juxingbodao

Ship navigation project

haar二维小波分解，输出四个分解系数子矩阵.将系数重构，得到最终的输出矩阵(haar two dimension wavelet decompositon)

说明：  它主要研究的是多因变量对多自变量的回归建模，特别当各变量内部高度线性相关时，用偏最小二乘回归法更有效。另外，偏最小二乘回归较好地解决了样本个数少于变量个数等问题。 偏最小二乘法是集主成分分析、典型相关分析和多元线性回归分析3种分析方法的优点于一身。它与主成分分析法都试图提取出反映数据变异的最大信息，但主成分分析法只考虑一个自变量矩阵，而偏最小二乘法还有一个“响应”矩阵，因此具有预测功能。(It is mainly a result of the study is a multi-variable regression modeling and more independent variables, especially when the internal height of each variable linear correlation, partial least squares regression method is more effective. In addition, partial least squares regression solves the number of samples is less than the number of variables and other issues. Partial least squares method is a set of principal component analysis, canonical correlation analysis and multiple linear regression analysis of the advantages of the three methods in one. It is the principal component analysis are trying to extract the maximum information reflects data variability, but the principal component analysis considers only one independent variable matrix, and partial least squares and a &quot response&quot matrix, predictive capabilities.)

Its IEEE papers on papr reduction in OFDM

一个简单的混沌程序，可以看到分叉的现象和振荡的现象。(A simple chaotic process, you can see the phenomenon of bifurcation and oscillation)

这个Matlab程序实现最速下降算法。最速下降法是一种最基本的算法，它在最优化方法中占有重要地位.最速下降法的优点是工作量小，存储变量较少,初始点要求不高；缺点是收敛慢，最速下降法适用于寻优过程的前期迭代或作为间插步骤，当接近极值点时，宜选用别种收敛快的算法. (Matlab implementation of this procedure steepest descent algorithm.)

Design a linear phase minimum order FIR Band Stop filter to stop the frequency band 0.45π to 0.55π with atleast 60dB attenuation at the edges. Further -3dB frequencies are required to be at 0.3π and 0.7π. 1. Plot the frequency response of the psuedo magnitude function using (a)rectangular window, (b)Hann window and (c) Hamming window. 2. Plot the frequency response for Rectangular and Hamming window for a length of 1024 to demonstrate the fact that the increase in the length, doesn t reduce undershoot and overshoot in Rectangular window.

录制自己的一段语音，时间控制在15秒到30秒左右；利用wavread函数对自己的语音进行采样，记住采样频率。 （1）画出原始语音信号的时域波形，而后以1秒为间隔，求出每秒数据的功率谱。 （2）根据语音信号特点，分别设计FIR及IIR滤波器，分别画出滤波器幅频和相频特性曲线。用设计的滤波器对信号滤波，画出滤波后时域波形。用sound函数回放语音信号。 （3）求出特征频段语音信号随时间变化的曲线（每间隔0.05秒求一次功率谱）。 (Record your own voice for some time (15 seconds to 30 seconds) use wavread function of the voice samples, remember the sampling frequency. (1) Draw the original speech signal in time domain waveform, and then to 1 second intervals, the data obtained power spectrum per second. (2) According to the voice signal characteristics, FIR and IIR filters were designed, namely to draw the amplitude and phase frequency characteristic curve. With the design of the filters on the signal filtering, to draw the filtered time-domain waveform. Playback with sound function of the speech signal. (3) Find the characteristic band speech signal curve over time (0.05 seconds intervals seeking a power spectrum.))

FCEUX的汉化版。希望大家积极下载！希望大家支持我！(FCEUX the finished version. Hope that we actively download! Hope you will support me!)

根据姚天任 现代数字信号处理 第四章，那个课后作业题编的，搞了好久，结果很理想，如果能搞成，1024点，结果会十分稳定。(music method)