【作者】 梁晓雯；

【导师】 朱近康；

【作者基本信息】 中国科学技术大学 ， 通信与信息系统， 2006， 博士

【摘要】 自适应分配技术是正交频分复用(OFDM)系统对抗衰落信道带来的损失的一种强有力的方法，它的基本原理是对通信链路中各种因素加以权衡，达到一种实时的平衡，以期在不浪费功率或不牺牲系统误码率的前提下，获得更高的频谱效率(bps／Hz)。目前已有大量文献研究了OFDM系统中自适应技术，但绝大多数是以理论上的性能为研究目标，而忽视了其计算有效性和可实现性。本文将计算有效性作为算法的一个重要的衡量标准，对一些经典的分配算法进行了计算量的研究，并提出了一些计算量较低、性能较好的自适应分配算法。本文的工作可以为算法的实现提供一定的理论依据。 本文首先选取了单用户系统和多用户系统中一些经典的算法进行研究，分析了它们的计算复杂度。选择的标准有二：一是选择理论上的最优算法，以便得到具有指导意义的结果；二是选择计算复杂度较低、易实现的算法，以便得到能实际使用的结果。在算法分析的基础上，本文对这些经典算法的计算量进行了仿真，并给出了在目前主流DSP上实现这些算法所需的时间计算量，为算法实现提供了一定的参考依据。 其次，本文提出了一种单用户系统中基于注水原理的子载波分组的自适应分配算法，并对其性能进行了理论推导和仿真分析。该算法根据相邻子信道之间的相关性，对子载波进行分组，在尽量减小容量损失的前提下，降低算法的运算复杂度。这项工作可以为算法实现中，容量和计算量之间的折衷问题提供理论依据。 本文还将子载波分组算法应用于多用户系统中，提出了一种多用户系统中子载波分组的自适应分配算法，并对其性能进行了分析。该算法根据相邻子信道之间的相关性，对子载波进行分组，并以组为单位将子载波分配给各用户，优化目标是在保证每个用户的服务质量(QoS)的前提下，使系统总的发送功率最小化。如果系统总的发送功率固定，那么优化目标是使系统支持的用户数最大化。该算法可以使自适应分配算法的计算量大大降低，同时也减小了传送比特分配向量所需的开销。 另外，本文对利用多用户分集效应进行的等功率分配算法进行了研究。在给定一个子载波最大能携带的比特数以后，本文推导了能够找到携带最大比特数的子载波的概率公式，并提出了两种等功率分配算法的实现流程，它们分别优化用户数和用户速率。最后给出了一些性能和计算量的仿真结果。 本文在最后一部分讨论了非理想信道条件对自适应分配性能的影响，并总结了一些在非理想信道条件下，自适应分配技术需采取的措施。更多还原

【Abstract】 Adaptive technique is an effective scheme to combat the fading channel in Orthogonal Frequency Division Multiplexing (OFDM) systems. The basic premise of adaptive techniques is a real-time balancing of the link budget in flat fading through adaptive variation of several parameters. Thus, without wasting power or sacrificing BER, the system can obtained a higher average link spectral efficiency (bps/Hz). A great amount of literature has been presented to demonstrate the results of various adaptive schemes, and most of them have focused on the improvement of performance theoretically, while ignoring the computational complexity and feasibility of the adaptive algorithms. Regarding them as the important parameters, this thesis has analyzed the computational complexity and feasibility of some typical adaptive allocation algorithms, and provided some adaptive algorithms with lower computational complexity and better performance. The results of this work can be referenced for employing adaptive allocation algorithms in practical systems.In this thesis, we first choose some typical adaptive allocation algorithms of single-user or multiple-user system, and analyze their computational complexity. These chosen algorithms must be either the optimal algorithms, which have theoretical significance, or the algorithms with lower computational complexity, which can be easily realized. The simulation results of the computational complexity of the chosen algorithms have been presented, and the time they cost in current DSP are calculated. This work can be used as a reference for employing adaptive allocation algorithms in OFDM systems.In order to reduce the computational complexity further, we provide a new adaptive allocation algorithm with blocked subcarriers for single-user systems, which is based on the water-filling theorem. The theoretical derivation and simulation analyses for its performance have been given. The subcarriers have been divided into blocks according to the correlation between the adjacent subcarriers and those subcarriers in a block use the same modulation mode. Thus, the amount of computation is decreased premising an acceptable tradeoff between the computational complexity and the cost of capacity.Then, we employ the subcarrier-blocked scheme into multi-user OFDM systems and analyze its performance. The subcarriers have been divided into blocks according to the correlation between the adjacent subcarriers and allocated to different user in block. The optimization target is to minimize the total transmit power while guaranteeing the quality of service for each user. If the total transmit power is fixed, then the optimization target is to maximize the number of user the system can support. The simulation results show that this algorithm can reduce the amount of computation greatly; meanwhile, the overhead to transfer the bit allocation vector is less.In this paper, equal-power-allocating algorithm has also been studied taking advantage of multiuser diversity. Given a maximal number of bit that a subcarrier can carry, we derivate a formula to demonstrate the probability of finding at least one suitable subchannel, in which the maximal number of bit can be transferred, among the same subchannels of multiple users. We also provide two kinds of scheme to realize the equal-power-allocating algorithm, optimizing the number of users and the users’ data rate respectively. At last, some simulation results are presented.In the last chapter, the effects of imperfect channel information on the performance of the adaptive allocation have been discussed, and some combating measures have been summarized.更多还原