基于QoS的LTE下行链路调度算法研究

发布时间：2018-06-16 16:33

本文选题：自适应资源分配 + 服务质量　；参考：《兰州交通大学》2017年硕士论文

【摘要】：无线多媒体业务的快速发展,要求LTE系统具有高吞吐量、低时延和满足用户服务质量要求(QoS)的特性。这就使得有限的频带资源与不断增长的数据业务需求之间的矛盾日益严重。OFDM作为LTE下行链路的核心技术之一,能够显著提高频带利用率和数据传输速率,使系统能更加灵活的调度资源。因此在OFDM技术的基础上,如何合理高效地调度资源并兼顾用户QoS是一个需要研究的问题。论文首先介绍了LTE系统架构、QoS机制以及LTE下行关键技术—OFDM技术原理、系统实现,分析了无线信道衰落特性,为后续建模仿真提供了理论支持。而后引入了两个资源优化分配准则,即速率自适应准则RA和余量自适应准则MA。对这两种准则下的一些资源分配算法进行了理论分析和性能仿真,其中包括单用户资源分配的注水算法、贪婪算法,多用户资源分配的MaxSNR+WPA算法、RR+EPA算法、Shen算法。在分析各个算法原理和优缺点的基础上,发现现有LTE下行链路资源分配仍有许多待解决的问题,如用户间公平性较差、无法保证QoS、效率低等。重点研究了在LTE下行多用户环境下,基于QoS的自适应资源分配算法。针对这个研究方向,首先基于RA准则将资源分配分为子载波分配与功率分配。在子载波分配过程中,首先依照RR算法为每个用户分配一个子载波,再结合用户目前信道状态信息、业务类型、速率约束因子分配剩余子载波,速率约束因子可根据用户业务类型动态调整,其决定了用户速率的上限;在功率分配过程中,由于已经完成了子载波的分配,此时RA优化问题可放松为一个凸优化问题,结合已完成的子载波分配方案剔除增益噪声比过小的子载波并以此作为一个约束条件简化该优化问题,化简至只有一个未知变量的情况下利用二分搜索迭代算法迭代出该问题的最优解,再利用该最优解求出系统中每个用户的动态注水线,此时采用注水思想完成功率分配。最后通过MATLAB仿真工具,从基站、用户、噪声、路径损耗等几个方面进行建模,结合城市宏小区(Uma)场景参数配置以及3GPP 36.942协议规范设置相应仿真参数,搭建静态LTE系统级仿真环境。在小区吞吐量、用户SINR、单一业务场景下的用户容量、混合业务场景下的用户容量这几个方面对提出算法进行了仿真分析。结果表明,提出算法能够在用户公平性与系统吞吐量之间进行平衡,资源分配更加灵活,可通过设置不同比例公平因子来控制用户传输速率上限,在不同业务类型下实现了差异化资源调度,分别从用户间公平性、用户通信链路质量、不同业务传输速率保证等几个方面满足了用户的服务质量(QoS)要求,同时具有较快的收敛速度。
[Abstract]:The rapid development of wireless multimedia services requires that LTE systems have the characteristics of high throughput, low delay and QoS. As one of the core technologies of LTE downlink, OFDM can significantly improve the bandwidth efficiency and data transmission rate, which makes the contradiction between the limited frequency band resources and the increasing demand for data services increasingly serious, as one of the core technologies of LTE downlink. Make the system more flexible scheduling resources. Therefore, on the basis of OFDM technology, how to reasonably and efficiently schedule resources and take account of user QoS is a problem that needs to be studied. Firstly, this paper introduces the QoS mechanism of LTE system architecture and the principle and implementation of OFDM technology, which is the key technology of LTE downlink. The characteristics of wireless channel fading are analyzed, which provides theoretical support for subsequent modeling and simulation. Then two optimal resource allocation criteria are introduced, that is, rate adaptive criterion RA and residual adaptive criterion MA. Some resource allocation algorithms under these two criteria are theoretically analyzed and simulated, including water injection algorithm for single user resource allocation, greedy algorithm, MaxSNR WPA algorithm for multi-user resource allocation, RR EPA algorithm and Shen algorithm for multi-user resource allocation. On the basis of analyzing the principle of each algorithm and its advantages and disadvantages, it is found that there are still many problems to be solved in LTE downlink resource allocation, such as poor fairness among users, poor QoS, low efficiency and so on. The adaptive resource allocation algorithm based on QoS in LTE downlink multiuser environment is studied. In this research direction, resource allocation is divided into subcarrier allocation and power allocation based on RA criterion. In the subcarrier allocation process, one sub-carrier is assigned to each user according to RR algorithm, and then the remaining sub-carrier is allocated according to the current channel state information, service type, rate constraint factor, etc. The rate constraint factor can be dynamically adjusted according to the user service type, which determines the upper limit of the user rate. In the power allocation process, the RA optimization problem can be relaxed as a convex optimization problem because the sub-carrier allocation has been completed. Combined with the completed subcarrier allocation scheme, the subcarriers with too small gain and noise ratio are eliminated and used as a constraint to simplify the optimization problem. In the case of only one unknown variable, the optimal solution of the problem is iterated out by using the binary search iterative algorithm, and then the dynamic water injection line of each user in the system is obtained by using the optimal solution, and the power distribution is accomplished by using the water injection idea. Finally, through MATLAB simulation tools, from the base station, user, noise, path loss and other aspects of modeling, combined with the urban macro cell Uma) scene parameters configuration and 3GPP 36.942 protocol specification to set the corresponding simulation parameters. Build static LTE system level simulation environment. The proposed algorithm is simulated and analyzed in the aspects of cell throughput, user SINR, user capacity in single service scenario and user capacity in mixed service scenario. The results show that the proposed algorithm can balance user fairness and system throughput, and the allocation of resources is more flexible, and the upper limit of user transmission rate can be controlled by setting different proportional fairness factors. The differentiated resource scheduling is realized under different service types, which satisfies the QoS requirements of users from the aspects of user fairness, user communication link quality, different service transmission rate assurance and so on. At the same time, the convergence rate is faster.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN929.5

【参考文献】