Subspace_Learning_Codes

Loop Transfer Recovery control design

说明：  本程序运行需一个小时左右。优化准则是旁瓣电平最小。取栅格间距为lemda/2,阵元间距为栅格间距的整数倍。放阵元的位置放1，否则放0(This program should be running an hour or so. Rule of optimization is side-lobe level minimum. Take grid spacing for lemda/2, spacing for grid spacing array (integer times. Put the position of the array (put 1, otherwise put 0 )

beam element is a finite element code for assembling of stiffness and mass matrix of bernulli beam with two dof at each node

说明：  基本的图像处理技术在MATLAB里得到非常好的应用(The basic image processing technology in the years to be a very good application of MATLAB)

一 维 时 域 有 限 差 分 法 MATLAB 代 码(One-dimensional finite difference time domain method MATLAB code)

我们用频率采样法设计具有线性相位的数字低通滤波器。(We use the frequency sampling design method of digital filter with linear phase. )

INTRODUCTION FOR MODEL PREDECTIVE

说明：  关于管道流动及充填问题中，为求解时间、压力等的常微分方程组的求解。(About the pipeline flow and filling problems, in order to solve the time, pressure, etc. for solving ordinary differential equations.)

SVM方法的基本思想是：定义最优线性超平面，并把寻找最优线性超平面的算法归结为求解一个凸规划问题。进而基于Mercer核展开定理，通过非线性映射φ，把样本空间映射到一个高维乃至于无穷维的特征空间（Hilbert空间），使在特征空间中可以应用线性学习机的方法解决样本空间中的高度非线性分类和回归等问题。svm 程序，即支持向量机的代码。(The basic idea of SVM method are: the definition of the optimal linear hyperplane, and the search algorithm for optimal linear hyperplane by solving a convex programming problem. Then based on Mercer nuclear expansion theorem, through a nonlinear mapping φ, the sample space is mapped to a high-dimensional and even infinite dimensional feature space (Hilbert space), so that in the feature space can be applied to solve the linear learning machine method, the sample space The highly nonlinear classification and regression problems. svm procedures that support vector machine code.)

说明：  对rosenstein求解最大李雅普诺夫指数算法,实现成功(Rosenstein solving of the largest Lyapunov exponent algorithm, to achieve successful)