动态不确定性条件下的认知无线电关键技术研究

Cognitive radio (CR) is considered as an emerging wireless communication framework which is of great promise to alleviate the spectrum scarcity. Nevertheless, the existing fundamental theory and key technologies in CR are essentially premised on a simplified quasi-static channel modeling. As a consequence, it can hardly deal with the dynamic time-variations and information uncertainty, which may be ubiquitous in practical applications. This research mainly focuses on the development of a promising CR framework in the presence of dynamic time-variations and uncertainty. A novel dynamic state-space model will be formulated to characterize the spectrum sensing in time-varying fading channels. By resorting to the Bayesian sequentially statistical inference and the random sampling theorem, the real-time joint estimations of time-varying channels and the working state of primary users may be obtained. On this basis, the sensing schedule strategy with an more efficient self-adaptation will be investigated. Meanwhile, by taking the time-variations and uncertainty into full considerations, a more realistic constrained optimization model is presented, and furthermore, an efficient cognitive resources allocation and schedule scheme will be proposed. This research will suggest a more flexible and efficient CR framework, which can address the practical difficulties posed by the complex realistic scenarios. It is supposed to this new framework may have significance both in theory and practice, which is also of great importance to promote the large-scale commercial progress of CR.

认知无线电技术是一种有望缓解现有频谱资源紧张局势的新兴无线通信技术。现有认知无线电基础理论与关键技术研究主要基于一种简化的准静态信道模型，因而难以有效地应对实际应用中普遍存在的复杂动态时变特性与信息不确定性。本课题重点研究一种面向未来实际应用的动态时变性下的新型认知无线电系统架构。通过建立一种新颖的动态状态空间模型，基于贝叶斯序贯检测与随机采样理论实现了时变信道与授权用户状态的联合实时估计；在此基础上，深入研究动态时变特性下的自适应认识无线电感知调度机制；同时充分考虑了无线信道的动态时变特性，建立一种更为有效的约束优化模型，提出一种能应对动态时变特性的高效能认知系统资源分配与调度方案。本课题将建立一种更高层面的认知无线电系统架构，旨在解决该技术应用至复杂实际场景中的所面临的诸多实现难题，对于推动认知无线电系统的实用化进程，将具有极其重要的理论与实践意义。

课题组根据认知无线电研究的最新发展，开展了以下几个方面的相关研究，研究内容及取得的成果介绍如下：.1）.设计了动态时变信道下频谱检测方案，通过建立一种新型的动态状态空间模型对时变衰落信道下的频谱感知问题进行有效建模。.2）.基于动态状态空间模型，设计一种新颖的联合频谱感知算法。该算法利用接收端观测信号，对时变信道增益与授权用户状态进行联合估计。.3）.分析时变多径信道下的实时信道估计与授权用户状态估计方案的可行性并评估其实际性能。.4）.对噪声方差不确定和时频双选性衰落信道等多种复杂无线环境下的频谱感知技术进行研究，基于随机有限集提出了认知无线电频谱感知的一个新的感知框架，即深度感知理论。.5）.研究了OFDM系统的动态时变信道，提出一种新型的优化算法。.6）.提供了一种时空两维的信道频谱接入方法，对频谱资源进行优化分配。.7）.提出了一种新环境下的感知策略，该策略能够在时变多路径平变信道中实现对信号的盲感知。利用两种不同的隐性Markov模型来抽象主用户状态和时变多路径的平变通道增益并进行实际感知。.8）.提出了时变信道下频谱感知问题的动态状态空间模型（DSM）。主用户（PU）状态和衰落信道被表征为隐藏状态。在此基础上，提出了一种新的实时估计PU状态和衰落信道增益的方法。.9）.针对信道衰减系数随着时间变化这一问题，本项目组在研究过程中提出了一种新的分布式应用频谱感知算法。.10）提出了一种新的认知无线电网络架构（CRN），用于时变平坦衰落（TVFF）信道中的动态频谱共享。