没有合适的资源?快使用搜索试试~ 我知道了~
温馨提示
Passive bistatic coherent radar system operates with distinct non-cooperative transmitter and receiver located at different sites. In such spatial separation, the receiving system needs to independently solve the problem of time and phase synchronization via direct-path signal. The impact of time synchronization error was explored and solved in passive bistatic coherent radar. The common model of digital sampling process is introduced to illustrate the effect of time synchronization error (TSE).
资源推荐
资源详情
资源评论
.
RESEARCH PAPERS
.
SCIENCE CHINA
Information Sciences
December 2010 Vol. 53 No. 12: 2664–2674
doi: 10.1007/s11432-010-4110-x
c
Science China Press and Springer-Verlag Berlin Heidelberg 2010 info.scichina.com www.springerlink.com
The impact of time synchronization error on passive
coherent pulsed radar system
HE You
1
, ZHANG CaiSheng
1,2 ∗
,DINGJiaHui
2
& TANG XiaoMing
1
1
Research Institute of Information Fusion, Naval A eronautical and Astronautical University,
Yantai 264001,China;
2
Nanjing Research Institute of Electronics Technology, Nanjing 210039,China
Received May 18, 2010; accepted August 9, 2010
Abstract Passive bistatic coherent radar system operates with distinct non-cooperative transmitter and re-
ceiver located at different sites. In such spatial separation, the receiving system needs to independently solve
the problem of time and phase synchronization via direct-path signal. The impact of time synchronization error
was explored and solved in passive bistatic coherent radar. The common model of digital sampling process is
introduced to illustrate the effect of time synchronization error (TSE). The periodicity of relative sample instant
variation of each pulse is deduced. Then, normalized interference power (NIP) is defined to evaluate the poten-
tial impact of TSE. The NIP expressions are derived when the difference between the adjacent relative sample
instant is uniform distribution or Gaussian distribution. It is found that there are new spurious frequency com-
ponents and the detection performance is decreased in the case of asynchronous sampling. At the same time,
the analytical Doppler frequency bias expression caused by TSE is presented. Theoretical results are confirmed
by simulation in the cases of different periods of relative sampling instant.
Keywords passive coherent radar (PCR), time synchronization error (TSE), normalized interference power
(NIP)
Citation He Y, Zhang C S, Ding J H, et al. The impact of time synchronization error on passive coherent pulsed
radar system. Sci China Inf Sci, 2010, 53: 2664–2674, doi: 10.1007/s11432-010-4110-x
1 Introduct ion
One premise of digital coherent signal processing for Doppler radar is keeping the coherence between
pulses after data sampling. Several restrictions on digital sampling parameters are proposed in [1, 2]. In
order to guarantee the coherence between pulses, the sampling clock and trigger signal must synchronize
to the pulse repetition frequency (PRF) of non-cooperative transmitting signal. The performance of time
synchronization will directly affect the coherent signal processing gain [3].
One of the most critical issues for coherent detection in passive coherent pulsed radar is time synchro-
nization. The only technique of time synchronization for passive coherent pulsed radar, however, is to
synchronize the sampling clock by direct-path signal in the passive receiving system [4–7]. In practice,
even if the direct-path signal can be received successively during the scanning of the transmitter’s an-
tenna, the signal-to-noise ratio of direct-path signal will affect the accuracy of PRF estimation. As a
∗
Corresponding author (email: caifbi2008@yahoo.com.cn)
HE You, et al. Sci China Inf Sci Decemb er 2010 Vol. 53 No. 12 2665
result, there is TSE between the time-base of sampling clock and target echo. The number of samples
per PRF may be not an integer, and the relative sample instant will drift, too. In the process of digital
signal processing, however, the independent variable in the sampling sequence has no information about
the jitter of timing clock.
Jenq [8, 9] first described and analyzed the spectrum of ideal sine wave in the case of non-uniformly
sampling. Its original idea is decomposing the non-uniformly sampled sequence into M subsequences,
which are subsequently treated as uniformly sampled data. The digital spectral representation of a non-
uniformly sampled sequence data is proposed in terms of its analog spectrum. Based on this research,
ref. [10] explored a more generalized digital spectral expression for non-uniformly sampled signal. Perfect
reconstruction algorithm of digital spectrum from non-uniformly sampled signals is proposed if timing
error in the receiving system is known [11]. Some special cases on the spectrum reconstruction algorithm
are presented in [3, 12].
