基于线形麦克风阵列的声源定位算法改进与噪声的实验分析
|
摘要
人们对“声”的研究由来已久。随着科学技术的不断发展和研究手段的不断进步,音频信号处理这一方向又出现了许多新的热点与焦点。基于麦克风阵列的音频信号处理就是其中之一。因其自身优越的性能,开始被应用于许多新的领域,如视频会议、语音识别、车载系统、机器人控制、助听器、地震报警等。
麦克风阵列的基本用途可以归纳为实现声源定位、跟踪与实现抗干扰语音增强两种。虽然这两种用途已经被人们广泛研究,但是依然有许多有待解决的问题。本文的主要工作就是对这两种基于麦克风阵列的语音信号处理方法所遇到的一些实际问题进行分析并提出改进解决方案。内容包括:
1.简单介绍了麦克风阵列用于声源定位的几种常用方法与用于语音增强时的原理,讨论了它们的各自特点。
2.对于延时累加可控波束形成方法中所遇到的对称双主瓣、空间角分辨率下降、频率选择性衰落等主要问题,提出了解决办法。计算机仿真结果说明,联合使用所提出的三个方法,可以提高室内线形全向麦克风阵列声源定位与跟踪的性能,而不必改变阵列拓扑结构、不增加硬件成本。
3.在用于语音增强时,波束形成方法对于宽频带信号所对应的不同频点的空间带宽不同所带来的音色影响问题,讨论了一种恒定波束宽度方法并对已有方法进行改进。提高了运算速度,减小了系统开销,同时也达到了比较好的控制波束宽度的效果,并给出了计算机仿真结果。
4.通过进行麦克风音频采集与数据处理实验所得到数据的归纳分析,总结出了工程实现中需要注意实际问题。得到了实际采集系统中噪声等因素对阵列信号处理的关键变量——信号相位的影响并归纳出一些解决方法。
Although the research no sound has a long history, there are many new focuses with the continuous development of science and technology and the progress on the means of research. One of them is investigation on microphone array that has been applied to many areas because of its advanced character, such as video-conferences, hearing aid design, robot control, speech recognition, earthquake alarming and so on.
The two common functions of microphone array are sound source localization/tracking and speech enhancement. A lot of work has been done about them, but the problems still exist. The contents of this thesis are as follows:
1. The general methods based on microphone array for DOA (Direction of Arrival) and speech enhancement was introduced.
2. The performance of linear microphone array was analyzed. And solutions to the problems that will occur when the microphone array serves for sound source localization and tracking in a closed environment was put forward. Using the three methods proposed together will improve the performance of the microphone array highly without either changing the topology of array or increasing additional hardware devices cost.
3. An adaptive method based on microphone array to involve the constant beamwidth was introduced. And another method that can be faster was proposed. Both of them give the solutions to the problem that different frequencies make the beamwidth multiply.
4. The data of a group of experiments was showed to find the character of real DAQ(data acquisition) environment. The noise which was involved by the circuit of DAQ system made the phases of signals inaccurate. The inaccurate phases impacted the result of the microphone array directly. Some solutions which were based on those experiments about inaccurate phase was put forward at the end of each experiments.
