我课设编写的非凸优化算法调试工具

AOA——toa角度估计算法仿真，基于MUSIC的(AOA-toa angle estimation algorithm simulation, based on MUSIC)

阵列信号处理的高分辨率估计，数学方法是部分子空间法，本科毕设要求参见标准测试模型，验证可用，各种kalman滤波器的设计，最终的权值矩阵就是滤波器的系数。

2代小波示意程序 2维小波变换经典程序 Daubechies小波基的构造 采用多孔trous算法(undecimated wavelet transform)实现小波变换 平移变换平移法(cycle_spinning)消除gibbs效应 提升法97经典程序 消失矩作用的程序 小波插值与小波构造 小波滤波器构造和消噪程序 小波谱分析mallat算法经典程序-2 Generation Wavelet indicated procedures classic 2-D wavelet transform procedures for Daubechies wavelet bases constructed using porous trous algorithm (undecimated wavelet transform) the realization of wavelet transform translational translational transform method (cycle_spinning) to eliminate Gibbs effect of lifting 97 vanishing moments the role of the classic procedure of wavelet interpolation procedure value structure of wavelet and wavelet de-noising filter structure and procedures of small Mallat algorithm for spectral analysis of the classical procedures

经典算法--距离多普勒算法，在MATLAB上编程实现，有很广的实用性-Classical algorithm- Range-Doppler algorithm, in MATLAB programming on implementation, there is a wide range of practical

8FSK系统，波形生成，相干解调，非相干解调等(8FSK systems, waveform generation, coherent detection, non-coherent demodulation, etc.)

基于方差一致性准则对图像进行四叉树分解！(variance quadtree)

在详细介绍各种命令的同时,向读者介绍了MATLAB在高等数学,线性代数,图形图像处理等方面的应用和外部接口.(In detail the various commands at the same time, introduced to readers MATLAB in higher mathematics, linear algebra, graphics, image processing and other applications and external interfaces.)

对时间序列或者信号进行fft谱分析，可以打开任意数据文件后分析，很方便使用。(FFT transfer of matlab code,it is useful for signal analysis.)

用于运动捕获数据降维后寻找极值点，可以找到极大值以及极小值点(Dimensionality reduction for motion capture data can be extracted extreme points, keyframe extraction or motion segmentation)

将整个数据集中的103个样本随机划分为训练集与测试集，其中训练集包含80个样本，测 试集包含23个样本； 建立极限学习机模型，并训练； 利用训练好的极限学习机模型对测试集中的23个样本进行预测； 输出结果并绘图（真实值与预测值对比图）； (The 103 random samples of the entire data set is divided into training set and test set, wherein the training set contains 80 samples, measuring Test set contains 23 samples Establish ELM model, and training Limit the use of the trained learning machine model test set of 23 samples to predict Output and graphics (real and predicted values comparison chart) )