UKF

Absorbing Boundary Conditions in Free Space

LU factorization of a square matrix using *no row exchanges*.

说明：  先进PID控制及其matlab仿真——配套程序(Advanced PID Control and matlab simulation- complete program)

matlab的经典算法和讲义，包括常用的算法，并附带程序(matlab classic algorithms and handouts, including the commonly used algorithm, and comes with the program)

蜂群算法计算适应度函数，它有利于初学者对人工蜂群算法的学习(Colony algorithm fitness function, it is conducive to beginners learning artificial bee colony algorithm)

MATLAB2016b实现FFT快速傅里叶变换(MATLAB implementation of FFT fast Fourier transform)

神经网络中用矩阵分块的方法处理数据量过大的情况，效果很好(Neural network is used in the method of matrix blocking deal with the situation of the large amount of data, the effect is very good )

《visual c++matlab图像处理与识别实用案例精选》源码(" Visual c++ matlab image processing and recognition of practical cases selected" source)

code for the mpeg decoder.

该算法用于自回归输入模型，是一种迭代的算法。其基本思想是基于对数据先进行一次滤波处理，后利用普通最小二乘法对滤波后的数据进行辨识，进而获得无偏一致估计。但是当过程的输出信噪比比较大或模型参数较多时，这种数据白色化处理的可靠性就会下降，辨识结果往往会是有偏估计。数据要充分多，否则辨识精度下降。模型阶次不宜过高。初始值对辨识结果有较大影响。(The algorithm used for autoregressive input model, it is a kind of iterative algorithm. The basic idea is based on data to conduct a filtering processing, after using ordinary least square method to identify the data filter, and then obtain unbiased consensus estimates. But when the process output signal-to-noise ratio is larger or model parameters are too, this kind of data white processs reliability will drop, identification results tend to be biased estimate. Data should fully, otherwise the identification accuracy down. Model order time shoulds not be too high. Initial value to identification results have great influence on.)