基于MFCC特征提取的全新节能机器人设计与实现

Design and Implementation of a New Energy-Saving Robot Based on MFCC Feature Extraction

对基于MFCC特征提取的全新节能机器人设计与实现进行研究与实验,并将面向服务的深度学习技术应用,其中构建MFCC特征提取的全新节能机器人智能语音识别。在梅尔滤波实验过程中,对原始小云语音信号时域波形及语谱图绘制,同时对其进行预加重,并分析其分帧的语音信号在短时间(通常10-30ms)内具有平稳性,通过关键词小云语音信号进行实验并加入短时间内的能量特征,并引入梅尔频率倒谱系数(MFCC)的一阶差分和二阶差分,对语音数据特征提取和降低运算维度感知敏感度,从而提高特征提取的有效性。在此基础上,能够更好地反映全新节能机器人在复杂环境下,它能够较好地表述语音信号特征并且能准确将语音信号特征进行提取。因此,提出基于MFCC特征提取的全新节能机器人设计与实现,该系统主要实现:核心技术从语音信号采集、预处理、特征提取然后对外界设备进行相关控制功能。

Studies and experiments the design and implementation of a new energy-saving robot based on MFCC feature extraction,and applies the service-oriented deep learning technology,in which a new energy-saving robot intelligent speech recognition based on MFCC feature ex⁃traction is constructed.In the Mel filtering experiment,the time-domain waveform and spectrum of the original voice signal of small cloud are drawn,and it is pre weighted at the same time,and the frame speech signal is analyzed to be stable in a short time(usually 10-30ms).Through the key words of small cloud speech signal experiment,the energy characteristics in a short time are added,the first-or⁃der difference and the second-order difference of Mel Frequency Cepstrum Coefficient(MFCC)are introduced to extract the features of speech data and reduce the sensitivity of operation dimension perception,so as to improve the effectiveness of feature extraction.On this basis,it can better reflect the new energy-saving robot in the complex environment,it can better express the speech signal features and ac⁃curately extract the speech signal features.Therefore,a new energy-saving robot design and Implementation Based on MFCC feature ex⁃traction is proposed.The main implementation of the system is as follows:the core technology is voice signal acquisition,preprocessing,feature extraction and then related control functions for external equipment.