机器人听觉实时自适应增益的研究与实现  

Research and implement of real-time self-adaptation gain on robot hearing

作　　者：谭平[1] 蔡自兴[1] 余伶俐[1] 

机构地区：[1]中南大学信息科学与工程学院,湖南长沙410083

出　　处：《华中科技大学学报（自然科学版）》2008年第S1期46-49,共4页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家基础研究基金资助项目(A1420060159)

摘　　要：针对现有人机语言交互,当距离相对较远时,其交互能力急剧变差的问题,提出使用语音实时自适应增益处理方法,采用嵌入式微处理器对语言信号进行自适应增益处理,提高机器人的听觉范围,并且在嵌入式系统上予以实现.实验结果表明增加自适应增益处理模块后,在较远距离下,能提高机器人的识别率,使机器人的语言交互范围扩大,提高交互性能.Nowadays,the human-robot conversation usually needs the person keep short distance to the robot,as the human body is farther away from the robot,the speech interactive performance is becoming worse sharply.To address this issue,we present the method of real-time self-adaptive gain processing on speech to improve the robot′s hearing region,which has been implemented on the embedded system.The result of experiment shows that the robot′s speech recognition rate is raised after using this method when the man is...

关 键 词：机器人听觉 自适应增益 嵌入式系统 人机交互 识别率 

分 类 号：TP24[自动化与计算机技术—检测技术与自动化装置]