基于自适应算法的汽车主动降噪系统的研究与设计

发布时间：2018-08-04 18:14

【摘要】：随着汽车产业的大力发展,乘车品质逐渐被消费者所注意,这就让汽车的噪音控制水平成为各大汽车厂商的一个瓶颈所在。汽车有着复杂的噪声成分,其噪声主要由低频率噪声组成,且夹有部分中间频率的噪声。之前的被动降噪技术,主要对高频率的噪声有很好的抑制作用,但对于低频噪声收效甚微。而主动降噪技术利用的是声波干涉及相互抵消的原理,产生的次级声波与原噪声幅值相等、相位反转,从而达到降噪目的。在此基础理论之上,本文做了如下工作:首先,对课题的研究背景以及主被动降噪技术做了一定的介绍,又对其中涉及到的声学基本原理做了分析和推导,使技术的实现有了理论的指导;同时,从汽车的自身结构出发,对其噪声的产生做了分析和介绍。其次,由于自适应滤波器具有实时性强,结构单一等优点,因此本文的主动降噪技术主要应用了该方法。自适应滤波器主要由两部分组成,横向滤波器模块和自适应修正模块,横向滤波器依靠自适应算法实时传来的误差输入,依据误差最小化原则来调节自身权值,以适应噪声环境的不断改变。与此同时,又对主动降噪的通信系统模型,包括前馈型、反馈型、多通道型以及混合型四类做了分析和介绍。最后,在实验阶段,对横向滤波器的长度对滤波效果的影响进行了论证;还在基于单频前馈双通道系统模型做了对固定步长和变步长的对比实验,并找到了本实验模型的理想变步长参数,相比于传统模型,本文的改进变步长算法降噪效果明显,且相比于其他单通道仿真而言,本文设计的单频双通道仿真模型可进一步模拟真实的多通道环境,为降噪理论能够获得实际应用做出了一定贡献;文中最后做了真实语音信号的仿真,从收敛性和处理时间两方面取得了满意的实验结果。
[Abstract]:With the rapid development of automobile industry, the quality of car has been paid more and more attention by consumers, which makes the noise control level of automobile become a bottleneck for major automobile manufacturers. Automobile has complex noise components, and its noise is mainly composed of low frequency noise, and there is a part of intermediate frequency noise. The previous passive noise reduction technique has a good effect on the high frequency noise, but it has little effect on the low frequency noise. The active noise reduction technique utilizes the principle of acoustic interference and mutual cancellation. The secondary acoustic wave is equal to the original noise amplitude and the phase is reversed, so the noise reduction can be achieved. On the basis of this basic theory, this paper has done the following work: first, the research background of the subject and the active and passive noise reduction technology have been introduced to a certain extent, and the basic principles of acoustics involved therein have been analyzed and deduced. The realization of the technology is guided by theory, and the noise generation of the automobile is analyzed and introduced from the view of its own structure. Secondly, because the adaptive filter has the advantages of real time and single structure, the active noise reduction technique is mainly used in this paper. The adaptive filter consists of two parts: the transversal filter module and the adaptive correction module. The transversal filter is based on the error input from the adaptive algorithm and adjusts its weight according to the principle of error minimization. To adapt to the changing noise environment. At the same time, the communication system models of active noise reduction, including feedforward type, feedback type, multi-channel type and hybrid type, are analyzed and introduced. Finally, in the experiment stage, the influence of the length of the transversal filter on the filtering effect is demonstrated, and the comparison experiment of the fixed step size and the variable step size is also done on the basis of the single-frequency feedforward double-channel system model. The ideal variable step parameters of the experimental model are found. Compared with the traditional model, the improved variable step size algorithm in this paper has obvious effect on noise reduction, and compared with other single channel simulation. The single frequency and double channel simulation model designed in this paper can further simulate the real multi-channel environment and make a certain contribution to the practical application of noise reduction theory. Finally, the simulation of real speech signal is done. The experimental results are satisfactory in terms of convergence and processing time.
【学位授予单位】：燕山大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：U461.4

【相似文献】