基于多参数组合优化的微弱信号检测方法

摘要

针对传统检测方法对混沌背景下微弱信号检测能力不足的问题，提出了基于多参数组合优化的微弱信号检测方法。该方法利用相空间重构和支持向量机(SVM)模型参数间的相互依赖、相互制约的关系，将相空间重构和 SVM 模型参数作为遗传算法(GA)的个体，以混沌时间序列预测均方根误差的倒数作为适应度函数，采用实数编码，通过循环迭代获得最优参数值进行建模、训练和预测，从一步预测误差中将混沌背景中微弱信号(包括瞬态信号和周期信号)检测出来。分别以 Lorenz 系统和实测的 IPIX 雷达数据作为混沌背景噪声进行实验研究，仿真验证表明，所提方法能够有效地将混沌背景噪声中瞬态信号和周期信号检测出来，与传统的参数求取方法相比，预测精度和检测性能均得到显著提高。%Aiming at the poor performance of weak signal detection of the traditional detection method in chaotic background,a weak signal detection method was proposed based on muti-parameter combination optimization. The phase space reconstruction parameters and SVM model parameters were taken as genetic algorithm individu-als by using the relationship of the interdependence and mutual restraint between them,while the reciprocal of the chaotic time series prediction root mean square error was taken as the fitness function of genetic algorithm, and the real number coding was adopting.Modeling,training and predicting by utilizing the obtained optimal pa-rameter values which were obtained through the loop iteration method,then the weak signal including transient signal and periodic signal embedded in the chaotic background were detected from the one-step predictive error. Lorenz system and IPIX Sea Clutter was used as chaotic background noise for experimental research respective-ly,simulation results showed that this method was highly effective to detect weak transient signal and periodic signal from a chaotic background.Compared with traditional parameters optimization algorithm,the prediction accuracy and detection performance of the new method was improved significantly.