DownSample

思维进化算法优化BP神经网络——非线性函数拟合(Mind Evolutionary algorithm optimization BP neural network- a nonlinear function fitting)

说明：  模型预测电流控制其控制思想基于以下原理：首先，功率变换器产生的开关状态个数有限，且通过建立模型可以预测每一个开关状态下被控变量的对应变化特性；其次，定义价值函数评价被控变量的预测值，为判别最优有限开关状态提供依据；最后，计算出每一种开关状态下的价值函数值，以此评价每一种开关状态下被控变量预测值的优劣，选择出最优的开关状态并应用。(The control idea of model predictive current control is based on the following principles: firstly, the number of switching states produced by power converter is limited, and the strain characteristics of controlled variables under each switching state can be predicted by establishing a model; secondly, the value function is defined to evaluate the predicted values of controlled variables, which provides a basis for judging the optimal finite switching state; lastly, the optimal finite switching state is calculated. The value function value under switching state is used to evaluate the predicted value of controlled variables under each switching state, and the optimal switching state is selected and applied.)

This code is used for modeling frequency response and phase change of accelerometer.

matlab数学手册，里面是这本书的word文件和源代码(Matlab mathematics is the book handbook, the word document and the source code)

用于滚动轴承微弱故障信号的诊断的方法：最小熵反卷积程序、自适应消噪法处理轴承信号、美国西储大学轴承数据处理等5个源文件(Fault diagnosis for rolling bearing weak signal approach: minimum entropy deconvolution procedure, adaptive noise cancellation treatment bearing signals, Case Western Reserve University, United States bearing data processing five source files)

说明：  Diversity Techniques over Fading Channels Selection Combining

说明：  这是一个关于盲源分离独立成分分析方法，效果很好(Here is a blind source separation independent component analysis)

改变Simulink中Scope输出波形的颜色(Change the color of the output waveform in Simulink Scope)

矢量控制系统中的PI调节器，此为磁链调节器(PI regulator vector control system, this flux regulator)

QR分解SMI算法的目的正是要避免直接来解线性方程，而是将自相关矩阵分解，并利用Givens旋转实现数据矩阵的QR分解，最终将权矢量的求解问题转化为三角线性方程组的求解问题。(QR decomposition SMI algorithm, whose objective it is to avoid directly to solution of linear equations, but will auto-correlation matrix decomposition, and the use of Givens rotation implementation of the QR decomposition of data matrix, the weight vector will eventually solve the problem into a triangular system of linear equations solving the problem.)