遗传算法在高频区目标二维散射分析中的应用

Application of genetic algorithm in high frequency 2-D scattering signatures analysis

针对高频区雷达目标的散射特征问题,提出利用联合时频分析———自适应高斯基表示方法(AGR)来简化目前的二维散射中心模型,并将遗传算法(GA)用于二维散射模型参量的估计.该方法充分利用遗传算法的内在并行性,解决了二维等效散射中心的估计问题.仿真结果表明,该简化的二维模型不仅可以区分局域性散射中心,也可以再现具有一定长度的展布式散射结构,利用该简化模型在无噪情况下可以准确估计各散射中心的参数,同时在低信噪比情况下也可以取得很好的估计性能,它对于去除背景噪声起到了很好的作用.通过验证,在简化的散射模型基础上的目标散射特征分析是准确和可行的.

To investigate the scattering features of radar target in high frequency, a joint time-frequency analysis method---the adaptive Gaussian basis representation (AGR) was introduced to simplify the 2-D scattering parameter model. To estimate the model parameters, Genetic algorithm (GA) was proposed. This method takes full advantage of GA's inherent parallelism and solves the estimation problem of the equivalent scattering center. The simulation results show that this simplified 2-D scattering model may not only discriminate localized scattering centers of the radar target, but also present distributed scattering centers of definite length. In terms of the extraction of scattering features using this 2-D scattering model, and this method can estimate all parameters of scattering centers correctly on condition of no noise, at this time it has a good performance of estimating the position of initial scattering centers and eliminating the noise in the low signal to noise rate (SNR). The analysis using the above 2-D scattering model is correct and feasible.