摘要
为更好地消除噪声,提出TD-BP神经网络拟合未知的噪声信道滤波器,讨论了它的模型与学习算法及其通用逼近性.基于TD-BP神经网络的噪声消除器具有神经网络分布式并行信息处理能力,利用TD-BP逼近噪声,然后抵消噪声得到有用信号,具有较好的鲁棒性和除噪性能.
This article introduces TD-BP network which can simulate sealed noise channel filter to cancel noise better. Its model, learning algorithm and universal approximation are discussed. Noise canceller based on TD-BP network has the ability to parallel-distributed in signal processing. During cancellation, TD-BP network is used to approximate noise, then the noise is cancellcd and the useful signal is recovered. The simulation result indicates that TD-BP network is robust and easy to use.
出处
《信阳师范学院学报（自然科学版）》
CAS
北大核心
2006年第4期455-457,共3页
Journal of Xinyang Normal University(Natural Science Edition)
