【作者】 Reza Ghazizadeh；

【导师】 范平志；

【作者基本信息】 西南交通大学 ， 通信与信息系统， 2009， 博士

【摘要】 近些年来,高速因特网接入,多媒体服务应用的快速发展,以及这些应用需求无处不在的特性(尤其是特殊应用领域),使得宽带无线接入(BWA)通信的需求大幅增加。目前,已出现了诸如IEEE802.11,IEEE802.16以及IEEE 802.20的宽带无线网络,以满足快速增长的无线接入需求。然而,宽带无线接入中存在一个重要的难题,即如何为各类需要不同服务质量(QoS)保证的业务提供要求的服务。业界普遍认为,调度控制算法在宽带无线接入网络内的QoS结构中起了至关重要的作用。所以,如何设计有效的调度控制算法是宽带无线接入网络研究的关键技术之一。本博士论文主要致力于研究自适应调度控制算法,以便对诸如IEEE802.11和IEEE 802.16.等的无线网络内的多媒体业务提供要求的QoS保证。首先,本文提出了一种无线局域网中基于竞争的介质接入过程中的动态QoS保障机制。在动态业务流优先级和传输时延估计的基础上,此保障机制不仅有更好的吞吐率和丢包率表现,并能使得这些参数在某个时间尺度下的变化范围更窄。除此之外,在不牺牲低优先级业务的前提下,与基于固定优先级的方法相比,这种机制提供了用户之间更好的公平介质共享,并大幅改善了系统性能。由于众多多媒体应用都有变化的比特率特征,因此设计实现一个有效的调度控制算法便成为了IEEE802.11e内基于轮询的介质接入过程中的一个挑战。在IEEE802.11e标准草案中提及的简单调度控制算法无法满足该业务所需要的QoS保证。所以,本文提出了一种新的与自适应传输机会相关的针对无竞争介质接入过程的业务调度控制算法并对其性能进行了评估。与文献中提到的现有的主流调度算法相比,本文提出的新的调度算法在吞吐率和时延性能上获得了极大的改善。根据实队列的大小和预测队列大小,使用动态的带宽选择,可以降低固定传输机会引入的带宽浪费。然后,本文提出了一种新的优先级分析模型,此模型可用于分析并增强指定QoS要求下的混合协调函数控制信道接入(HCCA)的性能。此模型同时也提供了高优先级和低优先级的数据包之间的差异服务。在这种模型中,节点是按照1/k相位分布(PH)马尔可夫到达过程(MAP)中的非抢占式包含空闲的优先级队列进行建模的。此外,通过在数据包到达模型中使用一个MAP,以及在服务和空闲过程和队列优先级模型中使用PH分布,可以使理论分析容易处理。接着,本文对提出的队列模型进行了扩展。该模型支持采用自适应调制和编码(AMC)的多速率非理想信道,而AMC在无线网络中是增加有效传输率的关键技术。为了对系统性能进行分析,本文采用了有限状态的马尔可夫链并将其节点视作有空闲和时限的MAP/PH/1队列。提出了一个新的定义在服务期间的称作无传输期的概念,它使服务器在传输信道恶劣时停止数据包传输。通过使用强大的矩阵几何方法,可以获得平稳系统状态概率分布。论文还给出了优化后的性能测试方法,以获得最好的跨层设计参数。为了在WiMAX环境中为分类业务流提供QoS保证,并在用户之间实现可用带宽的公平分配,本文提出了一个基于一种信用前提公平队列算法的双层信道感知调度控制算法。提出的算法在无线链路层中实现了一种无错、基于信用前提、联合AMC的多速率介质中的公平队列。此调度控制机制适用于WiMAX环境中定义的两个场景。首先,此种调度机制适用于点对多点的场景,作者在多类业务和信道条件下对此调度控制机制进行了研究。接下来,通过扩展此调度机制,使其适用于Mesh节点的中心调度控制场景,在此场景下某些用户可以在离基站两跳以上距离的情况下得到定位。论文还对系统的公平服务及时间介质接入性能进行了评估。仿真结果表明,两种场景下的信道传输速率选择机制以及补偿技术提供了一种在良好系统性能得以实现前提下的公平的多用户带宽共享机制。综上所述,本文提出了多个调度控制算法来公平地管理和分配不同传输介质下的系统资源,并论证了自适应调度控制机制在各种系统和信道条件下的优越性能。更多还原