In this paper, we concentrate on the impact of TSE on coherent detection and Doppler shift estimation
in passive coherent pulsed radar system. The common model of digital sampling process is described in
section 2. The period of relative sample instant variation of each pulse is explored via modeling in
section 3. In section 4, NIP is defined to assess the potential impact of the TSE on coherent detection
and Doppler shift estimation. The NIP expressions are derived when the difference between the adjacent
relative sample instant is uniform distribution or Gaussian distribution. Finally, conclusions are presented
in section 5.
2 Problem description
A scenario with non-cooperative radar illuminator considered in this paper is illustrated in Figure 1.
Assume that the transmitter is mechanically scanned in azimuth, and the receiver is stationary while the
target is moving. Both the transmitter and receiver are focusing on the moving target. The receiving
system intercepts the direct-path waveform transmission through a reference antenna when it tunes to
the transmitting frequency, and target reflection echoes are intercepted by the target antenna. Time
and phase synchronization need to be completed via direct-path signal. Then, optimum detection and
parameters estimation are commonly involved in passive coherent processing. Surveillance and early-
warning may be achieved in the area of interest [13, 14].
According to the Nyquist sampling theorem, the sampling frequency F
s
must not be less than 2F
c
,
where F
c
is the carrier frequency of transmitting signal. In practice, however, in coherent pulse radar,
the coherence is required among pulses after data sampling process. Generally speaking, the common
sampling process can be described as Figure 2. For convenience, the rising edge of each pulse is defined
as the starting instant of the corresponding PRF. Here, δt
i
denotes the interval between the first sample
instant of the ith pulse, and 0 δt
i
T
s
,whereT
s
is sampling interval. Then, δt
1
is fixed time delay
between the sampling clock and the rising edge of the first pulse caused by signal transmission. For the
purpose of coherent processing, it is desired that δt
1
= δt
2
= ···= δt
i
. In this way, we should guarantee
that F
s
= pf
r
, p =1, 2, 3,...,wheref
r
is the value of PRF.
In the passive coherent radar system, however, we need to estimate the PRF of the transmitting
signal via direct-path pulses. In theory, the direct-path signal is modulated by the rotating transmitter’s
antenna pattern, which will reduce the estimation accuracy of PRF, and hence TSE between the sample
frequency and PRF is introduced unconsciously.
Assume that the range sweep is initiated by a threshold detection of the direct-path pulse. In a passive
bistatic receiving system, we designate the first sample of each direct-path pulse crossed the threshold as
the trigger of sampling clock. There is a series of data samples in every pulse repetition interval (PRI).
With no loss of generality, the sampling process of target reflection echoes in the target dwell time can
be illustrated as Figure 3. It can be found that the first sample instant of each pulse is drifting relative
to the start of the direct pulse. Note that sample 1 is the first sample of direct-path pulse from PRI 1
to PRI 4. However, sample 4 will be the first sample of direct-path pulse from PRI 5 to PRI 8 and the
process will be repeated. For the case of a general direct path pulse shape, the net result of this asynch-
剩余10页未读,继续阅读
资源评论
weixin_38616120
- 粉丝: 7
- 资源: 944
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
最新资源
- 2025年 UiPath AI和自动化趋势:代理型AI的崛起及企业影响
- 基于Java的环境保护与宣传网站的设计与实现毕业论文.doc
- 2025年人形机器人产业发展蓝皮书-量产及商业化关键挑战
- 互联网金融发展指数 (第二期,2014年1月-2015年12月).zip
- 百度智能云千帆大模型平台推进企业多模态生成式AI应用
- 形状检测32-YOLO(v5至v9)、COCO、CreateML、Darknet、Paligemma数据集合集.rar
- qwewq23132131231
- 2024年智算云市场发展与生态分析报告
- 冒泡排序算法解析及优化.md
- MySQL中的数据库管理语句-ALTER USER.pdf
资源上传下载、课程学习等过程中有任何疑问或建议,欢迎提出宝贵意见哦~我们会及时处理!
点击此处反馈
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功