Wavelet Analysis Software Of 1/f Noise And Research On Long-Range Dependence Of It

The noise inside the electronic devices (especially the low-frequency noise such as 1/f noise) is the key factor limiting the sensitivity and precision of the devices. And at the same time, it is the important sensitive parameter to characterize the quality and reliability of the devices. Therefore, parameters extraction and analyzing for interior noise of electronic device are of great significance in engineering practices.The main work and conclusions are as follows:1. According to the practical requirements of analysising the parameters of low-frequency electronic noise, the former software which fits all the parameters of low-frequency spectrum was optimized. We adopted the method that find the corner frequency- f 0-first, and, then fit the low-frequency spectrum containing g-r component. The 6 parameters are obtained precisely, the precision of this fitting software are 3 percents higher than the former one in our Lab.2. Wavelet is very suit for 1/f noise analysis. Although there is a wavelet analysis tool box in MATLAB, it's still not so convenient and feasible for professional analysis of electrical noise. We developed a wavelet analysis for electronic noise version 1-WAEN 0.1(Wavelet Analyzer for Electronic Noise)-based on MATLAB GUI. Some methods for analyzing noise based on wavelet theory are available on WAEN 0.1. There are friendly interface face to users are on the software as well, and, it's convenient to use.3. Long-range dependence is that the two points apart from distance are still correlated. Long-range dependence is one of the characteristics of 1/f noise. As there are long-range dependence on stock market, data of network communication and DNA series, by far, a lot of researches are mainly focused on these areas above. We analyzed 1/f noise using the methods which used on other area research. The most effected and stable method-DFA- is filtered out. Hurst exponent which is the main representation for intensity of long-range dependence is precisely estimated for 1/f noise.