混沌海杂波背景下的微弱信号检测混合算法

发布时间：2018-04-05 23:12

本文选题：粒子群算法　切入点：支持向量机　出处：《物理学报》2015年04期

【摘要】：基于经验模态分解理论,提出了一种基于粒子群算法的支持向量机预测方法.采用总体平均经验模式分解法将混沌信号分解为若干固有模态函数和趋势分量,将复杂的非线性信号转化为具有不同尺度特征的平稳分量.利用粒子群算法对支持向量机的惩罚系数和核函数进行优化,结合支持向量机建立混沌序列的单步预测模型.从预测误差中检测淹没在混沌背景中的微弱信号(包括瞬态信号和周期信号).对Lorenz系统和实测IPIX雷达数据进行仿真实验,结果表明,该方法能够有效地从混沌背景噪声中检测出微弱目标信号,Lorenz系统得到的均方根误差0.000000339(-102.8225 d B时)比传统支持向量机方法的均方根误差0.049(-54.60 d B时)降低了5个数量级,从海杂波中检测出具有谐波特性的微弱信号,表明预测模型具有更低的门限和误差.
[Abstract]:Based on empirical mode decomposition theory, a support vector machine (SVM) prediction method based on particle swarm optimization (PSO) is proposed.The chaotic signal is decomposed into some inherent mode functions and trend components by the method of population average empirical mode decomposition, and the complex nonlinear signals are transformed into stationary components with different scale characteristics.Particle swarm optimization algorithm is used to optimize the penalty coefficient and kernel function of support vector machine (SVM), and a one-step prediction model of chaotic sequence is established by combining support vector machine (SVM).Weak signals (including transient signals and periodic signals) submerged in chaotic background are detected from prediction errors.The simulation results of Lorenz system and IPIX radar data show that,This method can effectively detect the weak target signal from chaotic background noise and the root mean square error obtained by Lorenz system is 0.000000339- 102.8225 dB), which is 5 orders of magnitude lower than that of traditional support vector machine (SVM) when the root mean square error is 0.049 ~ 54.60 dB).Weak signals with harmonic characteristics are detected from sea clutter, which indicates that the prediction model has lower threshold and error.
【作者单位】：南京信息工程大学气象灾害预报预警与评估协同创新中心江苏省气象探测与信息处理重点实验室;
【基金】：国家自然科学基金(批准号:61072133) 江苏普通高校研究生实践创新计划(批准号:SJZZ_0112) 江苏省产学研联合创新资金计划(批准号:BY2013007-02,BY2011112) 江苏省高校科研成果产业化推进项目(批准号:JHB2011-15) 江苏省“信息与通信工程”优势学科和江苏省“六大人才高峰”计划资助的课题~~
【分类号】：TN911.23;O415.5

【相似文献】