基于半参数化SLC 的雷达目标识别

针对雷达目标高分辨距离像(HRRP)统计识别中SLC(基于累计量的随机学习算法)在小样本情况下概率密度估计准确度下降的问题，该文提出一种基于半参数化SLC 的雷达目标识别方法。该方法利用半参数化概率密度估计思想对SLC 非参数化概率密度估计进行修正，有效利用了HRRP 各距离单元幅值近似服从 Gamma 分布的经验知识，达到参数化方法和非参数化方法优劣互补以提高小样本情况下雷达目标识别率的目的。最后基于5 种飞机模型HRRP 数据的仿真实验证明了该方法的有效性。 

Radar Target Recognition Based on Semiparametric Density Estimation of SLC

In order to solve the problem of the decline of accuracy when using the nonparametric methodStochastic Learning of the Cumulative (SLC) to estimate the density of High-Resolution Range Profile (HRRP) in radar target recognition under the condition that the samples are not enough, a radar target recognition approach based on the semiparametric density estimation of SLC is proposed in this paper. This method has the ability to make use of empirical knowledge which is known as the approximate Gamma distribution of amplitudes in each HRRP range cells, and the Gamma density estimate is then corrected by multiplying with SLC of a correction factor. Obviously, both advantages of parametric method and nonparametric method of SLC are merged in the semiparametric density estimation of SLC. Simulation results based on the HRRP dataset of five aircraft models demonstrate the effectiveness of the proposed approach.