Resonance-freq

提出了一种新的宽带DOA估计方法:频域子空间正交性测试方法(TOFS:Test of orthogonality of frequency sub- space)。该方法通过同时测试频域信号共同带宽内各频段噪声子空间与阵列流形之间的正交性来进行DOA估计。与宽带相干信号子空间方法不同,TOFS方法不需要任何初始值的预估及聚焦操作。与宽带非相干信号子空间方法也不同,TOFS方法同时测试各频段噪声子空间与阵列流形之间的正交性。本文仿真了TOFS与IMUSIC、CSSM、TOPS的性能比较。仿真结果表明TOFS方法在中等信噪比以上时有较好的性能,且避免了TOPS方法中常出现的伪峰。 (Proposed a new broadband DOA estimation methods: subspace orthogonal frequency domain test methods (TOFS: Test of orthogonality of frequency sub-space). The method by simultaneously testing the frequency domain signal bandwidth of each band noise common sub-orthogonality between space and the array manifold to be DOA estimates. And broadband coherent signal subspace different methods, TOFS method does not require any initial estimate and focusing operation values. And broadband incoherent signal subspace methods are different, TOFS method for simultaneous testing of each band orthogonal noise subspace and the array manifold between. This paper compares the performance TOFS simulation and IMUSIC, CSSM, TOPS is. Simulation results show that the method TOFS better performance when more moderate SNR and spurious peaks TOPS method avoids the often occur.)

说明：  将标准微粒群算法进行改进，从而解决标准PSO算法的容易陷入局部最优解的情况。(The standard particle swarm algorithm is improved so as to solve the standard PSO algorithm is easy to fall into local optimal solution of the situation.)

建立模糊控制的查询表,隶属度函数为高斯型(Fuzzy control query table)

通过预建的人脸图像库，实现人脸识别的matlab程序。(By pre-built library of face images, face recognition matlab program.)

说明：  灰色预测方法的特点表现在：首先是它把离散数据视为连续变量在其变化过程中所取的离散值，从而可利用微分方程式处理数据；而不直接使用原始数据而是由它产生累加生成数，对生成数列使用微分方程模型。这样，可以抵消大部分随机误差，显示出规律性。(The characteristics of the grey prediction method are as follows: First, it regards the discrete data as the discrete value of the continuous variable in its change process, so that the differential equation can be used to process the data; instead of using the original data directly, it generates the cumulative generation. Number, using a differential equation model for generating a series. In this way, most of the random errors can be offset, showing regularity.)

Exercise on kalman filter with model AR, ARMA, ecc.

source code for imageproccesin

神经网络训练程序、测试程序以及验证程序整体(Neural network training procedures, test procedures, and the overall validation process)

说明：  MATLAB/simulink光伏电池建模！！！！(Photovoltaic cell modeling)

说明：  这是一个国外的DST的Matlab工具箱，很全面的反映了dempster-shafer理论，是学习DST的权威工具箱(This is a foreign DST toolbox of Matlab, it is a comprehensive reflection of dempster-shafer theory, is to learn from the authority of DST Toolbox)