基于频谱使用特征的动态频谱共享系统性能分析及评估方法研究

Cognitive Radio is the key enabling technology in the future wireless communication network, also known as dynamic spectrum sharing(DSS). In a DSS communication system, secondary users(SUs) opportunistically access the white space spetrum that is not occupied by the primary user(PU). The fundmental question faced in applying DSS is how to analyze and evaluate the performance of the dynamic spetrum sharing communication system. Current research indicates that the main obstacle of solving this question is lack of valid and effcient model for spectrum occupancy behavior of the primary wirless system,as the performance of the DSS communication system is mainly decided by the spectum occupancy characteristics of the primary system..Some analytical performance models have been proposed based on Markov chain modeling in the literature. In these models, SUs and PUs access the spectrum by randomly selecting the available channels with equal probability. But actually, how SUs and PUs use the spectrum are controlled by the access strategy designed in their MAC layer in an ad-hoc network, or centralized radio resource management (RRM) layer in an infrastructure-based network. In this project, the spectrum occupancy characteristics is proposed to be modeled by a state transion probability matrix which consided the traffic characteristics and access rules of the wireless system. The matrix is also used to analyze the performance of DSS system based on every-channel-state Markov chain which is more powerful than current analysis model. The performance metrics, as blocking probability, dropping probability, collision probablility, carrieied traffic etc, can be calculated by this method. .To verify the proposed model, three typical wireless communication system, GSM, WLAN and TV broadcasting, will be measured and modeled by the hidden Markov chain in different scenes. Moreover we will give the framework of the scenario-driving evaluation method based on the measured models.

动态频谱资源共享是无线通信的战略发展方向之一，对其性能进行客观、合理有效的分析和评估是其应用的重要基础。研究显示：由于目前对无线通信系统的频谱使用行为缺乏理论分析和建模，使得客观评价和分析动态频谱资源共享系统（在授权系统频谱使用行为约束下工作）的性能非常困难。本项目考虑系统外在业务特性和系统内在决策行为（如接入策略、调度算法等），利用状态转移速率矩阵描述无线通信系统频谱使用特征的时变性，并以此为基础，通过细化的每信道状态表示的连续MarKov链分析，获得动态频谱资源共享系统的理论性能；通过对三种典型系统（广播电视系统、GSM蜂窝系统、无线局域网）在网络侧和空中接口侧的测量和隐Markov链参数分析，提出三种典型授权系统的频谱使用特征模型；并在此基础上，研究基于策略驱动的动态频谱资源共享系统性能评估方法。

动态频谱资源共享是无线通信的战略发展方向之一，对其性能进行客观、合理有效的分析和评估是其应用的重要基础。本项目在深入研究无线通信系统频谱使用特征机理的基础上，提出了融合接入策略的动态频谱共享系统性能分析数学模型，通过马尔科夫链的分析方法，可以获得主用户阻塞概率、次级用户阻塞概率、次级用户中断概率、次级用户切换概率、次级用户平均切换次数、信道利用率等关键性能指标。通过空中接口对三种典型授权系统（广播电视系统、GSM蜂窝移动通信系统和无线局域网）频段进行大量实际测量，采用隐马尔科夫过程的分析方法对授权系统进行建模，估计描述频谱使用特征的关键参数集。在此基础上，针对典型的认知中继网络的应用，推导出不同中继节点选择方案下精确中断概率、误符号率下界和各态历经容量上界的闭合表达式。在对授权系统频谱使用特征建模的基础上，开发了基于OPNET的可扩展动态频谱共享系统仿真平台，验证了动态频谱资源共享系统性能分析、授权系统频谱使用特征分析的合理性。