非高斯噪声背景下的自适应滤波算法研究

发布时间：2018-06-06 22:59

本文选题：非高斯噪声模型 + 自适应滤波算法　；参考：《新疆大学》2017年硕士论文

【摘要】：在数字信号处理过程中,信号受到噪声干扰的现象十分常见,这些被噪声干扰的信号在识别以及后续的分析处理和判断工作中给我们带来一定的困扰。所以,在含有噪声的混合信号中检测并成功提取出我们所需要的有用信号是一个很重要课题。目前在信号理论相关领域中,我们习惯把噪声假设成高斯噪声,然后根据它的分布函数特点把噪声处理掉。但是随着科技的不断提高,我们的设备越来越复杂,在通信过程中遇到的都是非高斯噪声。按照高斯噪声处理效果并不理想,提取出的信号容易失真甚至出现严重的破坏和损失,最后使整个系统性能退化。本文从自适应滤波原理入手研究并提出新的算法,用于消除和抑制非高斯噪声。首先,介绍了几种常见的噪声,然后分析高斯噪声和非高斯的区别,根据非高斯噪声的特点提出了几种数学模型。接着分析它们的概率密度函数以及指出了它们的统计特性。其次,根据自适应滤波算法的原理,我们介绍了常见的自适应滤波器和它的基本结构和应用,然后对自适应滤波算法进行了列举。本文采用的是自适应滤波的一种典型应用——自适应噪声抑制系统。在非高斯噪声的前提下,常见的一些算法对信号的处理并不理想,因此本文研究了新的算法。本文结合归一化LMS(NLMS)算法和LMS-Newton算法的优点提出了NLMS-Newton算法,并根据它的特点给予了一些改进。在经过模拟验证后可以看出,虽然该算法的运算量变大了,但是收敛速度和稳定性得到了明显的提升。最后,在计算机上应用仿真软件MATLAB对该新算法进行仿真,根据信号处理后图像对算法的稳定性能和收敛性能进行对比分析。
[Abstract]:In the process of digital signal processing, the phenomenon that the signal is disturbed by noise is very common. Therefore, it is an important task to detect and successfully extract the useful signals from the mixed signals with noise. At present, in the field of signal theory, we usually assume the noise as Gao Si noise, and then deal with the noise according to its distribution function. But with the development of science and technology, our equipments are more and more complex, and all of them are non-Gao Si noise in the communication process. According to the Gao Si noise processing effect is not ideal, the extracted signal is prone to distortion and even serious damage and loss, finally the whole system performance degradation. Based on the principle of adaptive filtering, a new algorithm is proposed to eliminate and suppress non-Gao Si noise. Firstly, several common noises are introduced, then the difference between Gao Si noise and non-Gao Si noise is analyzed, and several mathematical models are proposed according to the characteristics of non-Gao Si noise. Then their probability density functions and their statistical properties are analyzed. Secondly, according to the principle of adaptive filtering algorithm, we introduce the common adaptive filter, its basic structure and application, and then enumerate the adaptive filter algorithm. In this paper, an adaptive noise suppression system is used as a typical application of adaptive filtering. Under the premise of non-Gao Si noise, some common algorithms are not ideal for signal processing. Therefore, a new algorithm is studied in this paper. Combined with the advantages of the normalized LMS-NLMS) algorithm and the LMS-Newton algorithm, this paper puts forward the NLMS-Newton algorithm and gives some improvements according to its characteristics. After simulation, we can see that the convergence speed and stability of the algorithm have been improved obviously, although the computation quantity of the algorithm has become larger. Finally, the simulation software MATLAB is used to simulate the new algorithm, and the stability and convergence of the algorithm are compared and analyzed according to the image after signal processing.
【学位授予单位】：新疆大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN713

【参考文献】