fusechuli

说明：  matlab与科学计算的教程编写的课堂课件(MATLAB and science calculational)

wind turbine modelling

利用BP网络对非线性系统进行辨识，同时，利用模型参考自适应方法，设计NNMARC控制器(BP network identification of nonlinear systems, at the same time, the use of model reference adaptive method, design NNMARC controller)

MD_MUSIC,多维MUSIC算法，测向算法，有工程意义，但运算量大。(MD_MUSIC，DOA，computational expense)

说明：  网上找的一个有意思的程序。共享下。参阅下子(Internet is an interesting program. Sharing the next. See erupted)

光伏电池阵列输出功率受光照强度和温度变化的影响，因此最大功率点跟踪（MPPT）技术广泛应用于光伏系统中。在所有最大功率点(MPP)控制策略中，扰动观察（P&O）MPPT算法因易实现被广泛应用，然而它的缺点是在稳定工作状态下工作点通过MPP 时会导致能量振荡损耗，并且在光照强度或温度发生突变时表现较差的动态响应。在本文中，提出一种改进型变步长扰动观察MPPT 算法，此方法依据工作点动态调整步长变化，与传统固定步长方法比较，本文提出的方法能有效地提高MPPT 速度和转换效率，通过仿真和实验结果分析，验证了此改进算法的可行性。(PV array output power by light intensity and temperature changes, so the maximum power point tracking (MPPT) technology is widely used in photovoltaic systems. In all the maximum power point (MPP) control strategy, the perturbation and observation (P & O) MPPT algorithm because easy to implement is widely used, but its disadvantage is the steady-state operating point by the MPP causes energy oscillations loss, and in the light intensity or temperature mutated poor performance dynamic response. In this paper, we propose a modified variable step size perturbation and observation MPPT algorithm, this method is based on dynamically adjusting the operating point changes in steps, with the traditional fixed-step methods, the proposed method can effectively improve the speed and efficiency MPPT, The simulation and experimental results verify the feasibility of this algorithm.)

用于图像处理，可以实现去燥功能，还可以加噪等(For image processing, can function to dry, but also noise and other)

用于对基于空间点的三维重建,非常好用,请多多指教,QQ:526289(for space-based point of the three-dimensional reconstruction, very handy, exhibitions, QQ : 526289)

说明：  This provides good overview for post-processing by Matlab

这个是用于两类的带有类别的数据的线性分类器，是结构体类型如下判别 y(i) = 1 if W *X(:,i) + b >= 0 2 if W *X(:,i) + b < 0(his function classifies input data X using linear discriminant function: y(i) = argmax W(:,y) *X(:,i)+ b(y) y where parameters W [dim x nfun] and b [1 x nfun] are given in model and nfun is number of discriminant functions. In the binary case (nfun=1) the classification rule is following y(i) = 1 if W *X(:,i)+ b >= 0 2 if W *X(:,i)+ b < 0)