系统工程与电子技术 ›› 2019, Vol. 41 ›› Issue (9): 1984-1989.doi: 10.3969/j.issn.1001-506X.2019.09.10

• 传感器与信号处理 • 上一篇    下一篇

基于AEPSO-SVM算法的雷达HRRP目标识别

王彩云1, 黄盼盼1, 李晓飞2, 王佳宁2, 赵焕玥1   

  1. 1. 南京航空航天大学航天学院, 江苏 南京 210016;
    2. 北京电子工程总体研究所, 北京 100854

  • 出版日期:2019-08-27 发布日期:2019-08-20

Radar HRRP target recognition based on AEPSO-SVM algorithm

WANG Caiyun1, HUANG Panpan1, LI Xiaofei2, WANG Jianing2, ZHAO Huanyue1   

  1. 1. College of Astronautics, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China; 2. Beijing Institute of Electronic System Engineering, Beijing 100854, China

  • Online:2019-08-27 Published:2019-08-20

摘要:

针对雷达自动目标识别中的高分辨距离像(high resolution range profile,HRRP)识别问题,提出自适应进化粒子群(adaptive evolution particle swarm optimization, AEPSO)算法优化支持向量机(support vector machine,SVM)的目标分类识别方法。该算法利用非线性自适应惯性权重的调整以适应粒子寻优的非线性变化过程,采用分阶段调节加速因子增强粒子在进化过程中的学习能力,通过引入局部搜索算子在增加粒子多样性的同时有效避免了粒子陷入局部最优陷阱。通过改进的PSO算法优化SVM参数,建立分类识别器模型。将该AEPSO-SVM模型应用到雷达HRRP目标识别中,实验结果表明,该算法对于高分辨雷达目标识别精度高、鲁棒性强。

关键词: 雷达自动目标识别, 高分辨距离像, 粒子群算法, 支持向量机

Abstract:

Radar high resolution range profile (HRRP) identification in radar automatic target recognition based on an adaptive evolutionary particle swarm optimization (AEPSO) optimizing support vector machine (SVM) target classifier method is proposed. To enhance the learning ability of particles in the evolution process, the proposed method, using the staged adjustment acceleration factor, adapts to the nonlinear variation of the particle optimization by an adaptive inertia weight. According to introducing local search operators, the particle diversity, to avoid the local optimal trap, is increased effectively. The SVM parameters are optimized by the improved PSO algorithm to establish a classification recognizer model. The AEPSO-SVM model is applied to radar HRRP target recognition, and the simulation results demonstrate that the proposed method has better accuracy and robustness.

Key words: radar automatic target recognition, high resolution range profile (HRRP), particle swarm optimization (PSO), support vector machine (SVM)