神经网络及同步方程自适应噪声抵消方法研究
|
摘要
随着现代工业的发展,噪声控制问题日益引起了人们的关注。基于自适应控制策略的有源噪声抵消技术已经成为有源噪声控制领域的重要研究内容之一。在自适应有源噪声控制中,控制算法是直接影响自适应控制品质的重要因素。目前,国内外在控制算法的研究方面已经取得了丰硕的成果,但其中一些算法存在着明显的不足和特定性,因此运用新的信息分析手段对控制算法进行研究是一个非常有潜力而且具有重要意义的工作。
针对这一问题,论文选择丁以控制算法为切入点,主要研究了两类自适应有源噪声抵消方法:基于神经网络的自适应有源噪声抵消方法和基于同步方程方法的自适应有源噪声抵消方法。
为了有效的控制非线性噪声,论文深入研究了FXBPNN算法,该算法采用BP(back-propagation)神经网络作为自适应有源噪声抵消系统的控制器,它能够抵消非线性噪声,但运算量大。因此在这种算法的基础上,论文提出了一种运算量小的FEBPNN算法。并对FEBPNN算法、FXBPNN算法和FXLMS算法进行了计算机仿真比较分析。
另外,论文深入研究了基于同步方程方法的自适应有源噪声抵消算法,该算法通过估计附加滤波器避免了次级通道建模,但运算量也很大。因此在这种算法的基础上,论文提出了一种改进的基于同步方程方法的自适应有源噪声抵消算法,通过存储少量的主输入噪声和误差信号,避免了估计附加滤波器,因而有效地降低了运算量。通过计算机仿真,将两种同步方程方法与FXLMS算法进行比较分析。
最后,利用文中研究的几种自适应有源噪声抵消算法对真实的海试数据进行处理分析。
With modern industrial development, the subject of noise control has attracted people's growing attention and concern. Adaptive active noise cancellation based on the adaptive control strategy has become one of important researches in the field of active noise control. Control algorithm is an important factor of directly affecting the adaptive control quality. So far, a lot of great achievements have been reached in the study of control algorithm all over the world. However, there are obvious deficiency and specificity in many of these algorithms. Therefore, it is a very potential and meaningful work to improve the performance of control algorithm by means of new information analytical tools.
Aimed at this point, the control algorithm is chosen to be studied in this paper. Two kinds of adapative active noise cancellation algorithms which are based on neural network and simultaneous equations are discussed in this paper.
To control nonlinear noise effectively, FXBPNN algorithm is studied. This algorithm using BP (back propagation) neural network as a controller of an adapative active noise cancellation system can cancel nonlinear noise. But it is has large amount of computation. Therefore, on the basis of this algorithm, FEBPNN algorithm with lower computational load is proposed in this paper.Then, Computer simulations are carry out to compare the FEBPNN algorithm with FXBPNN algorithm and FXLMS algorithm.
In addition, an adaptive active noise cancellation algorithm based on the simultaneous equations method is introduced in this paper. This algorithm can avoid secondary path modeling by system identification of an auxiliary filter. As a result, the amount of computation is increased. In this paper, an improved adaptive active noise cancellation algorithm based on the simultaneous equations method is proposed. By storing a small number of an input signal and an error signal, it can avoid this identification. Therefore, the amount of computation can be reduced greatly. Computer simulations are carry out to compare the two algorithms with FXLMS algorithm.
At last, the performance of several adaptive active noise cancellation algorithm studied in the paper is proved through analysis and trealing with the sea trial data.
