小波包变换用于非高斯噪声下的信号检测

【作者】 林蓉；

【导师】 吕明；

【作者基本信息】 电子科技大学 ， 信号与信息处理， 2005， 硕士

【摘要】 信号检测方法是信号处理的一个重要研究方向,以前的信号检测方法多是基于高斯噪声的讨论,对非高斯噪声中的信号检测研究较少。但是在雷达,通讯等领域的实际问题中,高斯噪声的假设往往难以满足实际情况。以雷达系统为例,由于大气和人工产生的各种随机性的脉冲噪声对信道的干扰,总的噪声过程将会偏离高斯分布,而呈现出拖尾分布的特性。这时,基于高斯分布假设的经典最优监测器的性能将严重恶化。小波变换通过伸缩和平移等运算功能,对函数或信号进行多尺度细化分析,具有良好的局部化分析特性和多分辨分析特性,非常适合信号处理,成为一门兴起的信号处理技术。本文将小波包变换用于非高斯噪声统计特性的研究,提出一种新的非高斯分布噪声下的信号检测算法。第一章首先介绍了信号检测概况和发展以及作者的主要工作。第二章介绍了信号检测基本理论,其中包括二元和多元假设检测,统计信号检测系统的设计思想,二元假设检测判决准则等。讲述了信号的统计检测理论是以假设检测为工具,充分利用信号和噪声的统计特性,依据不同的判决准则来设计最佳接收机的理论。第三章介绍了小波基本理论,阐述了正交小波包的根本思想,它直接来源于多分辨分析和正交小波的构造。正交小波包除了本身具有的各种性质和应用之外,它成功地解决了小波变换固有地“高频低分辨”这一时-频分析上的缺陷。这也是本文用到的最基本的思想。第四章主要是对经典信号检测方法在各种噪声背景下的研究。首先介绍了经典信号检测系统模型。然后对含噪声信号通过经典信号检测模型的仿真,紧接着根据仿真数据和图形对经典信号检测系统的检测性能进行了分析,最后给出结论。第五章将小波包变换应用于信号检测中。首先阐述了非高斯噪声中检测信号用小波包变换方法的理论依据;然后通过仿真,证明更多还原

【Abstract】 Signal dection is one of the most important research branches in the signal process. The former detection methods are mostly based on the Gaussian noise, and soldom based on non-gaussian noise. However, in radar and communication fields, the supposed gaussian noise can not meet the requirements. Taking radar system as an example, the total noises in the communication channels with various random pulse-noise interference, do not meet gaussian distribution and it trails long. Therefore, the performance of the detection system based on gaussian noise is falling down rapidly. Wavelet transform analyses the signals or functions by the fine division which is coming from the retractility and movable characters of wavelet. It has the perfect performance of fine partly division and multiresolution, as it is fit for the signal process and becomes a new signal process technology. In the dissertation, it puts wavelet-packet decomposition into the study on non-gaussian noise. It offers a new signal dection method under non-gaussian noise backgroud. In the first chapter, it introduces the history of signal detection and the main work of the paper. In the second chapter, it introduces the basic theroy of signal detection, including the duality supposed detection and multiple supposed detection, the design principle for statistic signal detection system. It also gives the theory for the best receiver design according to the different adjudging rules on the statistic character of the signal and noise. In the third chapter, it introduces the basic theroy of the wavelet and the fundamental idea of the orhonormal wavelet-packet, which is coming from the multiresolution and the construction of orhonormal wavelet. The orhonormal wavelet-packet has many perfect characters更多还原