电子学报 ›› 2022, Vol. 50 ›› Issue (6): 1351-1358.DOI: 10.12263/DZXB.20210934

所属专题: 电磁频谱智能+

• 电磁频谱智能+ • 上一篇    下一篇

基于多级箱与深度森林的雷达信号分选算法

张春杰1,2, 刘俞辰1,2, 司伟建1,2   

  1. 1.哈尔滨工程大学信息与通信工程学院,黑龙江 哈尔滨 150001
    2.哈尔滨工程大学先进船舶通信与信息技术工业和;信息化部重点实验室,黑龙江 哈尔滨 150001
  • 收稿日期:2021-07-16 修回日期:2022-01-07 出版日期:2022-06-25
    • 作者简介:
    • 张春杰 女,1975年出生,黑龙江哈尔滨人.现为哈尔滨工程大学信息与通信工程学院副教授.主要研究方向为宽带信号检测、处理及识别.
      刘俞辰(通讯作者) 男,1997年出生,黑龙江哈尔滨人.现为哈尔滨工程大学硕士研究生.主要研究方向为电子侦察、雷达信号分选等.E-mail: liuyuchen11@hrbeu.edu.cn
      司伟建 男,1971年出生,黑龙江哈尔滨人.现为哈尔滨工程大学研究员.主要研究方向为阵列信号处理、宽带信号检测、信号分选、数字信道化等.
    • 基金资助:
    • 国家自然科学基金 (61801143); 黑龙江省自然科学基金 (LH2020F019)

The Radar Signal Deinterleaving Algorithm Based on Multi-Level Bin and Deep Forest

ZHANG Chun-jie1,2, LIU Yu-chen1,2, SI Wei-jian1,2   

  1. 1.College of Information and Communication, Harbin Engineering University, Harbin, Heilongjiang 150001, China
    2.Key Laboratory of Advanced Marine Communication and Information Technology, Ministry of Industry and Information Technology, Harbin Engineering University, Harbin, Heilongjiang 150001, China
  • Received:2021-07-16 Revised:2022-01-07 Online:2022-06-25 Published:2022-06-25
    • Supported by:
    • National Natural Science Foundation of China (61801143); Natural Science Foundation of Heilongjiang Province, China (LH2020F019)

摘要:

针对复杂电磁环境下大范围抖动、滑变等特殊类型雷达信号难以分选,脉冲重复周期(Pulse Repetition Interval,PRI)估计精度低,脉冲序列搜索效果不佳等问题,提出一种基于PRI多级箱与深度森林的雷达信号单参数分选算法.该算法利用PRI多级箱结构进行PRI变换,提升针对特殊类型雷达的分选正确率.在多级箱中的脉冲对个数与PRI变换结果中,提取雷达信号PRI边界特征,将不同电磁环境下特殊类型雷达信号PRI边界特征混合,通过平滑滤波器增强特征,训练深度森林预测完整雷达信号PRI范围,从而校正中心PRI估计值.最后依据PRI中心值与变化范围,搜索提取脉冲序列,完成分选.仿真实验表明,所提算法可在复杂电磁环境下,对大范围抖动、单线性滑变、双线性滑变、锯齿波滑变以及正弦滑变等特殊类型雷达信号进行有效分选,PRI范围预测效果在原始深度森林基础上提升14%,PRI估计误差降低75%.

关键词: 电子侦察, 特殊类型雷达信号, 多级箱, 深度森林, PRI估计, PRI范围

Abstract:

To address the difficulty of deinterleaving special radars with large pulse repetition interval(PRI) range and the low accuracy of PRI estimation in complex electromagnetic environment, a radar signal deinterleaving algorithm based on PRI multi-level bin and deep forest is proposed. The algorithm utilizes the PRI multi-level bin structure for PRI transformation to improve the detection rate for special radars. The PRI boundary features of radar signals are derived from the number of pulse pairs in the multi-level bin and the PRI transform results. The PRI boundary features of special radars in different environments is mixed and the features by smoothing filters are enhanced. The deep forest is trained to predict the complete PRI range and thus to correct the central PRI estimate. Finally, based on the central value of PRI and the PRI range, the pulses are searched and extracted. Simulation experiments show that the proposed algorithm can effectively deinterleaving jittered, unilinear, bilinear, sawtooth and sinusoidal radars with large PRI range. The PRI range prediction performance is improved by 14% and the PRI estimation error is reduced by 75%.

Key words: electronic reconnaissance, special radar, multi-level bin, deep forest, pulse repetition interval estimation, pulse repetition interval range

中图分类号: