【作者】 张国斌；

【导师】 冯穗力；

【作者基本信息】 华南理工大学 ， 通信与信息系统， 2011， 博士

【摘要】 随着无线通信技术的飞速发展,频谱资源变得越来越紧张。传统的静态的无线频谱管理方式使得部分频谱通常处于空闲状态,限制了频谱的使用效率。认知无线电(Cognitive Radio)技术被认为是有效利用空闲频谱、缓解无线频谱紧缺问题的一种新方法。认知无线电网络能够感知无线频谱环境、自动搜寻可用频谱,为提高频谱利用率开辟了崭新的途径。认知无线电系统基于动态频谱接入方式,资源管理和分配是认知无线电系统的关键技术和研究热点。本文对认知无线电系统的频谱分配、功率控制和接入控制进行了深入研究,主要创新性成果如下:(1)基于图论着色模型研究认知无线电非连续正交频分复用系统的子载波分配问题。本文对现有的认知无线电系统的图论模型进行改进,建立了适用于认知无线电NC-OFDM系统的子载波分配模型。在此模型下,将随机算法和分布式贪婪算法应用于认知无线电NC-OFDM系统的子载波分配中,同时针对这两种算法的缺点,将多用户OFDM系统的最小容量最大化算法引入认知无线电NC-OFDM系统中,提出适用于认知无线电NC-OFDM系统的最小容量最大化算法,该算法很好的兼顾了系统的吞吐量性能和用户公平性。(2)基于干扰温度模型研究认知无线电系统的功率分配问题。现有基于干扰温度模型的功率控制问题的求解方法存在不能得到最优解或计算量较大的问题。本文提出一种适用于上行认知无线电OFDM系统的功率分配算法,计算复杂度较低、并且能够得到最大化用户传输速率的最优功率分配。该算法在干扰温度和认知用户总功率限制条件下,反复使用置零迭代注水算法,直至获得最优功率分配。数学推导证明算法具有最优性,仿真结果验证了该结论,同时表明该算法的计算量比梯度方法减少50%以上。(3)研究非实时业务条件下认知无线电系统的动态资源分配问题。由于认知无线电系统可用频谱状态的时变性,静态分配方法并不能实现频谱的高效利用。本文采用POMDP理论模型对认知无线电系统的子载波状态进行建模,然后基于两种共享方式研究系统的动态资源分配。在填充式共享中,针对可用子载波的不断变化,将空闲子载波动态分配给认知用户,通过控制各认知用户在各子载波上的接入概率最大化系统的吞吐量并保证用户的QoS需求。仿真结果表明,动态子载波分配算法的吞吐量性能优于静态分配方法和CSMA接入方法。在下垫式共享中,首先对功率进行动态分配,仅在授权用户出现时才将认知用户的功率限制在干扰功率约束之下,然后按照一定的接入概率将子载波动态分配给各认知用户,在保证用户最低速率需求的前提下最大化系统容量。仿真结果表明,子载波与功率的联合动态分配算法能够获得明显的吞吐量性能增益。(4)研究实时业务条件下认知无线电系统的接入控制问题。针对无缓冲机制和有缓冲机制两种系统类型,采用马尔可夫链对系统状态进行建模,分别对用户的阻塞概率、中断概率和系统容量进行分析,在此基础上提出一种认知用户的接入概率控制算法,在满足用户中断概率限制的前提下最大化系统容量。分析结果表明,在使用该算法进行接入概率控制的系统中,用户的接入概率与系统吞吐量随中断概率限制的增大而增大,中断概率限制下的最大接入概率即为最大化系统容量的接入概率,用户中断概率限制的满足是以牺牲系统吞吐量作为代价的。同时,结果表明,引入缓冲机制能够降低用户的中断概率、提高用户的接入概率和系统的吞吐量性能。更多还原