【Abstract】 Rapid growth in high-speed Internet access and multimedia service applications, along with the ubiquitous availability, especially in hard-to-reach areas has increased the demand for broadband wireless access (BWA) communications. Emerging broadband wireless networks, such as IEEE802.11, IEEE802.16 and IEEE 802.20, are promising solutions to support the rapidly growing demands. However, one of the main challenges in BWA is to provide the quality of service (QoS) for heterogeneous traffic streams requiring various QoS guarantees. On the other hand, it is widely accepted that the scheduling algorithm is the key feature of QoS architecture in BWA networks. Consequently, the design of efficient scheduling algorithms is a critical issue. It is the aim of this dissertation to develop adaptive scheduling algorithms with QoS guarantees for multimedia traffic in broadband wireless networks, such as IEEE802.11 and IEEE 802.16.In the first part of this thesis, a dynamic QoS enabling mechanism in contention-based medium access over wireless local area networks is introduced. Based on dynamic traffic prioritization and transmission delay estimation, not only better throughput and packet delay are obtained but also variations of these parameters become narrower over a timescale. Furthermore, without starving the low priority traffic, the mechanism provides better fair medium sharing among the users and significant improvement in the system performance in comparison to the fixed priority-based method introduced in the standard.Since a lot of multimedia applications have variable bit rate characteristics, designing an efficient scheduling algorithm is a challenging task in polling-based access medium over IEEE802.11e. The simple scheduling algorithm introduced in the draft version of the standard can not satisfy the required QoS guarantees. Next, the operation and evaluation of a new traffic scheduling algorithm for contention-free medium access referred to adaptive transmission opportunity (ATX) are presented. Comparing to the dominant existing algorithms in literatures, the proposed new scheduling algorithm indicates a great performance improvement in term of throughput and delay. Using a dynamic bandwidth selection, according to real queue size and predicted queue size, the bandwidth wastage offered in the fixed transmission opportunity can be reduced.Then, a novel priority analytical model that is useful to analyze and enhance the performance of hybrid coordination function controlled channel access (HCCA) in provisioning required QoS is introduced. The model provides service differentiation between two classes of packets, namely, high priority and low priority packets. In this model, nodes are modeled as Markovian arrival process (MAP)/phase type (PH) distribution/1/k non-preemptive priority queues with vacation model. Using a MAP for packet arrival process, PH distribution for service and vacation processes, and the priority queuing model with the vacation make the analysis tractable.In the succeeding part, an extension of the proposed queuing model is presented. The model supports the multi-rate non-ideal channel using adaptive modulation and coding (AMC), which is the key feature in the wireless networks to increase the effective transmission rate. To analyze the performance of the system, the finite state Markov chain channel model is adopted and nodes are considered as MAP/PH/1 queues with vacation and time-limited. A new concept of non-transmission period, defined in the service period, enables the server to stop the packet transmission if the transmission channel is dirty. Using a powerful Matrix-geometric method, the stationary probability distribution for the system states is obtained. Finally, the performance measures followed by an optimization method to catch the best amount of parameters in the cross-layer design are presented.To provide QoS guarantees for classified traffic streams and to distribute available bandwidth fairly among the users in the WiMAX environment, a two-layer channel-aware scheduling algorithm based on an adaptive credit-based fair queuing (ACFQ) algorithm is presented. The proposed algorithm develops an error-free credit-based fair queuing (CBFQ) for the multi-rate medium with AMC at the radio link layer. The scheduler is adapted in two different scenarios defined in the WiMAX standard. First, the scheduler is applied to the point-to-multipoint (PMP) scenario and the operation of scheduler in the presence of various types of traffic and channel conditions is evaluated. In the next step, the scheduler is extended to apply to centralized scheduling scenario in mesh mode, where some users can be located more than one hop away from the base station. The system performance in terms of fair service and temporal medium access is evaluated. Simulation results show that the channel transmission rate selection scheme and compensation techniques, in both scenarios, provide a fair sharing of bandwidth among the users while excellent system performance is achieved.In conclusion, this dissertation has provided some scheduling algorithms to manage and allocate the system resources fairly in different types of transmission medium, and shows significant positive effect of adaptive schedulers under various system and channel conditions.更多还原