PSKCreate

FDTD 的吸收边界条件讲解，详细讲解了边界条件的推导过程。(Absorbing boundary conditions for numerical simulation of waves.)

等间距线性阵列，阵元间距d= λ /2，等幅加权，目标方位φr = 0°，无噪声。 分别仿真计算阵元数M=8 、16时的阵列方向图，定性分析产生的原因。（可通过3dB波束宽度计算公式说明） (Equally spaced linear array, the array element spacing d = λ/2, equal amplitude weighted target azimuth φr = 0, no noise. , Respectively, the number of array elements M = 8 and 16 when the array pattern, qualitative analysis to produce simulation. (Available by 3dB beamwidth formula noted))

Multiple instance metric learning automatically labeled bags of faces

直接序列扩频（Direct Sequence Spread Spectrum）工作方式，简称直扩方式（DS方式）。就是用高速率的扩频序列在发射端扩展信号的频谱，而在接收端用相同的扩频码序列进行解扩，把展开的扩频信号还原成原来的信号。(irect Sequence Spread Spectrum (Direct Sequence Spread Spectrum), hereinafter referred to as Direct expansion mode (DS). Is a high rate of spread spectrum sequence extension in signal spectrum, and at the receiving end with the same spread spectrum code sequence for expansion, the expansion of the spread spectrum signal back into the original signal. )

这是计算dihedral的散射程序，对研究电磁散射的同行很有用(This is the calculation of the dihedral scattering process, the study of electromagnetic scattering useful peer)

说明：  matlab源程序代码，应用于非线性方程求根等(Roots of nonlinear equations)

matlab time frequency tfrcw.m

说明：  介绍了几个用于弱信号捕获的不同的信号处理方法，并对其捕获性能进行了比较。(A number of differential signal processing schemes have been proposed to address weak signal processing.)

simple watermark test

相关向量机(RVM)的matlab源程序，包含快速算法，内含代码使用说明。 RVM采取是与支持向量机相同的函数形式稀疏概率模型，对未知函数进行预测或分类。 优点： (1) 不仅仅输出预测目标量的点估计值,还可以输出预测值的分布. (2) 使用更少数量的支持向量,从而显著减少输出目标量预测值的计算时间. (3) RVM不需要估计过多的参数. (4) RVM对是否满足Mercer 定理的核函数没有限制,适应性更好(Relevance Vector Machine (RVM) of the matlab source code, including fast algorithm that contains code for use. RVM support vector machine is taken the same functional form sparse probabilistic model to predict the unknown function or classification. Advantages: (1) is not only the amount of output predicted target point estimates, but also the distribution of predicted values ​ ​ can be output. (2) using a smaller number of support vectors, thereby significantly reducing the output target amount predicted value calculation time. (3) RVM does not require excessive parameter estimation. (4) RVM meets Mercer theorem on the kernel function is not limited, and better adaptability)