机器人双耳声源定位  

Sound source localization for robotics by binaural methods

作　　者：甘宜超[1,2] 孙骞[1] 辛丽[1] 钱广秀 高寒旭 贾丽[3] 贾刚[3] 常发亮[4] 赵增顺[1,4] GAN Yi-chao;SUN Qian;XIN Li;QIAN Guang-xiu;GAO Han-xu;JIA Li;JIA Gang;CHANG Fa-liang;ZHAO Zeng-shun(School of Electronics Communications and Physics,Shandong University of Science and Technology,Qingdao 266590,Shandong P ro vince,China;School of Electronic Engineering,Xidian University,Xi'an 710000,China;School of Control Science and Engineering,Shandong University,Jinan 250061,China;Dongying Branch Company,China Unicom ,Dongying 257000,Shandong Province,China)

机构地区：[1]山东科技大学电子通信与物理学院,山东青岛266590 [2]西安电子科技大学电子工程学院,西安710000 [3]中国联通东营分公司,山东东营257000 [4]山东大学控制科学与工程学院,济南250061

出　　处：《信息技术》2017年第7期9-13,18,共6页Information Technology

基　　金：国家自然科学基金(61174190);山东省自然科学基金(ZR2014FM002);浙江省"仪器科学与技术"重中之重学科开放基金(JL150508);中国博士后科学基金特别资助(2015T80717)

摘　　要：机器人拥有听声辩位能力是机器人智能化的重要表现。研究机器人声源定位具有广泛的应用前景。对于基于双耳听觉的声源定位,文中根据听觉系统中的"双耳效应"及"耳廓效应"等线索进行定位,首先建立了拟人机器人头部模型,通过头部转动实现前后方位的判定。其次,运用时间差(ITD)算法模型和声级差(ILD)算法模型分别实现了信号的水平方位的定位。并且应用互相关算法进行信号时延的估算。最后根据耳廓效应,根据频谱的变化实现了垂直方位的定位。本文通过定位算法编写程序,并且通过输入的声源信号实现声源方位的确定及误差分析。Robot auditory perception ability is an important manifestation of robot intelligent.Sound source localization technology is one of the important research directions of robot audition and has good prospects for application.For binaural approaches,this paper uses the binaural effect and pinnae principle clues in the auditory system to sound source localization in robotics.The head model of humanoid robot is established,and the position of the front or back azimuth is determined by the rotation of the head.Then it uses ITD algorithm and the ILD algorithm to achieve the azimuth.Time-delay estimator is calculated according to the cross Correlation.The program is written based on the positioning algorithm and the location of the sound source is determined according to the input of the signal.

关 键 词：双耳听觉 ITD ILD 时延估计 

分 类 号：TP319[自动化与计算机技术—计算机软件与理论]