麦克风阵列的基本用途可以归纳为实现声源定位、跟踪与实现抗干扰语音增强两种。虽然这两种用途已经被人们广泛研究,但是依然有许多有待解决的问题。本文的主要工作就是对这两种基于麦克风阵列的语音信号处理方法所遇到的一些实际问题进行分析并提出改进解决方案。内容包括:
1.简单介绍了麦克风阵列用于声源定位的几种常用方法与用于语音增强时的原理,讨论了它们的各自特点。
2.对于延时累加可控波束形成方法中所遇到的对称双主瓣、空间角分辨率下降、频率选择性衰落等主要问题,提出了解决办法。计算机仿真结果说明,联合使用所提出的三个方法,可以提高室内线形全向麦克风阵列声源定位与跟踪的性能,而不必改变阵列拓扑结构、不增加硬件成本。
3.在用于语音增强时,波束形成方法对于宽频带信号所对应的不同频点的空间带宽不同所带来的音色影响问题,讨论了一种恒定波束宽度方法并对已有方法进行改进。提高了运算速度,减小了系统开销,同时也达到了比较好的控制波束宽度的效果,并给出了计算机仿真结果。
4.通过进行麦克风音频采集与数据处理实验所得到数据的归纳分析,总结出了工程实现中需要注意实际问题。得到了实际采集系统中噪声等因素对阵列信号处理的关键变量——信号相位的影响并归纳出一些解决方法。
Although the research no sound has a long history, there are many new focuses with the continuous development of science and technology and the progress on the means of research. One of them is investigation on microphone array that has been applied to many areas because of its advanced character, such as video-conferences, hearing aid design, robot control, speech recognition, earthquake alarming and so on.
The two common functions of microphone array are sound source localization/tracking and speech enhancement. A lot of work has been done about them, but the problems still exist. The contents of this thesis are as follows:
1. The general methods based on microphone array for DOA (Direction of Arrival) and speech enhancement was introduced.
2. The performance of linear microphone array was analyzed. And solutions to the problems that will occur when the microphone array serves for sound source localization and tracking in a closed environment was put forward. Using the three methods proposed together will improve the performance of the microphone array highly without either changing the topology of array or increasing additional hardware devices cost.
3. An adaptive method based on microphone array to involve the constant beamwidth was introduced. And another method that can be faster was proposed. Both of them give the solutions to the problem that different frequencies make the beamwidth multiply.
4. The data of a group of experiments was showed to find the character of real DAQ(data acquisition) environment. The noise which was involved by the circuit of DAQ system made the phases of signals inaccurate. The inaccurate phases impacted the result of the microphone array directly. Some solutions which were based on those experiments about inaccurate phase was put forward at the end of each experiments.
引文
[1]李昌立.第四讲 语音信号处理的现状与展望.[J]物理,2005,34(2),303-306
[2]C.R.McGeoch,A.D.Frost,C.R.Mingins,The Excitation of Surface Waveguides and Radiating Slots by Strip-Circuit Transmission Lines.[J].Microwave Theory and Techniques,IEEE Transactions on,1956,4(4),218-222
[3]A.T.Waterman,A Rapid.Beam-Swinging Experiment in Transhorizon Propagation.[J].Antennas and Propagation,IEEE Transactions,1958,6(4),338-340
