基于人工神经网络的认知无线电频谱感知研究

认知无线电是一种新的智能无线通信技术，通过感知周围环境特征，并采用构建理解的方法进行学习。频谱感知技术是认知无线电系统设计的重要组成部分，本文提出了基于人工神经网络（ANN）算法的认知无线电感知技术。由于单用户检测存在检测精度不高的问题，本文通过采用ANN的自主学习能力设计频谱检测分类器，对以往分类器所收集的环境信息进行学习和储存，从而对以后感知新的环境时起到先验信息的作用，能快速准确地判决出主用户的存在与否。通过与能量检测和循环平稳检测对不同调制方式信号进行信号识别仿真实验可以看出，相比原有的单用户频谱检测技术，ANN具有更高的有效性和可靠性。

Neural network based on spectrum sensing in cognitive radio

in recent years, cognitive radio is a new intelligent wireless communication technology, its approach learns the perception of surrounding environment characteristics by constructing the comprehension. Spectrum sensing is an important part in cognitive radio system, this paper presents a artificial neural network ( ANN ) algorithm used in cognitive radio sensing technology. Due to single user detection`s low precision,this paper adopts the ANN`s autonomous learning ability to design the spectrum detection classifer,it can learn and storn the previous collection of environmental information as a role of prior information. It can be fast and accurately judge the presense of the user. In part 4 in the contrast with the energy detection and Cyclostationary detection in different modulation signals recognition simulation experiment , compared with the original single-user spectrum detection technology, ANN has high availability and reliability.