抗冲击噪声干扰的主动噪声控制算法研究

发布时间：2018-03-03 12:23

本文选题：主动噪声控制　切入点：自适应滤波原理　出处：《西南交通大学》2016年硕士论文　论文类型：学位论文

【摘要】：噪声污染仍然是全球公认的四大污染之一。面对日益严峻的噪声污染问题以及传统的无源噪声控制方法在低频段噪声时的不足,有源噪声控制的方法越来越得到人们的重视。目前,有源噪声控制理论虽然在国内外研究的比较深入,但能够完全满足工程实际应用还有一定的距离,这是因为还有一些问题需要更进一步研究与解决。如当噪声是冲击噪声时,一般的主动噪声控制系统很难达到理想的降噪效果或者根本不能降噪。然而在很多实际工作场合,如ICU病房中的呼吸机、婴儿保育箱、泵机、破碎机等工作场合中的噪声就是冲击噪声。为了使主动噪声控制系统对冲击噪声有很好的抑制作用,克服传统的基于二阶矩的主动噪声控制算法在冲击噪声时不容易收敛的缺点,本文提出了几种针对冲击噪声的主动噪声控制自适应算法,能够使主动噪声控制系统在冲击噪声时有着较快的收敛速度,较低的稳态误差以及较低的计算复杂度。本文主要内容以及创新点如下：首先,本文简要介绍了主动噪声控制理论的发展以及当前国内外研究现状,介绍了常见的主动噪声控制系统三大结构,即：前馈型主动噪声控制系统结构、反馈型主动噪声控制系统结构和多通道型主动噪声控制系统结构,分析了它们各自的优缺点以及所适用的工作环境。还详细分析了冲击噪声的数学模型。其次,本文介绍了常见的主动噪声控制算法,即：滤波最小均方(FxLMS)算法、滤波最小均方P范数(FxLMP)算法、滤波基于对数的最小均方(FxlogLMS)算法以及基于M估计的主动噪声控制算法(FxLMM),分析了各自的优缺点以及针对冲击噪声时,几种算法的性能。再次,在滤波x基于对数的最小均方(FxlogLMS):算法的基础上,提出了一种改进的算法(CRFxLMS算法)。该算法相较与FxlogLMS算法,在同样的冲击噪声情况下,计算复杂度相对较低,同时,该算法引入了凸组合的方法,能够保证提高整个算法的收敛速度的同时,将稳态误差降到更低。通过计算机仿真以及实测数据的验证,证明了算法的性能。最后,本文还介绍了另一种改进的主动噪声控制算法,即基于LMS算法的凸组合最大相关熵(CMCCFxLMS)算法,它是基于最大相关熵(MCC)理论提出来的。由已有的论文可知,基于LMS算法的最大相关熵算法对冲击噪声有很好的抑制作用。为了更进一步提高其性能,将凸组合的方法引入该算法,改进凸组合混合参数的更新方式,并将其应用到主动噪声控制中。使该算法在冲击噪声的情况下有着在大步长时较快的收敛速度,小步长时较低的稳态误差的特点。经过Matlab软件仿真证明了该算法的性能。
[Abstract]:Noise pollution is still recognized as one of the four major pollution in the world. Facing the increasingly serious problem of noise pollution and the shortcomings of traditional passive noise control methods in low frequency band noise, The method of active noise control has been paid more and more attention. Although the theory of active noise control is deeply studied at home and abroad, there is still a certain distance to be fully satisfied with the practical application of active noise control. This is because there are still some problems that need to be further studied and solved. For example, when noise is impulse noise, it is difficult for a general active noise control system to achieve an ideal noise reduction effect or can not be de-noised at all. However, in many practical work situations, For example, the noise in the workplace such as ventilator, baby care box, pump machine, crusher and so on in the ICU ward is the impact noise. To overcome the disadvantage that the traditional active noise control algorithm based on second moment is not easy to converge under impulse noise, this paper proposes several adaptive active noise control algorithms for impulse noise. It can make the active noise control system have faster convergence speed, lower steady-state error and lower computational complexity under impulse noise. The main contents and innovations of this paper are as follows: first, This paper briefly introduces the development of active noise control theory and the current research situation at home and abroad, and introduces three main structures of active noise control system, I. E. feedforward active noise control system structure. The structure of the feedback active noise control system and the multi-channel active noise control system are analyzed, their respective advantages and disadvantages and the applicable working environment are analyzed. The mathematical model of impact noise is also analyzed in detail. This paper introduces some common active noise control algorithms, I. e., filtering least mean square FxLMSalgorithm, filtering least mean square P-norm algorithm and FxLMPalgorithm. The filtering algorithm based on log-based minimum mean square (LMS) and the active noise control algorithm based on M estimation are analyzed. The advantages and disadvantages of each algorithm and the performance of several algorithms for impulse noise are analyzed. Based on the log-based minimum mean square (LMS) algorithm of filtering x, an improved algorithm is proposed. Compared with FxlogLMS algorithm, the computational complexity of this algorithm is lower than that of FxlogLMS algorithm, and at the same time, the computational complexity of the algorithm is lower than that of FxlogLMS algorithm. The algorithm introduces the convex combination method, which can improve the convergence speed of the whole algorithm and reduce the steady-state error to a lower level. The performance of the algorithm is proved by computer simulation and the verification of the measured data. This paper also introduces another improved active noise control algorithm, the convex combinatorial maximum correlation entropy (CMCCFxLMS) algorithm based on the LMS algorithm, which is based on the maximum correlation entropy (MCC) theory. The maximum correlation entropy algorithm based on LMS algorithm has a good effect on suppressing impulse noise. In order to further improve its performance, the convex combination method is introduced into the algorithm to improve the updating method of the mixed parameters of convex combination. It is applied to active noise control. The algorithm has the characteristics of fast convergence speed in large step size and low steady-state error in small step size in the case of impulse noise. The performance of the algorithm is proved by the simulation of Matlab software.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TB535

【相似文献】