PCAandKPCAprogram

kernel bandwith optimization

一个比较典型的PID处理程序，在使用单片机作为控制cpu时，请稍作简化，具体的PID 参数必须由具体对象通过实验确定。由于单片机的处理速度和ram资源的限制，一般不采用浮点数运算， 而将所有参数全部用整数，运算到最后再除以一个2的N次方数据（相当于移位），作类似定点数运算，可 大大提高运算速度，根据控制精度的不同要求，当精度要求很高时，注意保留移位引起的“余数”，做好余 数补偿。这个程序只是一般常用pid算法的基本架构，没有包含输入输出处理部分。 (A typical PID process, using the microcontroller as a control cpu, make some simplification, the specific PID parameters must be determined experimentally by a specific object. As the chip processing speed and ram resource constraints, generally do not use floating-point operations, and all parameters of all integers, and divided by an operator to the last N-th power of 2 data (the equivalent of the shift) for a similar fixed-point operations, can greatly improve the speed of operation, according to the different requirements of precision control, when high precision, attention to retain the shift caused by the " remainder" and on the remainder of compensation. This procedure is commonly used basic structure pid algorithm, does not include input and output processing section.)

用matlab实现RLS滤波，最小二乘估计算法是以最小误差平方和为优化目标的。(RLS filter using matlab, least squares estimation algorithm is the smallest sum of squared errors for the optimization objectives.)

用matlab实现从原始视频中提取关键帧，并且自动生成新视频，新视频只由关键帧组成从而实现视频总结。(Using matlab realize extracted from the original video key frame, and automatically generate a new video, new video key frame composed by only enabling the video summary.)

一本详细介绍Matlab的书籍，很好很全面，科研必备书籍 (A detailed Matlab books, very good very comprehensive, scientific research must Books)

FGM薄板有限元程序（应力矩阵的简化程序）(failed to translate)

This program detetc the skin color and track using a web cam.

时频分析的工具，完成时频分析，希望能够对大家有帮助，共同进步！(Time-frequency analysis tools, complete the time-frequency analysis in hopes of help, make progress together!)

说明：  Matlab写得神经网络的训练实例，包括BP，RBF，SOFM等多种网络，适合初学者(err)

这个文章说明了格雷算法在心电信号处理上的应用(The paper proposes a learning algorithm for real-time analyzing the gray relations between templates and test electrocardiogram (ECG) signal patterns. Without analyzing relations between the test ECG pattern and every template in the database for the recognition, several training patterns for the learning algorithm are chosen from an ECG waveform database with patient diagnosis information, and then the training result is used to analyze the test ECG.)