University of Nebraska - Lincoln University of Nebraska - Lincoln
DigitalCommons@University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln
Faculty Publications from the Department of
Electrical and Computer Engineering
Electrical & Computer Engineering, Department
of
2000
Doppler Estimation Using a Coherent Ultrawide-Band Random Doppler Estimation Using a Coherent Ultrawide-Band Random
Noise Radar Noise Radar
Ram M. Narayanan
University of Nebraska-Lincoln
Muhammad Dawood
University of Nebraska-Lincoln
Follow this and additional works at: https://digitalcommons.unl.edu/electricalengineeringfacpub
Part of the Electrical and Computer Engineering Commons
Narayanan, Ram M. and Dawood, Muhammad, "Doppler Estimation Using a Coherent Ultrawide-Band
Random Noise Radar" (2000).
Faculty Publications from the Department of Electrical and Computer
Engineering
. 147.
https://digitalcommons.unl.edu/electricalengineeringfacpub/147
This Article is brought to you for free and open access by the Electrical & Computer Engineering, Department of at
DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Faculty Publications from
the Department of Electrical and Computer Engineering by an authorized administrator of
DigitalCommons@University of Nebraska - Lincoln.
868 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 48, NO. 6, JUNE 2000
Doppler Estimation Using a Coherent Ultrawide-Band
Random Noise Radar
Ram M. Narayanan, Senior Member, IEEE, and Muhammad Dawood, Student Member, IEEE
Abstract—The University of Nebraska has developed an ultra-
wide-band (UWB) coherentrandom noise radar operating overthe
1–2 GHz frequency range. The system achieves phase coherenceby
using heterodyne correlation of the received signal with a time-de-
layed frequency-shifted replica of the transmit waveform. Knowl-
edge of the phase of the received signal and its time dependence
due to target motion permits the extraction of the mean Doppler
frequencyfrom which the targetspeed can be inferred.Theoretical
analysis, simulation studies, and laboratory measurements using a
microwave delay line showed that it was possible to estimate the
Doppler frequency from targets with linear as well as rotational
motion. Field measurements using a photonic delay line demon-
strated the success of this technique at a range of about 200 m
at target speeds of up to 9 m/s. Analysis shows that the accuracy
with which the Doppler frequency can be estimated depends not
only on the phase performance of various components within the
system, but also upon the random nature and bandwidth (BW) of
the transmit waveform, and the characteristics of unsteady target
motion.
Index Terms—Doppler estimation, random noise radar, ultra-
wide-band radar.
I. INTRODUCTION
D
OPPLER radars estimate target velocity by measuring
the frequency shift between the transmit and receive fre-
quencies. Thus, these systems can be used to identify moving
targets and separate these from stationary targets and slowly
varying clutter. These systems maintain phase coherence by
using the same stable master oscillator (STAMO) for mixing
and frequency conversion in the transmitter and the receiver.
The University of Nebraska, Lincoln, has developed a
technique that succeeds in injecting coherence in a radar
system that transmits wide-band random noise. Phase coher-
ence is obtained using heterodyne correlation of the received
signal with a time-delayed frequency-translated replica of the
transmit waveform. This ensures that the reflected signal, when
mixed with the time-delayed transmit signal, yields the same
intermediate frequency, thereby preserving the phase contained
within the reflected signal. This system operates over 1–2
GHz frequency band, thereby achieving 1-GHz instantaneous
bandwidth yielding a down-range resolution of 15 cm. The
phase coherence in the system has been used to configure the
radar as a Doppler radar for measurement of target velocity, the
results of which are described in this paper.
Manuscript receivedJune 24, 1999;revised February 25, 2000. This workwas
supported by the Office of Naval Research under Contract N00014-1-97-0200.
The authors are wih the Department of Electrical Engineering, Center for
Electro-Optics, University of Nebraska, Lincoln, NE 68588-0511 USA.