【Abstract】 With the rapid development of the wireless communication technology, the frequency resource becomes more and more insufficient. The traditional static wireless frequency management makes much frequency in idle state in usual, which limits the usage efficiency of the frequency resource. Cognitive radio is considered an effective method, which employs spare frequency effectively and solves the insufficiency problem of the wireless frequency. Cognitive radio network can sense the wireless frequency environment intelligently and search for available frequency automatically. Therefore cognitive radio pioneers a new way to enhance the frequency utilization efficiency.Cognitive radio system is based on dynamic frequency access. So resource allocation and management is a pivotal technology and research hotspot. The paper makes deep research in frequency allocation, power control and access control of cognitive radio system. The main innovation results are as follows:1. Research on subcarrier allocation of cognitive radio NC-OFDM system based on graph coloring model. The paper improves the current graph model based on channel allocation and builds the subcarrier allocation model adapted to cognitive radio NC-OFDM system. Based on the improved model, the paper applies the rand algorithm and the existing distribution greedy algorithm for subcarrier allocation in cognitive radio NC-OFDM system. Also the paper introduces the max-min algorithm in multiuser OFDM system and proposes the max-min algorithm adapted to cognitive radio NC-OFDM system, which can not only obtains better throughput performance but good fairness performance.2. Research on power allocation in cognitive radio system based on interference temperature model. In the existing power control based on interference temperature some problems exist. For example, some methods are too simple to attain the optimal solution, and also some have too large computation amount. Aimed at the disadvantage of the existing methods, the paper proposes an optimal power allocation algorithm adapted to uplink cognitive radio OFDM system which has low computation complexity and can maximize the user transmission rate. The algorithm applies the zero-setting iterative water-filling algorithm repeatedly until the optimal power allocation is achieved. The mathematical derivation proves the optimization performance of the algorithm, which is also verified by the simulation result. At the same time the simulation result indicates that the computation complexity of the algorithm reduces by over 50% compared with the gradient method.3. Research on the dynamic resource allocation in cognitive radio system in the condition of non real time service. Currently a lot of document researches on the static resource allocation in cognitive radio system, but because of the time and space change of the available frequency, the static allocation method can’t realize the efficient frequency utilization. The paper researches on dynamic subcarrier and power allocation in cognitive radio system in the condition of non real time service based on overlay and underlay sharing mode. Firstly we build and analyze the model of the subcarrier state based on POMDP theory. Then under the overlay and underlay mode we research on the dynamic resource allocation in cognitive radio system. Under the overlay mode we allocate the idle subcarrier to cognitive user dynamically and control the access probability of cognitive users to spare subcarrier to maximize the system throughput and guarantee the QoS (Quality of Service) requirement of users. The simulation results indicate that the throughput performance of dynamic allocation algorithm is better than static allocation and CSMA access method. Under the underlay mode, first we allocate power dynamically by limiting the cognitive user power only when the primary user arrives. Then we allocate the subcarriers dynamically according to the access probability of cognitive users in each subcarrier states to maximize the system throughput in the precondition of guaranteeing the minimal rate requirement of cognitive users. The simulation results show that the associated subcarrier and power allocation algorithm can attain obvious throughput performance plus.4. Research on access control in cognitive radio system in the condition of real time service. Aimed at the system with buffer mechanism and without buffer mechanism respectively, we build the system state model by continuous time Markov chain and analyze the block probability, forced termination probability and throughput performance respectively. Then we propose the access probability controlling algorithm to satisfy the limitation of the user forced termination probability and also maximize the system throughput. The analysis results indicate that in the system with access probability control, the access probability of users and system throughput are enlarged with the forced termination probability limitation and the largest access probability under the forced termination probability is the access probability value able to maximize the system throughput. Therefore in conclusion the satisfaction of user forced termination probability limitation sacrifices the system throughput performance. Also the results show that the introduction of the buffer mechanism can reduce the forced termination of users and enhance the access probability of users and system throughput performance.更多还原