针对这一问题,论文选择丁以控制算法为切入点,主要研究了两类自适应有源噪声抵消方法:基于神经网络的自适应有源噪声抵消方法和基于同步方程方法的自适应有源噪声抵消方法。
为了有效的控制非线性噪声,论文深入研究了FXBPNN算法,该算法采用BP(back-propagation)神经网络作为自适应有源噪声抵消系统的控制器,它能够抵消非线性噪声,但运算量大。因此在这种算法的基础上,论文提出了一种运算量小的FEBPNN算法。并对FEBPNN算法、FXBPNN算法和FXLMS算法进行了计算机仿真比较分析。
另外,论文深入研究了基于同步方程方法的自适应有源噪声抵消算法,该算法通过估计附加滤波器避免了次级通道建模,但运算量也很大。因此在这种算法的基础上,论文提出了一种改进的基于同步方程方法的自适应有源噪声抵消算法,通过存储少量的主输入噪声和误差信号,避免了估计附加滤波器,因而有效地降低了运算量。通过计算机仿真,将两种同步方程方法与FXLMS算法进行比较分析。
最后,利用文中研究的几种自适应有源噪声抵消算法对真实的海试数据进行处理分析。
With modern industrial development, the subject of noise control has attracted people's growing attention and concern. Adaptive active noise cancellation based on the adaptive control strategy has become one of important researches in the field of active noise control. Control algorithm is an important factor of directly affecting the adaptive control quality. So far, a lot of great achievements have been reached in the study of control algorithm all over the world. However, there are obvious deficiency and specificity in many of these algorithms. Therefore, it is a very potential and meaningful work to improve the performance of control algorithm by means of new information analytical tools.
Aimed at this point, the control algorithm is chosen to be studied in this paper. Two kinds of adapative active noise cancellation algorithms which are based on neural network and simultaneous equations are discussed in this paper.
To control nonlinear noise effectively, FXBPNN algorithm is studied. This algorithm using BP (back propagation) neural network as a controller of an adapative active noise cancellation system can cancel nonlinear noise. But it is has large amount of computation. Therefore, on the basis of this algorithm, FEBPNN algorithm with lower computational load is proposed in this paper.Then, Computer simulations are carry out to compare the FEBPNN algorithm with FXBPNN algorithm and FXLMS algorithm.
In addition, an adaptive active noise cancellation algorithm based on the simultaneous equations method is introduced in this paper. This algorithm can avoid secondary path modeling by system identification of an auxiliary filter. As a result, the amount of computation is increased. In this paper, an improved adaptive active noise cancellation algorithm based on the simultaneous equations method is proposed. By storing a small number of an input signal and an error signal, it can avoid this identification. Therefore, the amount of computation can be reduced greatly. Computer simulations are carry out to compare the two algorithms with FXLMS algorithm.
At last, the performance of several adaptive active noise cancellation algorithm studied in the paper is proved through analysis and trealing with the sea trial data.
