新方法自动识别复杂雷达信号PRI模式

94 浏览量更新于2024-08-26 1 收藏 659KB PDF 举报

本文提出了一种新颖的方法，旨在解决复杂雷达信号中的脉冲重复间隔(PRI)模式自动识别问题。随着现代雷达技术的进步，信号特征变得更加复杂，传统的识别方法可能面临挑战。作者Yuwen Tang、Minghao He、Xiaojie Tang和Jun Han（*表示通讯作者，联系邮箱dspbuilder_t@163.com）来自中国空军预警学院和西安空军工程大学，他们专注于改进PRI特征提取和模式识别的精度与鲁棒性。首先，该研究在时域内对PRI序列进行深入分析，特别关注了三个关键特征：粗糙度、归一化平均跳跃能量和Sp曲线特性。粗糙度反映了序列的不规则变化程度，平均跳跃能量则衡量了相邻PRI之间的差异，而Sp曲线特性则通过绘制PRI值分布的统计图形，揭示了信号的周期性和异常模式。这些特征的选择是基于对雷达信号特性的深入理解，能够有效捕捉不同PRI模式的差异。接下来，决策树算法被用来作为分类器，其结构有助于通过比较和组合特征来区分各种不同的PRI模式。决策树的优势在于其易于理解和解释，且在处理多类问题时具有较高的准确性。通过这种方法，算法能够在处理复杂信号时保持较高的识别率，并表现出良好的抗伪脉冲和脉冲缺失的鲁棒性。模拟实验结果显示，新方法在实际应用中表现出显著的优势，不仅能够准确地识别出不同类型的PRI模式，而且对于雷达发射机的识别也有所帮助，这对于雷达操作模式识别和战术性能分析至关重要。这项研究为复杂雷达信号的PRI模式自动识别提供了一种实用且有效的工具，有望促进雷达信号处理领域的进一步发展。

A NOVEL METHOD FOR AUTOMATICALLY

IDENTIFYING PRI PATTERNS OF COMPLEX

RADAR SIGNALS

Yuwen Tang

, Minghao He

, Xiaojie Tang

, Jun Han

, Xikun Fan

Air Force Early Warning Academy, Wuhan, China

Air Force Engineering University, Xi’an, China

*dspbuilder_t@163.com

Keywords: PRI, FEATURE EXTRACTION, MODULATION RECOGNITION, RADAR EMITTER

RECOGNITION

Abstract

Aimed at the problem of automatically identifying PRI

patterns of complex radar signals, a novel method is proposed

in this paper. The roughness, normalized mean jump energy,

and

-curve characteristic are extracted as feature set based

on the analysis of PRI sequences in time domain. Finally, the

decision tree is employed to recognize different PRI patterns.

The simulation results show that the algorithm has high

recognition accuracy and good robustness to spurious pulses

and pulse missing.

1 Introduction

The radar pulse repetition interval (PRI) is a time series

parameter of the radar pulse train. It is a key parameter for

analysing the radar's main tactical performance and identifying

the radar operation mode. However, with the development of

modern radar technologies, the PRI modulations are becoming

more and more complex, which puts forward higher

requirements for radar emitter signal recognition technologies.

Therefore, it is very necessary to find a new reliable algorithm

for PRI pattern recognition.

The algorithm studied in this paper is performed on the rough-

sorted radar signal PRI sequences, so that, pulse deinterleaving

processing is not considered. Existing PRI pattern recognition

methods can be roughly divided into three major categories:

one is the methods based on statistical histograms or plane

transformations. Typical examples include cumulative

difference histogram (CDIF)[1], sequential sequence

difference histogram (SDIF)[2], and PRI transform[3,4], etc.

However, the histogram-based methods have a good

performance only when deinterleaving the Constant and

Stagger PRI, and are ineffective for other patterns[5].

Moreover, these methods require manual analysis of a large

amount of pulse data, and it is difficult to identify small data

samples effectively. On the other hand, although the PRI

transform can handle small data samples, it needs constantly

adjust the width of the plane to obtain a clear image and it is

difficult to achieve automatic processing[6]. Another kind of

methods are based on the shape features of autocorrelation

function[5,7,8]. However, the autocorrelation function figure

of PRI sequences will change with the modulation parameters,

and it is sensitive to the pulse missing. The third kind of

methods are based on sequence differential characteristics. For

example, the [9] proposed a method based on the first-order

difference of PRI sequences and uses multi-layer perceptron to

identify. But its feature dimension is up to 64, and the

calculation cost is too high. The [10] obtained a good result by

combining the histogram-based methods and the sequence

difference methods, but the recognition accuracy decreases

significantly when both pulse missing and spurious pulses are

present[11].

The above literatures provide a good idea for the PRI pattern

recognition. This paper starts with the analysis of the time-

domain features of the PRI patterns and extracts three features

of the PRI sequences, namely, roughness, normalized mean

jump energy, and

-curve characteristic. Finally, the

decision tree is employed to complete the recognition.

Simulation results show that the proposed method has a good

ability to resist spurious pulses and pulse missing.

1.1 Complex PRI patterns

PRI is the first-order difference of the arrival time of pulse

front. The PRI modulation function can be expressed as

( ) , 1,2

n n n

F n TOA TOA x n N   

(1)

Where

TOA

represents the arrival time of the

-th pulse

front. In order to realize its rich operation modes, modern

multi-function radars have developed many complex PRI

patterns, and each pattern for a different function. The

common PRI patterns are mainly divided into six types:

Constant, Stagger, Jitter, Dwell and Switch(D&S), Sliding,

and Periodic. Since the histogram-based methods have been

able to recognize the Constant and Stagger PRI patterns, this

paper mainly focuses on the other four.

1.1.1 Jitter: Jitter is mainly used for anti-reconnaissance and

anti-jamming. It is the pattern that the PRI jittered around a

constant value randomly within a certain range.

( ) ( ), 1,2F n PRI n n N



  

(2)

下载后可阅读完整内容，剩余4页未读，立即下载

weixin_38690149

粉丝: 7
资源: 909

新方法自动识别复杂雷达信号PRI模式

论文研究-基于改进的Kalman滤波的雷达信号PRI估计 .pdf

一种基于全卷积神经网络的雷达PRI调制样式识别方法.pdf

研究论文-复杂雷达信号模拟源的设计

雷达信号分选的经典matlab代码 包含PRI变换、CDIF、SDIF三种方式

Matlab 信号分选 PRI变换 序列差直方图算法(SDIF) 累积差直方图算法(CDIF) 固定重频信号 抖动重频信号

信噪比不稳定条件下辐射信号的改进识别方法

基于多级箱与深度森林的雷达信号分选算法.docx

基于FPGA的雷达脉冲信号分选预处理研究.pdf

使用频率和形状特征的雷达信号PRJ识别方法

电子情报分析：雷达信号的参差周期与识别

最新资源

雷达信号分选的经典matlab代码包含PRI变换、CDIF、SDIF三种方式

Matlab 信号分选 PRI变换序列差直方图算法(SDIF) 累积差直方图算法(CDIF) 固定重频信号抖动重频信号