基于机器人运动控制语音分类模型设计研究  

Design and Research of Speech Classification Model Based on Robot Motion Control

作　　者：王林 罗宏伟 李文龙 WANG Lin;LUO Hong-wei;LI Wen-long(Research Center for Digital Economy&New Technology Application,Wuhan Business University,Wuhan Hubei 430056,China;School of Mechanical Science&Engineering,Huazhong University of Science and Technology,Wuhan Hubei 430074,China)

机构地区：[1]武汉商学院数字经济与新技术应用研究中心,湖北武汉430056 [2]华中科技大学机械科学与工程学院,湖北武汉430074

出　　处：《计算机仿真》2025年第3期482-486,520,共6页Computer Simulation

基　　金：2021年湖北省教育厅哲学社会科学研究项目(21G110)。

摘　　要：通过LMS模型算法,对语音信号进行滤波处理,去除噪声。使用一阶FIR高通数字滤波器来实现预加重,将滤波后的语音进行加重处理,使特征更明显;再对语音信号进行分帧加窗处理,每秒语音信号分成33-100帧,分帧加窗后进行提取特征信号,利用BP神经网络算法对语音信号进行识别分类,通过分析出来的数据进行处理,得到与资料库中的数据进行取样比对,进行决策分类。结果表明,在实际应用中,语音分类的精准度达到75%,在具有较强噪音时,精确度仍能达到65%。LMS model algorithm is used to filter the speech signal and remove the noise.The first-order FIR high-pass digital filter is used to achieve pre-weighting,and the speech after filtering is weighted to make the features more obvious.Then the speech signal is divided into frames and window processing,and the speech signal is divided into 33-100 frames per second.After frame segmentation and window processing,the feature signal is extracted.The BP neural network algorithm is used to identify and classify the speech signal.The results show that the accuracy of speech classification can reach 75% in practical applications,and the accuracy can stll reach 65%when there is strong noise.

关 键 词：最小均方差算法 神经网络模型 语音分类 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]