基于卷积神经网络的宽带合作频谱感知

利用认知无线电网络中的多个次用户所能提供的大量频谱观测数据，提出一种基于卷积神经网络（convolutional neural network，CNN）的宽带合作频谱感知方案。宽带频谱感知旨在灵活地检测跟踪目标宽带授权频段上的可供使用的频谱空穴，该方案考虑利用宽带频谱上被占用的子带与未被占用的子带之间在信号能量及占用位置方面所体现出的类别差异，通过设计一种CNN模型并基于此对频谱观测数据进行训练学习，得到频带占用模式分类模型，从而实现宽带合作频谱感知。仿真结果证明，与传统的基于能量检测技术和典型机器学习（machine learning，ML）分类算法的宽带合作频谱感知方案相比，该方案在检测性能上具有较大的优势，特别是在低信噪比环境下的检测性能。

CNN based wideband cooperative spectrum sensing algorithm

Using the large amount of spectrum observation data provided by multiple secondary users in the cognitive radio network, we propose a frequency band occupancy patterns identification scheme based on the convolutional neural network (CNN) for wideband cooperative spectrum sensing. Wideband spectrum sensing aims to flexibly detect and track the available spectrum holes on the target band. This scheme considers the continuity of the PU''s occupancy in the wide licensed frequency band, and then we design a convolutional neural network model, and spectrum observation data is trained and learned based on this model. Finally, the trained classification model is used to detect the unknown signal to identify the occupancy pattern, thereby realizing broadband spectrum sensing. Theoretical derivation and simulation results prove that the proposed scheme has greater advantages in detection performance than the traditional wideband cooperative spectrum sensing scheme based on energy detection technology and typical machine learning (ML) classification algorithms. Especially in low signal-to-noise ratio, the effect is more prominent in the environment.