摘要
通过比较几种常用调制分类器的优缺点,提出了一种新的调制分类器设计方案。它采用高效稳健的遗传算法,先筛选出一高质量的最小特征子集,再对RBF神经网络的隐层节点中心值进行进化优选。仿真结果证实了该方法的优越性。此外,还对调制分类器用于移动数字电视的研究方向提出了建议。
A new design of modulation classifier is proposed based on comparisons to the advantages and disadvantages of some commonly used modulation classifiers. Firstly, an efficient and robust genetic algorithm is employed to sift out qualified and the least feature subset. Secondly, genetic algorithm is used to optimize the centers of the hidden units of RBF networks. The simulation results testify the effectiveness of the proposed method. Finally, the application of modulation classifier in mobile digital TV is suggested.
出处
《电视技术》
北大核心
2006年第2期27-29,37,共4页
Video Engineering
基金
总装备部重点研究项目基金资助(编号:6130320)
关键词
调制分类器
遗传算法
神经网络
移动数字电视
modulation classifier
genetic algorithm
neural network
mobile digital TV
