线性神经网络在语音干扰效果评估中的应用

Application of Linear Neural Network in Evaluating Jammed Speech Communication Effect

客观干扰效果评估方法研究是进行语音干扰效果评估方法研究的一个主要方向,在系统介绍线性神经网络结构、功能的基础上,具体分析了带延迟线的线性神经网络的信号预测能力,并对语音信号进行了预测,得到利用线性神经网络预测误差进行语音干扰效果评估的方法,给出了该方法的具体运算步骤,并通过一组数据验证了该参数的一致性。

Objectively evaluating speech communication effect is a common way in evaluation. The method to predict signal using the linear network with tapped delay line is analyzed, based on introducing the structure and the function of linear neural network. A new method of evaluating jammed speech communication effect using linear network is presented, through analyzing the network predicting result to speech. The process of the method is showed, and the consistency is validated using test data.