[4]W.H.Von Aulock,Properties of Phased Arrays.[J].IEEE:Proceedings of the IRE,1960,48(10),1715-1727
[5]洪鸥,麦克风阵列语音增强技术及其应用.[J].微计算机信息,2006(1),142-144
[6]Markus Buck,Tim Haulick,Hans-Jorg Pfleiderer.Self-calibrating microphone arrays for speech signal acquisition:A systematic approach.[J].Signal Processing,2006,86(6),1230-1238
[7]杜江,朱柯.智能麦克风阵列语音分离和说话人跟踪技术研究.[J].电子学报,2005,33(2),382-384
[8]GriffithsLJ,JimCW.An alternative approach to linearly constrained adaptive beamforming[J].IEEE Transactions on Antenna Propagation,1982,30(1):27-34
[9]姚文冰,姚天任,韩涛,稳健语音识别技术发展现状及展望[J],信号处理,2001,17(6),484-497
[10]赵贤宇,王作英.用于语音识别的鲁棒自适应麦克风阵列算法.[J].清华大学学报,2004(44),1433-1436
[11]Mumolo Enzo,Nolich Massimiliano,Vercelli Gianni.Alogrithms for acoustic localization based on microphone array in service robotics.[J].Robotics and Automomous System,2003,42,69-88
[12]Widrow Bernard,Luo Fa-Long,Microphone arrays for hearing aids:An overview.[J].Speech Communication,2003,39,139-146
[13]J.Z.Stafsudd,S.Asgari,R.Hudson K.Yao,E.Taciroglu.Localization of short-range acoustic and seismic wideband sources:Algorithms and experiments.[J].Journal of Sound and Vibration,2008,312,74-93
[14]乔渭阳,唐狄毅,U lfM ichel.基于麦克风阵列测量的机翼脱落涡噪声研究.[J].西北工业大学学报,2001,19(2),200-204
[15]张吴,朱林,葛利嘉.麦克风阵技术的现状及其应用前景.[J].重庆通信学院学报,2006,20(4),49-13
[16]泛华测控服务与产品手册[M],54
[17]刘庆华,周德新.基于语音激励信息的声源定位算法研究.[J].计算机工程与应用,2006(33),181-187
[18]贾智骏,蒋伟康.基于传声器阵列的电泵噪声源识别研究.[J].噪声与振动控制,2004,24(6),49-51
[19]Henning Puder.Adaptive signal processing for interference cancellation in hearing aids.[J].Signal Processing,2006,86(6),1239-1253
[20]周浩,顾晓东,蒋兴舟.基于MUSIC算法的矢量阵波达方向估计.[J].微计算机信息,2007,(02X),212-213,266
[21]Jae-Woong Choi,Yang-Hann Kim.Spherical beam-forming and MUSIC methods for the estimation of location and strength of spherical sound sources.[J].Mechanical Systems and Signal Processing,1995,9(5),569-588
[22]杨维,陈俊仕,李世明等.移动通信中的阵列天线技术[M],北京:清华大学出版社,北京交通大学出版社,2005,10
[23]金荣洪,耿军平,范瑜.无线通信中的智能天线[M],北京:北京邮电大学出版社,2006
[24]Tsoulos G,McGeehan J,Beach M.Space division multipile access Field Trails.[J].Radar,sonar having,1998,145(1),73-78
[25]Dimitris G.Manolakis,Vinay K.Ingle,Stephen M.Kogon,Statistical and adaptive signal processing[M],The McGraw-Hill Companies,2000
[26]韩纪庆,张磊,郑铁然.语音信号处理[M],北京:清华大学出版社,2005
[27]Cigada Alfredo,Francessco Ripamonti,Vanali Marcello,The delay & sum algorithm applied to microphone array measurements:Numerical analysis and experimental validation.[J].Mechanical Systems and Signal Processing,2007,21,2645-2664
[28]居太亮,彭启琮,邵怀宗等.基于任意麦克风阵列的声源二维DOA估计算法研究.[J].通信 学报,2005,26(8),129-133
[29]林志斌,徐柏龄.基于球麦克风阵列的三维空间多声源定位.[J].南京大学学报(自然科学),2006,42(4),384-394
[30]王学清,时银水,朱岩.空间六元麦克风阵列及其定位精度.[J].哈尔滨工业大学学报,2006,38(8),1392-1393
[31]何成林,杜利民,马昕.基于子带广义旁瓣相消器的麦克风阵列语音增强.[J].计算机应用研究,2006,23(4),208-210
