基于TDNN的线性噪声消除器

Linear noise canceller based on TDNN

提出基于TDNN神经网络的线性噪声消除器,讨论了它的模型与学习算法及其通用逼近性.该噪声滤除器具有神经网络分布式并行信息处理能力,利用TDNN逼近噪声,然后从测量中消去噪声得到有用信号,具有较好的鲁棒性和除噪性能.

Linear noise canceller based on TDNN was proposed, and its model, learning algorithm and universal approximation were discussed. The noise canceller has the ability of parallel-distributed signal processing, which is the advantage of neural networks. In the cancellation process, TDNN is used to approximate noise, and the noises are cancelled to recover the useful signals. The system is robust and is easy to use as a noise canceller.