引文
[1]王江荣.封闭空间噪声控制系统算法研究.兰州理工大学硕士论文.2006:11-20页
[2]朱江.有源噪声控制算法的研究及其应用.大连海事大学硕士论文.2007:21-26页
[3]钱燕.客车车内噪声的主动控制技术研究.南京林业大学硕士论文.2006:44-46页
[4]周尧.往复式机械管道噪声有源控制.西南交通大学硕士论文.2003:1-4页
[5]孙利峰.多通道有源噪声控制快速算法研究及实现.西北工业大学硕士论文.2007:1-14页
[6]王杰.互耦多通道自适应有源噪声抵消技术研究.哈尔滨工程大学博士论文.2004:94-104页
[7]孙进烨.自适应有源控制快速算法原理及性能研究.西北工业大学硕士论文.2006:1-13页
[8]董淑斌.自适应有源耳罩工程实现的关键技术研究.西北工业大学硕士论文.2005:1-28页
[9]夏道明.基于神经网络的自适应噪声抵消的研究.武汉理工大学硕士论文.2002:1-40页
[10]陈元琳.基于人工神经网络的动态系统仿真模型和算法研究.大庆石油学院硕士学位论文.2006:15-27页
[11]刘学广.车内低频噪声多次级声源有源消声系统研究.吉林大学博士学位论文.2004:1-12页
[12]鲍雪山.舰艇自噪声自适应有源抵消技术研究.哈尔滨工程大学博士论文.2007:124-134页
[13]李传光,姜丽飞,李长青.RBF网络在履带车辆舱室内有源降噪的实验研究.北京理工大学学报.2004,24(4):311-319页
[14]张奇志,贾永乐,周雅莉.有源噪声控制的递归神经网络方法.控制 与决策.2001,16(3):369-373页
[15]K.Fujii,M.Muneyasu and J.Ohga.Simultaneous Equations Method Not Requiring Secondary Path Filter,Proc.of Active'99,Fort Lauderdale,1999,12:941-948P
[16]张玉麟.小波自适应控制系统及其在噪声主动控制中的应用研究.重庆大学博士学位论文.2002:4页
[17]张玉麟,陈伟民,黄尚廉.小波变换自适应滤波器及其在噪声主动控制中的应用.控制与决策.2002,17(1):236-240页
[18]张菊香,岳阳.基于自调整量化因子模糊控制器的有源噪声控制.噪声与振动控制.2002,22(6):198-202页
[19]陈克安.有源噪声控制.国防工业出版社,2003:12-27页
[20]陈克安,马远良.自适应有源噪声控制--原理、算法及实现.西北工业大学出版社,1993:82-107页
[21]孙旭,陈端石.次级通道模型误差下滤波X型最小均方差算法性能的分析.声学学报.2003,28(1):234-237页
[22]张奇志,贾永乐.噪声有源稳定自适应控制.振动工程学报.2002,15(1):163-168页
[23]何振亚.自适应信号处理.科学出版社,2002:31-44页
[24]董淑斌.自适应有源耳罩工程实现的关键技术研究.西北工业大学硕士论文.2005:29-37页
[25]侯宏,杨建华,孙进才.神经网络BP算法在有源消声中的应用.西北工业大学学报.1999,17(1):76-79页
[26]温激鸿.基于BP神经网络的自适应有源消声系统.长沙国防科技大学机电及自动化学院机电工程研究所.电子技术应用.2000,12:23-27页
[27]丁士圻.人工神经网络基础.哈尔滨工程大学出版社,2004:45-70页
[28]蒋威,唐健,秦素梅.神经网络BP算法在噪声主动控制中的应用.沈阳工业学院学报.2002,21(4):111-116页
[29]戴葵,李伯民.神经网络设计.机械工业出版社,2002:197-207页
[30]樊荣.基于神经网络的自适应噪声抵消的研究.太原理工大学硕士学 位论文.2006:41-50页
[31]关碧华,周雅莉,尹建民,张奇志.噪声有源控制的模糊神经网络方法.北京机械工业学院学报.2002,17(3):233-238页
[32]张奇志,曾成,关碧华,尹建民,周雅莉.基于神经网络的风机管道噪声有源控制的实验研究.噪声控制.2002,7:216-221页
[33]Q.Z.Zhang,Y.L.Jia.Active noise hybrid feedforward/feedback control using neural network compensation.Journal of Vibration and Acoustics.2002(124):100-104P
[34]周雅莉,张奇志,贾永乐.基于DSP的自适应有源噪声控制系统.计算机应用.2002,2:163-167页
[35]Kensaku Fujii,Juro Ohga.Method to update the coefficients of the secondary path Filter under active noise control.Signal Processing.2001,81:381-387P
[36]Kensaku Fujii,Mitsuji Muneyasu,and Juro Ohga.Active Noise Control System Using the Simultaneous Equation Method without the Estimation of Error Path Filter Coefficients.Electronics and Communications in Japan.2002,85(12):299-305P
[37]Fujii K,Ohga J.Examination of the adaptation offixed-coefficient filters in active noise control.J Commun Tech Rep 1998:97-101R
[38]MITSUJI M,KAGAWA K,FUJII K.Performance improvement for distributed active noise control systems based on simultaneous equations method.IEICE Trans.Fundamentals,2005,E88-A(7):1760-1764P.
