Optimisation of wireless multimedia networks with MIMO antennas: a cross-layer approach

下一代无线通信的最终目标是在诸如个人数字助理（PDA）的移动的终端和计算机服务器之间提供无处不在的无缝连接，使得用户可以在任何时间任何地点享受高质量的多媒体服务而无需有线。然而，诸如信道衰落和干扰之类的基本物理挑战已经对无线电资源施加了压力，使得难以实现可靠的无线通信。为了响应快速增长的数据速率需求，人们特别关注的是所谓的MIMO天线（或者在发射机和接收机处都具有多个天线的配置），该协议巧妙地利用了天线之间的空间特征差异来再生传输带宽。本项目的目的是设计一种智能调度协议，最大限度地利用MIMO下行链路信道其中通过对信道变化的适应，从多天线基站到许多多天线移动的终端进行通信。具体来说，我们的目标是最小化基站的总功率，同时满足用户的错误率和延迟的限制。该问题对于在无线信道中提供高质量的多媒体服务是重要的。所提出的研究联合处理在每个分组时间（即，媒体接入控制（MAC）层中的主要功能）以及它们如何使用MIMO传输来接入信道（即，所提出的联合MAC-PHY优化的优点可以解释如下。用户通常位于不同的位置，他们独立褪色。在特定的时间常数上，某个用户可能处于深度衰落中，但有些用户可能不是，这有助于所谓的多用户分集。智能MAC调度可以通过适应用户之间的信道时间变化来最大化信道利用率。另一方面，从物理层的角度来看，MIMO传输可以被优化，以充分利用用户之间空间上的信道变化。为了最大化信道利用率，正在体验良好信道的用户应该获得对信道的访问。然而，对于具有MIMO天线的信道，良好的感觉是不好定义的，因为用户可能不具有对于所有天线一致地比另一用户的更好的信道。此外，当我们在下行链路中促进频谱重用时，用户间干扰将进一步恶化良好信道的定义。此外，就延迟而言，即使是具有坏信道的用户也应该在他/她的数据到期之前获得对信道的访问。这些相互冲突的目标使优化变得复杂。为了实现我们的目标，我们需要克服上述挑战，并最终能够联合设计可以观察和适应信道变化的调度和传输。可以预期，通过增强PHY和MAC之间的交互，系统吞吐量可以大大提高，远远超过目前使用的系统。

The ultimate goal of future-generation wireless communications is to provide ubiquitous seamless connections between mobile terminals such as personal digital assistant (PDA), and computer servers so that users can enjoy high-quality multimedia services at anytime anywhere without wires. Fundamental physical challenges such as channel fading and interference, however, have put strains on the radio resources that make reliable wireless communications difficult to achieve. In response to the rapidly increasing demand of the data-rate requirements, of particular attention is the so-called MIMO antenna (or a configuration that has multiple antennas at both the transmitter and the receiver), which cleverly exploits the difference between spatial signatures among antennas to reproduce transmission bandwidth.The aim of this project is to devise a smart scheduling protocol that maximises the utilisation of a MIMO downlink channel where communication occurs from a multi-antenna base station to many multi-antenna mobile terminals, through adaptation to the channel variations. Specifically, our aim is to minimise the total power of the base station while satisfying the users' error-rate and delay constraints. The problem is important for providing high-quality multimedia services in wireless channels. The proposed research jointly deals with scheduling of the users that access the channel for each packet time (i.e., the main function in the medium access control (MAC) layer) and how they access the channel using MIMO transmission (i.e., the main function in the physical (PHY) layer).The merit of the proposed joint MAC-PHY optimisation can be explained as follows. Users generally situated at different locations and they fade independently. At a particular constant in time, a certain user may be in deep fade but some may not be, and this contributes the so-called multiuser diversity. Intelligent MAC scheduling can maximise the channel utilisation by adapting to the channel variations in time among users. On the other hand, from the PHY point of view, MIMO transmission can be optimised to take full advantage of the channel variations in space among the users.To maximise the channel utilisation, users who are experiencing a good channel should gain access to the channel. However, for a channel with MIMO antenna, a sense of goodness is ill defined as a user may not have unanimously better channels for all antennas than another user's. Besides, as we are promoting spectral reuse in downlink, inter-user interference will further exacerbate the definition of a good channel. In addition, as far as delay is concerned, even a user who has a bad channel should gain access to the channel before his/her data expire. These mutually conflicting objectives complicate the optimisation.To accomplish our objective, we need to overcome the above challenges and in the end be able to design the scheduling and transmission jointly that can observe and adapt to the channel variations. It is anticipated that through enhanced interaction between PHY and MAC, the system throughput can be improved far exceeding systems in use today.