[32]齐娜。传声器阵列恒定束宽波束形成.[J].电声技术,2006,24-25
[33]Taylor.T.T.Design of line source antennas for narrow beamwidth and Low sidelobes[J],IEEE Trans Antennas and Propagation,1955,3,16-28
[34]程佩青,数字信号处理教程(第二版)[M],北京:清华大学出版社,2001
[35]Robert J.Mailloux,Phased array antenna handbook(second edition)[M],Artech house inc,2005
[36]徐占朝,张盛林,赵国志,王金民.传声器的技术特性与种类.[J].中国医学教育技术,2006,16(1),58-60
[37]华子兴.实用传声器现状.电声技术,1994(3),20-22
[38]冯纪强,徐晨,张维强.一种改进的语音谱减去噪算法.[J].计算机工程与应用,2007,43(23).80-82
[39]王冬霞,殷福亮.联合波束形成与谱减法的麦克风阵列语音增强算法.[J].大连理工大学学报,.2006,46(1),121-126
[40]冯国友,戴扬,沈海斌等.孤立词语音识别中端点检测加速器的设计与实现[J],电子器件,2007,30(3),1098-1101
[41]刘晓明,覃胜,语音端点检测的仿真研究[J],系统仿真学报,2005(17),1974-1976
[42]Parham Aarabi.The fusion of distributed microphone arrays for sound localization[J].Applied Signal Processing,2003,4,338-347
[43]张吴,朱林,葛利嘉.麦克风阵技术的现状及其应用前景.[J].重庆通讯学院学报.2001,20(4),9-13
[44]Parham Aarabi,Sam Mavandadi.Robust sound localization using conditional time-frequency histograms.[J].Information Fusion 2003,4,111-122
[45]杨飞,刘增力,严云保,肖笛.基于小波多分辨分析的DOA估计.[J].通信技术.2006,5,67-68
[46]PengKe,LuJing,Xu BoLing.Robust adaptive beamforming in the frequency domain by using a small microphone array.[J].南京大学学报(自然科学),2004,4(40)
[2]C.R.McGeoch,A.D.Frost,C.R.Mingins,The Excitation of Surface Waveguides and Radiating Slots by Strip-Circuit Transmission Lines.[J].Microwave Theory and Techniques,IEEE Transactions on,1956,4(4),218-222
[3]A.T.Waterman,A Rapid.Beam-Swinging Experiment in Transhorizon Propagation.[J].Antennas and Propagation,IEEE Transactions,1958,6(4),338-340
[4]W.H.Von Aulock,Properties of Phased Arrays.[J].IEEE:Proceedings of the IRE,1960,48(10),1715-1727
[5]洪鸥,麦克风阵列语音增强技术及其应用.[J].微计算机信息,2006(1),142-144
[6]Markus Buck,Tim Haulick,Hans-Jorg Pfleiderer.Self-calibrating microphone arrays for speech signal acquisition:A systematic approach.[J].Signal Processing,2006,86(6),1230-1238
[7]杜江,朱柯.智能麦克风阵列语音分离和说话人跟踪技术研究.[J].电子学报,2005,33(2),382-384
[8]GriffithsLJ,JimCW.An alternative approach to linearly constrained adaptive beamforming[J].IEEE Transactions on Antenna Propagation,1982,30(1):27-34
[9]姚文冰,姚天任,韩涛,稳健语音识别技术发展现状及展望[J],信号处理,2001,17(6),484-497
[10]赵贤宇,王作英.用于语音识别的鲁棒自适应麦克风阵列算法.[J].清华大学学报,2004(44),1433-1436
[11]Mumolo Enzo,Nolich Massimiliano,Vercelli Gianni.Alogrithms for acoustic localization based on microphone array in service robotics.[J].Robotics and Automomous System,2003,42,69-88
[12]Widrow Bernard,Luo Fa-Long,Microphone arrays for hearing aids:An overview.[J].Speech Communication,2003,39,139-146
[13]J.Z.Stafsudd,S.Asgari,R.Hudson K.Yao,E.Taciroglu.Localization of short-range acoustic and seismic wideband sources:Algorithms and experiments.[J].Journal of Sound and Vibration,2008,312,74-93
