摘要
针对现有基于聚类算法的信号调制识别在低信噪比时识别率低的缺点,文中采用聚类算法提取信号特征参数,通过变梯度Polak-Ribiere BP修正算法对神经网络进行训练,以提高收敛速度,改善在低信噪比条件下网络识别性能,实现对基于星座图调制方式信号的调制识别,仿真结果表明,在低信噪比条件下,调制识别率和单独采用聚类算法或基于BP算法的神经网络识别时比较提高30%以上,在信噪比为4d B条件下识别率可达到90%,且系统易于实现,在信号调制识别中具有广泛的应用前景。
To improve the recognition rate of the signal,a modulation recognition method is proposed based on the clustering algorithm under the low SNR.The characteristic parameter of the signal is extrac-ted by using a clustering algorithm,neural network is trained by using the algorithm of variable gradient correction BP so as to enhance the rate of convergence.The performance of recognition under the low SNR is improved,and the modulation recognition of the signal is realized based on the modulation system of the constellation diagram.Simulation results show that the recognition rate based on this algorithm is enhanced over 30% compared with methods of adopting clustering algorithm or neural network based on BP algorithm alone under the low SNR.The recognition rate can reach 90%when the SNR is 4 dB,and the method is easy to be achieved so that it has broad application prospect in the modulating recognition.
出处
《中山大学学报（自然科学版）》
CAS
CSCD
北大核心
2015年第2期24-29,共6页
Acta Scientiarum Naturalium Universitatis Sunyatseni
基金
国家自然科学基金资助项目(61072070
61301179)
科技型中小企业创新资金项目(14C26214402603)
广东省科技计划资助项目(2011B010200030
2012B010100038)
关键词
变梯度修正BP算法
聚类算法
特征值的提取
神经网络
调制识别
algorithm of variable gradient correction BP
clustering algorithm
feature extraction
neural network
modulation recognition