[39]T.Kohno,Y.Ohta and A.Sano.Adaptive active noise control algorithm without explicit identification of secondary path dynamics,IEICE Trans.Fundam,2003,J86-A:9-18P
[40]M.Muneyasu,Y.Wakasugi,O.Hisayasu and K.Fujii.Hybrid active noise control systems based on the simultaneous equations method,IEICE Trans.Fundam,2001,E84-A:479-481P
[41]K.Fujii,S.Hashimoto and M.Muneyasu.Application of frequency domain processing techniques to the simultaneous equations method,Proc.ICSV, 2003,10:275-282P
[42]Y.Kajikawa and Y.Nomura,An active noise control system without secondary path model,IEICE Trans.Fundam,1999,J82-A:209-217P.
[43]王宗锋,张奇志.基于同步方程方法的有源噪声控制.噪声与控制.2006,4:66-69页
[44]Y.Kajikawa and Y.Nomura.Frequency Domain Active Noise Control System without a Secondary Path Model via Perturbation Method.IEICE Trans.on Fundamentals,vol.,Dec.2001,E-84A(12):3090-3098P
[45]尹雪飞,陈克安,相敬林.自适应有源控制器的算法实现.噪声与振动控制.2004,3:151-155页
[46]郑宝玉.自适应滤波器原理.电子工业出版社,2006:236-249页
[2]朱江.有源噪声控制算法的研究及其应用.大连海事大学硕士论文.2007:21-26页
[3]钱燕.客车车内噪声的主动控制技术研究.南京林业大学硕士论文.2006:44-46页
[4]周尧.往复式机械管道噪声有源控制.西南交通大学硕士论文.2003:1-4页
[5]孙利峰.多通道有源噪声控制快速算法研究及实现.西北工业大学硕士论文.2007:1-14页
[6]王杰.互耦多通道自适应有源噪声抵消技术研究.哈尔滨工程大学博士论文.2004:94-104页
[7]孙进烨.自适应有源控制快速算法原理及性能研究.西北工业大学硕士论文.2006:1-13页
[8]董淑斌.自适应有源耳罩工程实现的关键技术研究.西北工业大学硕士论文.2005:1-28页
[9]夏道明.基于神经网络的自适应噪声抵消的研究.武汉理工大学硕士论文.2002:1-40页
[10]陈元琳.基于人工神经网络的动态系统仿真模型和算法研究.大庆石油学院硕士学位论文.2006:15-27页
[11]刘学广.车内低频噪声多次级声源有源消声系统研究.吉林大学博士学位论文.2004:1-12页
[12]鲍雪山.舰艇自噪声自适应有源抵消技术研究.哈尔滨工程大学博士论文.2007:124-134页
[13]李传光,姜丽飞,李长青.RBF网络在履带车辆舱室内有源降噪的实验研究.北京理工大学学报.2004,24(4):311-319页
[14]张奇志,贾永乐,周雅莉.有源噪声控制的递归神经网络方法.控制 与决策.2001,16(3):369-373页
[15]K.Fujii,M.Muneyasu and J.Ohga.Simultaneous Equations Method Not Requiring Secondary Path Filter,Proc.of Active'99,Fort Lauderdale,1999,12:941-948P
[16]张玉麟.小波自适应控制系统及其在噪声主动控制中的应用研究.重庆大学博士学位论文.2002:4页
[17]张玉麟,陈伟民,黄尚廉.小波变换自适应滤波器及其在噪声主动控制中的应用.控制与决策.2002,17(1):236-240页
[18]张菊香,岳阳.基于自调整量化因子模糊控制器的有源噪声控制.噪声与振动控制.2002,22(6):198-202页
[19]陈克安.有源噪声控制.国防工业出版社,2003:12-27页
[20]陈克安,马远良.自适应有源噪声控制--原理、算法及实现.西北工业大学出版社,1993:82-107页
[21]孙旭,陈端石.次级通道模型误差下滤波X型最小均方差算法性能的分析.声学学报.2003,28(1):234-237页
[22]张奇志,贾永乐.噪声有源稳定自适应控制.振动工程学报.2002,15(1):163-168页
[23]何振亚.自适应信号处理.科学出版社,2002:31-44页
[24]董淑斌.自适应有源耳罩工程实现的关键技术研究.西北工业大学硕士论文.2005:29-37页
[25]侯宏,杨建华,孙进才.神经网络BP算法在有源消声中的应用.西北工业大学学报.1999,17(1):76-79页
