【作者】 张凯；

【导师】 叶芃；

【作者基本信息】 电子科技大学 ， 工程硕士（专业学位）， 2022， 硕士

【摘要】 现代电子信息技术朝着高速高带宽的方向发展,不断提高对系统指标要求的同时,也对信号的全面分析也提出了新的要求,仅从时域或者频域的分析已经满足不了人们对信号特征的提取。本论文将从混合域仪器出发,解决了传统示波器只能从单一的信号域对信号进行采集恢复分析的局限,而且能够同时对时域、频域、时频联合域进行分析,形成对信号进行测量的多维度测量仪器。所以多域分析是当前信号特征分析的必不可少的一部分。在这样的要求下,本论文从多信息域分析的角度出发,将时域、频域、调制域、时频联合分析域等方面出发,构建一台以多域分析为目的的仪器。本论文主要从时频域的角度出发,以小波变换实现的时频分析为主线,从理论和实现两个方面,对小波变换这一主题进行阐述说明;同时将数字下变频技术与小波变换结合,实现对信号进行更加灵活的分析,具体的内容如下:1、对小波变换的原理说明,并对小波变换常用实现结构进行重现,对其实现方式进行评估,包括资源消耗、计算速度、分析精度等各方面进行对比。2、对小波变换的实现方式进行改进。小波变换在实现时资源消耗大、计算时间长是其主要的限制因素,融合利用常用的小波变换实现方式,优化其实现的资源消耗,并与常用的实现结构进行效果对比。3、将改进的小波变换实现方案在FPGA上实现,对参数的量化结果进行参数分析,并对FPGA上实现的小波变换进行精度对比。4、将数字下变频技术与小波分析结合,通过数字下变频技术改变信号频率与采样率之间的关系,能够改变不同分析频点的时域分辨率和频域分辨率,使得小波变换效果能够更加灵活。数字下变频技术的实现利用了CORDIC算法对信号进行频谱搬移,利用CIC滤波器实现高效的多倍率抽取滤波,利用半带滤波器的系数特点以及过渡带特性实现两倍抽取滤波,最终实现信号的数字下变频功能。将数字下变频后的信号进行小波变换,可以实现改变分析频点的时频分辨率,增大小波变换分析的时间长度。通过对上述问题的研究,实现了基于小波变换的信号时频联合分析功能;同时将数字下变频技术与小波变换结合,实现了对信号更加灵活有效的时频分析。更多还原

【Abstract】 The modern electronic information technology is developing towards a high-speed and wide-bandwidth direction,it continually improves its requirement for systematic index,meanwhile,it also puts forward a new request for comprehensive analysis of signals.Only from time domain or frequency domain to analyze cannot meet peoples’ extraction of signal characteristics.This study will start from a mixed domain instrument to slove the limitation that the traditional oscilloscopes just collect,resume and analyze signals from a single signal domain,and it can simultaneously analyze the time domain,frequency domain and time-frequency joint domain to form a multidimensional measuring instrument to measure signals.Therefore,multi-domain analysis is an indispensable part of current signal characteristic analysis.Under this requirement,from the perspective of multi-information analysis,this study aims to construct a multi-domain analysis instrument from the time domain,frequency domain,modulation domain and time-frequency joint analysis domain.This study’s topic mainly from the time and frequency domain,regarding the time-frequency analysis through the wavelet transform as its main line,will state the theme of the wavelet transform from the aspects of theory and implementation,at the same time,it also will combine the digital down-conversion with the wavelet transform to analyze signals more flexibly.The concrete contents are following:1.Stating the theory of the wavelet transform and reproducing the implementation structure commonly used by the wavelet transform,evaluating its model of implementation,including the comparisons among the resource consumption,calculation speed and analysis accuracy and so on.2.Improving the implementation model of the wavelet transform.The main limited factors of implementing the wavelet transform is its large resource consumption and long calculating time.Mixing and utilizing the commonly used methods of the wavelet transform,optimizing its resource consumption of realization,and comparing its effects of realization structure with that of commonly used.3.Implementating the improved scheme of the wavelet transform on the FPGA.Analyzing the quantization results of partial parameters and contrasting the the accuracy of the wavelet transform realized on the FPGA.4.Combining the digital down-conversion and the wavelet analysis can change the relationship between the signal frequency and the sampling frequency through the digital down-conversion;the time-domain and the frequency-domain resolution of different analysis frequency points can be changed as well,and make the the analysis of the wavelet transform more flexible.The realization of the digital down-conversion utilizes the CORDIC algorithm to shift the frequency spectrum of signals;uses the CIC filter to realize the high-efficiency,multiplicity-ration decimation filter;takes advantage of the coefficient characteristics of half band filter and the transition characteristics to achieve double decimation filter,finally the function of digital down-conversion will be achieved.The wavelet transform of digital down-conversion can change the time-frequency resolution of analyzing frequency points and increase the time of the wavelet transformat analysis.Through investigating the above questions,Realize the time-frequency joint analysis function based on the wavelet transform,in the meanwhile,through the combination of the digital down-conversion and the wavelet transform to achieve more flexible and efficient time-frequency analysis of signals.更多还原