INVERTER_3Ph

this program is the simulink of logistic map

神经网络理论与MATLAB7实现，对一些算法进行了实现，(Neural Network Theory and MATLAB7 achieved, a number of algorithms to achieve,)

说明：  matlab 6.5辅助神经网络设计原代码，有丰富详尽的示例。(neural network design supporting the original code, a rich detailed examples.)

Code Source DFIG Model

录制一段自己的语音信号，并对录制的信号进行采样；画出采样后语音信号的时域波形和频谱图；给定滤波器的性能指标，采用窗函数法和双线性变换设计滤波器，并画出滤波器的频率响应；然后用自己设计的滤波器对采集的信号进行滤波，画出滤波后信号的时域波形和频谱，并对滤波前后的信号进行对比，分析信号的变化；回放语音信号；最后，设计一个信号处理系统界面。(Section of their own voice recording signal, and recording the signal sampling draw sample after the speech signal time-domain waveform and frequency spectrum given filter performance indicators, using a window function method and bilinear transformation design of filters, and draw the frequency response filter then use the filter of their own design for the signal acquisition filtering, after filtering the signal to draw the time-domain waveform and spectrum, and filtering of signals before and after contrast, analysis of signal changes playback voice signal Finally, the design of a signal processing system interface.)

说明：  利用自相关函数估计基音周期，函数jiyinzhouqi，输入参数是分析语音文件名和帧移，帧移在5~10ms(use of the correlation function is estimated Pitch, jiyinzhouqi function, input parameters to analyze voice file names and frame shift, frame shift in the 5-10ms)

DCT变换：变换域系数选取采用阈值编码，：阈值越小，图像恢复效果越好，但压缩效果差，一般取在5-10 即可很好的获得恢复图像信息。(DCT transform: transform domain coefficients selected using the threshold coding: the smaller the threshold value, the better the image restoration, but the compression is poor, and generally in the 5-10 to get a good recovery image information.)

these are the DBMS ppt for basic concept of management system

计算大型的稀疏矩阵的本征值和本证函数，和eigs差不多(Large sparse matrix computing eigenvalues ​ ​ and functions of the card, and almost eigs)

add pink noise to signal with certain SNR