[26]温激鸿.基于BP神经网络的自适应有源消声系统.长沙国防科技大学机电及自动化学院机电工程研究所.电子技术应用.2000,12:23-27页
[27]丁士圻.人工神经网络基础.哈尔滨工程大学出版社,2004:45-70页
[28]蒋威,唐健,秦素梅.神经网络BP算法在噪声主动控制中的应用.沈阳工业学院学报.2002,21(4):111-116页
[29]戴葵,李伯民.神经网络设计.机械工业出版社,2002:197-207页
[30]樊荣.基于神经网络的自适应噪声抵消的研究.太原理工大学硕士学 位论文.2006:41-50页
[31]关碧华,周雅莉,尹建民,张奇志.噪声有源控制的模糊神经网络方法.北京机械工业学院学报.2002,17(3):233-238页
[32]张奇志,曾成,关碧华,尹建民,周雅莉.基于神经网络的风机管道噪声有源控制的实验研究.噪声控制.2002,7:216-221页
[33]Q.Z.Zhang,Y.L.Jia.Active noise hybrid feedforward/feedback control using neural network compensation.Journal of Vibration and Acoustics.2002(124):100-104P
[34]周雅莉,张奇志,贾永乐.基于DSP的自适应有源噪声控制系统.计算机应用.2002,2:163-167页
[35]Kensaku Fujii,Juro Ohga.Method to update the coefficients of the secondary path Filter under active noise control.Signal Processing.2001,81:381-387P
[36]Kensaku Fujii,Mitsuji Muneyasu,and Juro Ohga.Active Noise Control System Using the Simultaneous Equation Method without the Estimation of Error Path Filter Coefficients.Electronics and Communications in Japan.2002,85(12):299-305P
[37]Fujii K,Ohga J.Examination of the adaptation offixed-coefficient filters in active noise control.J Commun Tech Rep 1998:97-101R
[38]MITSUJI M,KAGAWA K,FUJII K.Performance improvement for distributed active noise control systems based on simultaneous equations method.IEICE Trans.Fundamentals,2005,E88-A(7):1760-1764P.
[39]T.Kohno,Y.Ohta and A.Sano.Adaptive active noise control algorithm without explicit identification of secondary path dynamics,IEICE Trans.Fundam,2003,J86-A:9-18P
[40]M.Muneyasu,Y.Wakasugi,O.Hisayasu and K.Fujii.Hybrid active noise control systems based on the simultaneous equations method,IEICE Trans.Fundam,2001,E84-A:479-481P
[41]K.Fujii,S.Hashimoto and M.Muneyasu.Application of frequency domain processing techniques to the simultaneous equations method,Proc.ICSV, 2003,10:275-282P
[42]Y.Kajikawa and Y.Nomura,An active noise control system without secondary path model,IEICE Trans.Fundam,1999,J82-A:209-217P.
[43]王宗锋,张奇志.基于同步方程方法的有源噪声控制.噪声与控制.2006,4:66-69页
[44]Y.Kajikawa and Y.Nomura.Frequency Domain Active Noise Control System without a Secondary Path Model via Perturbation Method.IEICE Trans.on Fundamentals,vol.,Dec.2001,E-84A(12):3090-3098P
[45]尹雪飞,陈克安,相敬林.自适应有源控制器的算法实现.噪声与振动控制.2004,3:151-155页
[46]郑宝玉.自适应滤波器原理.电子工业出版社,2006:236-249页