ai_RayCh

Snake Algorithem,Snake Method Example

此文件包含三个MATLAB程序，用于对线性系统AR模型的权向量估计：Kalman滤波估计，RLS和LMS，均为较常用的估计方法，并附有相应的仿真图(There are three programs in this files:kalman,RLS and LMS to pridict the victors of linear systems . These three wawys are common and useful in signal processing)

说明：  基于RLs的自适应滤波器的设计 仿真通过(Adaptive filter based on RLs)

对测量专业中不同时期观测的数据进行精度是否相同进行统计检验(The measurement data of observation in different periods in the professional statistical test accuracy are the same )

非二位元LDPC 16 CODE錯誤率 非二位元LDPC 16 CODE錯誤率(Non-two yuan LDPC 16 CODE error rate of non-two yuan LDPC 16 CODE error rate)

说明：  利用lanczos输注特征值，特征向量，尽请指正(Lanczos infusion using eigenvalue)

实现模糊动态神经网络的一种matlab程序，直接粘贴到命令窗口可运行(a program could realize DFNN in matlab)

(有源代码)数值分析作业,本文主要包括两个部分,第一部分是常微分方程(ODE)的三个实验题,第二部分是有关的拓展讨论,包括高阶常微分的求解和边值问题的求解(BVP).文中的算法和算例都是基于Matlab计算的.ODE问题从刚性(STIFFNESS)来看分为非刚性的问题和刚性的问题,刚性问题(如大系数的VDP方程)用通常的方法如ODE45来求解,效率会很低,用ODE15S等,则效率会高多了.而通常的非刚性问题,用ODE45来求解会有很好的效果.从阶次来看可以分为高阶微分方程和一阶常微分方程,高阶的微分方程一般可以化为状态空间(STATE SPACE)的低阶微分方程来求解.从微分方程的性态看来,主要是微分方程式一阶导系数大的时候,步长应该选得响应的小些.或者如果问题的性态不是太好估计的话,用较小的步长是比较好的,此外的话Adams多步法在小步长的时候效率比R-K(RUNGE-KUTTA)方法要好些,而精度也高些,但是稳定区间要小些.从初值和边值来看,也是显著的不同的.此外对于非线性常微分方程还有打靶法,胞映射方法等.而对于微分方程稳定性的研究,则诸如相平面图等也是不可缺少的工具.值得提出的是,除了用ode系类函数外,用simulink等等模块图来求解微分方程也是一种非常不错的方法,甚至是更有优势的方法(在应用的角度来说).((Source code) numerical analysis homework, this docment includes two parts, the first is ordinary differential equations (ODE) of the three examples, the second part is about the expansion of the discussion, including the higher-order ordinary differential & boundary value solution Problems (BVP). the text of the algorithm and numerical examples are based on the Matlab. ODE from the rigid (STIFFNESS) look into the issue of non-rigid and rigid problem, rigid problems (such as large coefficient VDP equation) such as using the always method ODE45 used to solve the problems , efficiency will be low, with ODE15S the other hand, many of the high efficiency. and the usual problem of non-rigid, there will be used to solve ODE45 very good results. Judging from the order can be for high-order differential equations and first-order ordinary differential equations, higher-order differential equations can be transformed into a general state space (STATE SPACE) used to solve the low-order different)

标准的子空间跟踪算法，配合卡尔曼滤波算法，可用于移动用户跟踪(past-d subspace tracking )

利用matlab进行五点三次法平滑处理(时域和频域)(Using matlab 5.3 times France smoothing process (time domain and frequency domain))