesis reveals the exact/approximate dependence between the system performance and the system designing parameters, which gives designing insights for networks with SIR -threshold based scheduling. Next, in a cognitive radio (CR) network consisting of secondary users (SUs) and primary users (PUs), this thesis assumes that the PUs are able to react to the SU's collision or interference by adapting its transmit power level and/or adjusting its channel access probability. Such PUs are referred to as reactive PUs, to di erentiate from their non-reactive counterparts. This thesis proposes a 4-state Markov chain to model the channel occupancy of the reactive PU. Under a practical scenario where the PU's channel occupancy status is only partially known to the SUs, this thesis studies the opportunistic spectrum access (OSA) design for the SU, by adopting two methods to protect the reactive PU's transmission, respectively. One is the conven- tional short-term conditional collision probability (SCCP) constraint, and the other is a long-term PU throughput (LPUT) constraint. By exploiting a separation principle to separate the SU's channel selection from its spectrum sensor design and channel access decision on the selected channel without loss of optimality, an optimal OSA policy under the SCCP constraint and a suboptimal OSA policy under the LPUT constraint are obtained in this thesis, respectively. It is shown that the conventional SCCP constraint, which is e ective in protecting the non-reactive PUs, is not able to properly protect the reactive PUs, while the LPUT constraint can provide su - cient protection for the reactive PU's transmission. Interesting tradeo between SU throughput maximization and PU transmission protection are also revealed in this thesis.
DRNTU::Engineering::Electrical and electronic engineering