基于改进Hopfield神经网络的自适应二维噪声对消器  

Adaptive 2-D Noise Canceller Using Improved Hopfield Neural Network

作　　者：黄健[1] 张冰[1] 

机构地区：[1]江苏科技大学电子信息学院,江苏镇江212003

出　　处：《电声技术》2008年第2期75-78,共4页Audio Engineering

摘　　要：提出了基于连续型Hopfield神经网络(CHNN)的自适应二维噪声对消器,讨论了神经网络的结构和原理及相应的自适应滤波算法,并从理论上进行了论证。仿真结果表明相对于采用最小均方算法的二维线性噪声对消器,CHNN噪声对消器能更有效实现二维噪声的消除,保持原信号的完整性,获得较好的去噪声效果。A novel method to design adaptive 2-D noise canceller based on CHNN（Continuous Hopfield Neural Network） is proposed. The structure and principal of neural network are investigated with its relevant adaptive filtering algorithm, whieh are proved theoretically. Simulation results demonstrate that CHNN noise eanceller is more efficient than 2-D linear canceller using LMS algorithm in cancellation of noise, and it keeps the integrality of original signal with preferable noise cancelling effects.

关 键 词：连续型HOPFIELD神经网络 自适应滤波器 噪声对消器 多维信号处理 

分 类 号：TB53[理学—物理]