面向能效优先的宽带协作频谱感知关键技术研究

Spectrum sensing in cognitive radio has been researched widely due to its promising capability in utilizing the spectrum resources of less active systems so as to solve the spectrum shortage problem in public mobile communication systems. However, the power efficiency received less consideration in current research, while power efficiency is another very important factor which needs to be considered seriously to satisfy the requirements of green network. By considering this, this project makes further research and achieves optimization on the wideband cooperative spectrum sensing algorithms, spectrum access-sending strategy and inter-node sensing information transmission scheme by introducing the philosophy of energy-efficient design. The research of this project includes: the modeling of energy-efficiency distribution of mobile terminals and evaluation criteria of energy-efficiency, energy-efficient wideband cooperative sensing algorithms, the selection of both sensing order and sensing-access methodology for energy-efficient wideband spectrum sensing, and inter-node transmission scheme of sensing information. Our research is believed to be able to lower the terminal energy consumption while satisfying all the sensing-ability requirements, so as to construct an energy-efficient green radio network, improve the spectrum efficiency and promote end-user perception of the network. Three scientific issues, namely, the modeling of energy-efficiency of spectrum sensing, the design of energy-efficient cost function, and the mapping between sensing information transmission scheme and its power consumption have been solved, by means of technical solutions including location estimation of primary users, clustered cooperation of secondary users, sequantial sensing sampling, sequantial judgement of sensing stage, and rateless channel coding. The main innovations of this project are as follows: the evaluation criteria and the mathematical model of terminal energy-efficiency, the construction of hybrid cost function considering both opportunistic throughput and energy efficiency, the design of the energy-efficient sensing-access strategies and selection of sensing order as a sequential decision making problem so as to achieve optimal solution, and the energy-efficient inter-node transmission scheme of sensing information based on fountain code.

针对认知无线电频谱感知技术在终端能效方面考虑较少，无法满足建设绿色通信网络要求的问题，引入能效优先思想，提出对宽带协作频谱感知的三个核心环节（感知判决、感知策略、感知信息传输）进行优化设计的方法。建立终端能效模型和评价指标，研究能效优先的宽带协作频谱感知算法、低能耗协作频谱感知策略和协作感知传输方案等。采用主用户位置估计、感知用户分簇协作、序贯压缩采样、感知阶段序列决策和无码率信道编码等技术途径，解决频谱感知能效评价模型构建、能效优先认知代价函数设计、感知传输机制与能耗换算关系三个科学问题。在保证感知性能前提下，有效降低终端能耗，以达到提高频带利用率、建设能效优先的绿色网络和提升终端用户体验的目标。具有终端能效评价的数学模型、将能效准则引入感知判决构造出兼顾机会吞吐量和能效的复合认知代价函数、面向能效优先的感知顺序和接入策略联合优化策略和基于喷泉码的高能效感知信息传输方案等创新点。

根据建设绿色通信网络的目标，针对认知无线电宽带频谱感知技术在降低能耗方面的问题展开研究，主要研究内容包括终端能耗的成因与能效评价指标研究、能效优先的宽带协作频谱感知算法研究、宽带协作频谱感知的低能耗感知策略研究和高能效宽带协作频谱感知的传输方案研究等。在项目的资助下，已发表学术论文16篇，录用3篇；申请国家发明专利13项，授权4项；培养硕士研究生12名。全面超出项目的预期成果。关键研究成果及意义包括：. 1. 构建了感知能耗优化问题的模型，并得出相关参数的优化值：提出了基于感知时间最优化的高能效频谱感知方法，通过求解此优化问题获得满足特定感知性能且能耗最低的最优感知时间闭式表达式；提出基于判决门限最优化的高能效频谱感知方法，通过求解该优化问题得到满足特定误判概率下的最小感知能耗。. 2. 基于压缩感知理论，提出一系列能效优先的宽带协作频谱感知算法：提出采用非重构的方式减少观测数目并降低终端能耗的方法；提出了基于最大似然估计检测算法，并且推导出了检测性能界；提出分布式协作序贯压缩的随机信号检测算法，降低了能耗并提高了检测的实时性；提出了一种可变步长序贯压缩频谱快速感知算法，该算法建立了下一次判决所需观测值数目步长因子与当前对数似然比距门限距离之间的函数关系。. 3. 建立了宽带协作频谱感知的低能耗感知策略：提出了以概率密度函数的变化情况作为判断主用户是否出现的依据的高能效频谱感知方法；针对频谱感知性能不稳定的问题，应用控制理论，提出提高系统误码率稳定性的策略；将粗糙集理论引入频谱感知决策系统，提出基于发射功率最小的决策准则进行约简的决策规则，较好地节省能量。. 4. 研究了高能效宽带协作频谱感知的传输方案：提出采用分簇及信道编码方式的感知信息传输方案；提出通过重要性抽样和梯度算法的优化设计对喷泉码的编解码算法进行改进，进一步降低发射功率；基于改进的线性注水算法和人工智能算法，提出能够最大化频谱效率的资源分配算法。