[14]乔渭阳,唐狄毅,U lfM ichel.基于麦克风阵列测量的机翼脱落涡噪声研究.[J].西北工业大学学报,2001,19(2),200-204
[15]张吴,朱林,葛利嘉.麦克风阵技术的现状及其应用前景.[J].重庆通信学院学报,2006,20(4),49-13
[16]泛华测控服务与产品手册[M],54
[17]刘庆华,周德新.基于语音激励信息的声源定位算法研究.[J].计算机工程与应用,2006(33),181-187
[18]贾智骏,蒋伟康.基于传声器阵列的电泵噪声源识别研究.[J].噪声与振动控制,2004,24(6),49-51
[19]Henning Puder.Adaptive signal processing for interference cancellation in hearing aids.[J].Signal Processing,2006,86(6),1239-1253
[20]周浩,顾晓东,蒋兴舟.基于MUSIC算法的矢量阵波达方向估计.[J].微计算机信息,2007,(02X),212-213,266
[21]Jae-Woong Choi,Yang-Hann Kim.Spherical beam-forming and MUSIC methods for the estimation of location and strength of spherical sound sources.[J].Mechanical Systems and Signal Processing,1995,9(5),569-588
[22]杨维,陈俊仕,李世明等.移动通信中的阵列天线技术[M],北京:清华大学出版社,北京交通大学出版社,2005,10
[23]金荣洪,耿军平,范瑜.无线通信中的智能天线[M],北京:北京邮电大学出版社,2006
[24]Tsoulos G,McGeehan J,Beach M.Space division multipile access Field Trails.[J].Radar,sonar having,1998,145(1),73-78
[25]Dimitris G.Manolakis,Vinay K.Ingle,Stephen M.Kogon,Statistical and adaptive signal processing[M],The McGraw-Hill Companies,2000
[26]韩纪庆,张磊,郑铁然.语音信号处理[M],北京:清华大学出版社,2005
[27]Cigada Alfredo,Francessco Ripamonti,Vanali Marcello,The delay & sum algorithm applied to microphone array measurements:Numerical analysis and experimental validation.[J].Mechanical Systems and Signal Processing,2007,21,2645-2664
[28]居太亮,彭启琮,邵怀宗等.基于任意麦克风阵列的声源二维DOA估计算法研究.[J].通信 学报,2005,26(8),129-133
[29]林志斌,徐柏龄.基于球麦克风阵列的三维空间多声源定位.[J].南京大学学报(自然科学),2006,42(4),384-394
[30]王学清,时银水,朱岩.空间六元麦克风阵列及其定位精度.[J].哈尔滨工业大学学报,2006,38(8),1392-1393
[31]何成林,杜利民,马昕.基于子带广义旁瓣相消器的麦克风阵列语音增强.[J].计算机应用研究,2006,23(4),208-210
[32]齐娜。传声器阵列恒定束宽波束形成.[J].电声技术,2006,24-25
[33]Taylor.T.T.Design of line source antennas for narrow beamwidth and Low sidelobes[J],IEEE Trans Antennas and Propagation,1955,3,16-28
[34]程佩青,数字信号处理教程(第二版)[M],北京:清华大学出版社,2001
[35]Robert J.Mailloux,Phased array antenna handbook(second edition)[M],Artech house inc,2005
[36]徐占朝,张盛林,赵国志,王金民.传声器的技术特性与种类.[J].中国医学教育技术,2006,16(1),58-60
[37]华子兴.实用传声器现状.电声技术,1994(3),20-22
[38]冯纪强,徐晨,张维强.一种改进的语音谱减去噪算法.[J].计算机工程与应用,2007,43(23).80-82
[39]王冬霞,殷福亮.联合波束形成与谱减法的麦克风阵列语音增强算法.[J].大连理工大学学报,.2006,46(1),121-126
[40]冯国友,戴扬,沈海斌等.孤立词语音识别中端点检测加速器的设计与实现[J],电子器件,2007,30(3),1098-1101
[41]刘晓明,覃胜,语音端点检测的仿真研究[J],系统仿真学报,2005(17),1974-1976
[42]Parham Aarabi.The fusion of distributed microphone arrays for sound localization[J].Applied Signal Processing,2003,4,338-347
[43]张吴,朱林,葛利嘉.麦克风阵技术的现状及其应用前景.[J].重庆通讯学院学报.2001,20(4),9-13
[44]Parham Aarabi,Sam Mavandadi.Robust sound localization using conditional time-frequency histograms.[J].Information Fusion 2003,4,111-122
[45]杨飞,刘增力,严云保,肖笛.基于小波多分辨分析的DOA估计.[J].通信技术.2006,5,67-68
[46]PengKe,LuJing,Xu BoLing.Robust adaptive beamforming in the frequency domain by using a small microphone array.[J].南京大学学报(自然科学),2004,4(40)