OFDM系统的智能决策引擎研究

发布时间：2018-02-09 20:20

本文关键词： 认知无线电资源分配智能决策引擎遗传算法案例推理　出处：《电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：智能决策引擎是认知无线电(Cognitive Radio,CR)系统的核心模块,通过自身的优化决策和学习推理功能,实现系统资源的最佳配置。OFDM(Orthogonal Frequency Division Multiplexing)技术是一种高效的多载波调制技术,已广泛应用到当前主流的无线通信技术中,适合应用于认知无线电系统。本文从人工智能技术在OFDM系统资源分配应用的角度出发,研究了认知无线电智能决策引擎系统的设计。智能决策引擎能够根据感知的外界信息,结合通信业务要求、历史经验以及约束规则,利用内部的优化决策和学习推理模块进行智能决策,自适应地调整和配置系统资源。本文的研究主要分为以下四个部分:本文第一部分主要研究OFDM系统资源分配以及人工智能算法。分析了OFDM系统动态资源分配的优化准则:边界自适应(Margin Adaptive,MA)和速率自适应(Rate Adaptive,RA)准则;同时研究了四种优化决策算法和四种学习推理算法的工作原理及应用,并对这些优化算法和学习算法进行对比分析,为后续的资源分配和智能决策引擎的研究提供理论指导。本文第二部分研究了以最小化发射功率为目标的OFDM系统资源分配算法。在分析遗传操作算子的不同取值对遗传算法搜索性能影响的基础上,提出了一种基于改进遗传算法、以最小化系统发射功率为目标的资源分配算法,然后分别使用TDMA-OFDM、FDMA-OFDM、多用户贪婪注水算法、基本遗传算法和改进遗传算法进行子载波、比特和功率分配,并对这几种算法的性能和复杂度进行了对比分析。本文第三部分研究了以最大化容量为目标的OFDM系统资源分配方法。在研究最优以及次优子信道和功率分配算法原理的基础上,提出了一种基于遗传算法、最大化系统容量为目标的子信道和功率联合分配算法。重点研究了在满足系统目标(最大化系统容量和用户传输速率比例公平性要求)和约束条件(发射功率和误码率约束)下,遗传算法染色体、目标函数和评价函数的设计,并仿真验证了多目标优化问题中不同加权因子取值对遗传算法搜索性能的影响;对两种次优算法、遗传算法和改进遗传算法的性能进行了综合对比分析。本文第四部分主要研究了基于遗传算法和案例推理的OFDM系统智能决策引擎的设计。给出了智能决策引擎优化模块和学习模块的协同工作机制,并通过几个实例场景仿真验证了系统的运行机理。智能决策引擎在满足系统通信目标和约束条件的情况下,根据检测的信道信息,采用遗传算法和案例推理进行智能决策,自适应地调整系统的子载波、比特、发射功率和调度周期等配置参数,来适应外界环境的变化。
[Abstract]:Intelligent decision engine is the core module of cognitive radio cognitive radio (CR) system. The optimal configuration of system resources. OFDM orthogonal Frequency Division Multiplexing (OFDM) is an efficient multicarrier modulation technology. Has been widely used in the current mainstream wireless communication technology, suitable for the cognitive radio system. This article from the artificial intelligence technology in the OFDM system resource allocation application point of view, In this paper, the design of cognitive radio intelligent decision engine system is studied. The intelligent decision engine can combine the requirements of communication service, historical experience and constraint rules according to the perceived external information. Using the internal optimization decision and learning reasoning module to make intelligent decision, The research of this paper is divided into four parts: the first part of this paper mainly studies OFDM system resource allocation and artificial intelligence algorithm, and analyzes the dynamic resource allocation of OFDM system. The criteria are boundary adaptive margin adaptive MAand rate adaptive rate adaptive rama; At the same time, the working principle and application of four optimization decision algorithms and four learning reasoning algorithms are studied, and these optimization algorithms and learning algorithms are compared and analyzed. In the second part of this paper, we study the resource allocation algorithm of OFDM system with the aim of minimizing the transmit power, and analyze the different values of genetic operators. On the basis of the effect on the search performance of genetic algorithm, This paper presents a resource allocation algorithm based on improved genetic algorithm, which aims at minimizing the transmission power of the system, and then uses TDMA-OFDM FDMA-OFDM, multi-user greedy water flooding algorithm, basic genetic algorithm and improved genetic algorithm to carry out subcarriers, respectively. Bit and power allocation, The performance and complexity of these algorithms are compared and analyzed. In the third part of this paper, the resource allocation method of OFDM system aiming at maximizing capacity is studied. On the basis of studying the principle of optimal and suboptimal subchannel and power allocation algorithm, A genetic algorithm is proposed. A joint subchannel and power allocation algorithm with maximum system capacity as its target. The emphasis is placed on the study of system objectives (maximization of system capacity and user transmission rate proportional fairness requirements) and constraints (transmit power sum). Error rate constraint, The design of chromosome, objective function and evaluation function of genetic algorithm, and the simulation results show the influence of different weighting factors on the search performance of genetic algorithm. The performance of genetic algorithm and improved genetic algorithm are compared and analyzed synthetically. In the 4th part of this paper, the design of intelligent decision engine of OFDM system based on genetic algorithm and case-based reasoning is studied, and the intelligent decision engine is given. The cooperative working mechanism of the module and the learning module, The mechanism of the system is verified by the simulation of several examples. The intelligent decision engine adopts genetic algorithm and case-based reasoning to make intelligent decision according to the detected channel information and meets the system communication objectives and constraints. The configuration parameters of the system such as subcarrier bit transmit power and scheduling period are adaptively adjusted to adapt to the change of the external environment.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN929.53

【相似文献】