Sub-Nyquist spectrum sensing and learning challenge-猫眼课题宝

DOI：

10.1007/s11704-021-1275-y

发表时间：

2021-06

期刊：

Frontiers of Computer Science

影响因子：

4.2

通讯作者：

中科院分区：

计算机科学3区

文献类型：

作者：

Yue Gao;Zihang Song;Han Zhang;Sean Fuller;A. Lambert;Z. Ying;P. Mähönen;Yonina C. Eldar;Shuguang Cui;M. Plumbley;C. Parini;A. Nallanathan

频谱资源在某些频带中未被充分利用的事实已经激发了动态频谱接入（DSA）方法，该方法使得配备有认知无线电（CR）设备的未授权次级用户（SU）能够接入频谱而不会对初级用户（PU）造成显著干扰。目前，毫米波和太赫兹频段无线通信带宽的不断增加对DSA的主要实现手段频谱感知技术的性能提出了更高的要求。由于高速模数转换器（ADC）的高功耗、高成本和硬件复杂性，传统的奈奎斯特速率采样和处理往往不切实际。为了克服采样率瓶颈，已经提出了几种亚奈奎斯特采样方法[1-9]、恢复算法[10-18]和信道检测方法[1 - 9 -23]。此外，基于机器学习的频谱感知的最新进展已经得到了表征，这为CR设备提供了进一步的智能，在复杂的无线电环境下具有更好的适应性和更高的灵活性[24-30]。然而，对亚奈奎斯特频谱感测的性能要求产生了许多不同的挑战，这些挑战包括但不限于以下内容：·对于压缩采样器，成功重构稀疏信号所需的采样率由信号的实际稀疏度阶数（占用信道与总感测带宽的比率）确定。另一方面，基于贪婪算法的频谱重构需要频谱稀疏度的先验知识作为输入。然而，由于环境的不确定性，

The fact that the spectrum resource is underutilised in certain bands has motivated the dynamic spectrum access (DSA) approach, which enables unlicensed secondary users (SUs) equipped with cognitive radio (CR) devices to access the spectrum without causing significant interference to primary users (PUs). Nowadays, the increasing bandwidth for wireless communication in millimetre-wave and Terahertz frequency bands puts higher requirements on the performance of spectrum sensing technique, the primary enabler of DSA. Traditional Nyquist-rate sampling and processing tend to be impractical due to high power consumption, high-cost, and hardware complexity of high-speed analogue to digital converters (ADCs). To overcome the sampling rate bottleneck, several sub-Nyquist sampling methods [1–9], recovery algorithms [10–18] and channel detection methods [19–23] have been proposed. Moreover, the recent advancements in machine-learning-based spectrum sensing have been characterised, which has provided further intelligence to CR devices with better adaptivity and higher flexibility under complex radio environments [24–30]. Still, the performance demands placed on sub-Nyquist spectrum sensing creates many different challenges, which comprise, but are not limited to, the following:• For compressive samplers, the necessary sampling rate to successfully reconstruct a sparse signal is determined by the actual sparsity order (the ratio of the occupied channel to the total sensing bandwidth) of the signal. On the other hand, spectrum reconstruction based on a greedy algorithm requires prior knowledge of spectrum sparsity as an input. However, due to the uncertainty in the environ-

Sub-Nyquist spectrum sensing and learning challenge

文献摘要

被引文献

相似文献