Cross-Layer Design for Quality of Service provisioning in AMC/ARQ-based wireless networks
In the last years, the explosive development of wireless services and applications has produced an unprecedented revolution in wireless communications systems. The impact of wireless fading channels on the quality of service (QoS) provisioning for such heterogeneous mobile users is one of the most c...
| Autor: | |
|---|---|
| Tipo de recurso: | tesis doctoral |
| Estado: | Versión publicada |
| Fecha de publicación: | 2012 |
| País: | España |
| Institución: | CBUC, CESCA |
| Repositorio: | TDR. Tesis Doctorales en Red |
| OAI Identifier: | oai:www.tdx.cat:10803/84121 |
| Acceso en línea: | http://hdl.handle.net/10803/84121 |
| Access Level: | acceso abierto |
| Palabra clave: | Informàtica 004 |
| Sumario: | In the last years, the explosive development of wireless services and applications has produced an unprecedented revolution in wireless communications systems. The impact of wireless fading channels on the quality of service (QoS) provisioning for such heterogeneous mobile users is one of the most challenging issues for nextgeneration wireless networks. In order to support the diverse QoS requirements of wireless applications, innovative techniques have been proposed at the physical (PHY) layer. Among them, the rate-adaptive modulation and coding (AMC) scheme has received significant research attention, resulting in its adoption in most state-of-the-art wireless communications standards. With the aim of enhancing the link reliability to guarantee the QoS constraints of new applications, most modern communication systems use error control strategies at the data link control (DLC) layer. Among them, we can distinguish two basic approaches: the forward error correction (FEC) scheme, in which an errorcorrection code is used, and the automatic repeat request (ARQ) scheme, in which a code with good error-detection capability is used. In the latter case, when a received codeword is detected in error, the packet is retransmitted until it is correctly received (infinitely persistent ARQ) or until a preset number of retransmissions have taken place (truncated ARQ). In order to achieve the advantages of both strategies, most state-of-the-art wireless communications standards use combinations of ARQ and FEC. Moreover, recent proposals make use of one or multiple intermediate relay stations to forward data from a source node to the corresponding destination node. As a result, a performance improvement is obtained due to spatial diversity, which is generated by transmitting signals from different locations (source and relay(s)), thus providing independently faded versions of the signal at the receiver. Cross-layer design in wireless networks, where one allows the stack protocol layers to interact and share information, has become increasingly popular over the past few years. In particular, many recent cross-layer proposals coincide in combining AMC at the PHY layer with an ARQ protocol at the DLC layer. These cross-layer designs improve the spectral efficiency by jointly exploiting the adaptability of AMC to the wireless channel conditions and the error-correcting capability of ARQ-based error control strategies. The main goal of this dissertation is to provide a unified view of the crosslayer design, analysis and optimization of AMC/ARQ-based wireless systems to allow the joint optimization of both the PHY and the DLC layers. The adopted approach for tackling this problem will rely on the use of discrete time Markov i chains (DTMCs) to jointly consider the packet arrival, the queueing process and the PHY layer. To that end, a novel first-order two-dimensional Markov model of the PHY layer is developed, which takes into account the wireless channel characteristics and de AMC scheme. The availability of accurate models for the PHY layer characterization is one of the key motivations of this work in order to guarantee a correct analysis of the QoS metrics at the DLC layer. Using this model, the interactions between the PHY and DLC layers are analyzed either when infinitely persistent or truncated ARQ-based error control protocols are implemented. Additionally, a cooperative scheme is also proposed, in which the relay node is in charge of retransmitting the erroneously received packets at the destination. This DTMC-based model allows the analytic derivation of various system performance metrics, namely, throughput, average packet delay and packet loss rate (both due to buffer overflow and due to exceeding the maximum number of allowed retransmissions). For the sake of comparison with non Markov-based analytical tools, the infinitely persistent ARQ protocol is also analyzed through the effective bandwidth/capacity theory. Both analytical frameworks are compared, showing the superiority of the Markov-based approach, which more faithfully reproduces the real system behaviour at the cost of higher complexity in the analysis. The proposed analytical framework allows the formulation of cross-layer multidimensional design strategies, aiming at the maximization of the average throughput of the system while satisfying prescribed QoS requirements in the form of average packet loss rate and average delay. Finally, an explicit analysis of the impact on the system performance of the delay in the channel state information (CSI) feedback is also presented. |
|---|