kalmanf

说明：  多极坐标搜素，一个局部与全局最优的兼容方法(Multi-polar search element, a local method is compatible with the global optimum)

等间距线性阵列，阵元数M=8，阵元间距d= λ /2，等幅加权，目标方位φr =0°。产生空间白噪声矢量（复高斯分布）。 设阵列信号矢量元素s(m)=exp[j mψr]，其信号幅度为1，该阵元接收机附加的高斯噪声为n(m)=nmr+jnmi，其中实虚部均为独立同分布N(0, σ2)的高斯随机数，则该通道合成信号x(m)=s(m) + n(m)，其中信噪比为： SNR = 10 lg[1/(2σ2)] = – 3 – 10 lg(σ2) (dB) (The spacing of the linear array, and the array element number M = 8, the array element spacing d = λ/2, equal amplitude weighted target azimuth φR = 0. Space generated white noise vector (complex Gaussian distribution). Let array signal vector elements s (m) = exp [j mψr], the signal amplitude, the element receiver additional Gaussian noise n (m) = nmr+jnmi which the real and imaginary parts are independent and identically distributed N ( Gaussian random numbers of 0, σ2), the channel the synthesis signal x (m) = s (m) the PPLS n (m), where SNR is: SNR = 10 lg [1/(2σ2)] =- 3- 10 lg (σ2) (dB))

坐标系转变，通过运行，很好的代码，哈哈哈(Coordinate system transformation, by running, a good code)

stationary test 檢驗數據是否為穩定 藉此判定ARIMA程序中的d值(stationary test)

小波模极大源代码，可用来进行小波降噪或小波模态参数识别，该代码为斯坦福大学源码。(wavelet modulus maxima source code, which can be used for wavelet denoising or wavelet modal parameter identification, The code for the source of Stanford University.)

一阶AR模型使用LMS算法，使用M=2，M=3两种滤波器，步长已选择(A first-order AR model uses the LMS algorithm, M = 2, m = 3 two filters, the step size has been selected)

有源电力滤波器的仿真模型，matlab中simulnk建模，测试可用。(Active power filter APF simulnk model)

modulation par déplacement de phase multi- niveaux MPSK

An implementation of Lucas-Kanade optical flow method for 3-D images(This is an implementation of Lucas-Kanade optical flow method for three dimensional images like NIFTI, DICOM etc. A demo with test dataset is given.)

说明：  利用matlab提取出频域和时域信号的29个特征，主运行文件feature_extraction，fre_statistical_compute和time_statistical_compute分别提取频域和时域的特征，生成的29个特征保存在生成的feature矩阵中。(Using MATLAB to extract 29 features of frequency-domain and time-domain signals, the main running files feature extraction, frequency-domain and time-domain features are extracted respectively, and the 29 features are saved in the generated feature matrix.)