Publisher Item Identifier S 0018-926X(00)05803-8.
Fig. 1. Block diagram of ultrawide-band random noise radar system.
Section II provides a description of the University of Ne-
braska’s 1–2 GHz coherent random noise radar system. In Sec-
tion III, we developthe basic theory of Doppler estimation using
the system. Results of computer simulations are shown in Sec-
tion IV, which support the theoretical analysis. Experimental re-
sults for the short-range laboratory measurments and the long-
range field measurements are shown in Sections V and VI, re-
spectively. An analysis of error sources and their effects on
Doppler performance is provided in Section VII. Section VIII
provides a discussion of the results and presents conclusions.
II. D
ESCRIPTION OF COHERENT RANDOM NOISE RADAR
SYSTEM
A block diagram of the polarimetric random noise radar
system is shown in Fig. 1 [1]. This system was originally
designed to detect and identify shallow buried objects, such
as landmines. The noise signal is generated by a noise source
OSC1, which provides a wide-band noise signal with a
Gaussian amplitude distribution and a constant power spectral
density in the 1–2-GHz frequency range, with a power output
of 0 dBm. This output is split into two in-phase components
in power divider PD1. One of these outputs is amplified in
a 34-dB gain power amplifier AMP1, which has a 1-dB gain
compression point greater than
40 dBm. Thus, the average
power output of AMP1 is
30 dBm (1 W), but the amplifier
0018–926X/00$10.00 © 2000 IEEE
NARAYANAN AND DAWOOD: DOPPLER ESTIMATION USING COHERENT RANDOM NOISE RADAR 869
(a) (b)
(c) (d)
Fig. 2. Simulated Doppler spectra of linear motion at velocities of (a) 1.1 m/s; (b) 1.8 m/s; (c) 2.3 m/s; and (d) 2.3
6
1 m/s.
is capable of faithfully amplifying noise spikes that can be as
high as 10 dB above the mean noise power. The output of the
amplifier is connected to a broad-band (1–2 GHz) transmit
horn antenna ANT1. The E/H plane beamwidths and gain of
ANT1 at the center frequency of 1.5 GHz are 23
,34 , and 17
dB, respectively.
The other output of the power divider PD1 is fed to a com-
bination of a fixed and variable delay lines: DL1 and DL2,
respectively. These delay lines are used to provide the neces-
sary transmit delay so that it can be correlated with the re-
ceived signal scattered from objects at appropriate distance cor-
responding to the delay. The variable delay line is a seven-bit
programmable stepped delay line that can be varied from 0 to
19.812 ns in 0.156 ns steps. The fixed delay line is physically
realized by a low-loss linear phase shifter in the 1–2 GHz fre-
quency range.
In order to perform coherent processing of the noise signals,
the delayed replica of the transmit waveform is mixed in MXR1
with an IF signal produced by a 160-MHz phase-locked os-
cillator OSC2, which is phase-locked to an internal 1–5 MHz
crystal. MXR1 is a lower side-band upconverter that yields an
output within the 0.84–1.84-GHz frequency range, while the
upper side-band output at 1.16–2.16 GHz is internally termi-
nated. This coherent noise signal is split by power divider PD3
into two identical channels, which can also be configured as the
copolarized and the cross-polarized channels.
We will now discuss the signal processing of one of the chan-
nels since other channel operation is essentially identical. One
of the outputs of PD3 is amplified in AMP4, a 19-dB gain ampli-
fier. Since this signal is noiselike, amplifier AMP4 is chosen so
as to provide a linear output of
10 dBm minimum. This signal
is used as the local oscillator (LO) input to a biasable mixer,
MXR2, whose RF input is obtained from receive antenna ANT2
and a 20-dB gain low-noise amplifier AMP2. The receive an-
tenna is a dual-polarizedlog-periodic antenna of constant7.5 dB
gain over the 1- to 2-GHz frequency range. Amplifier AMP2 is
used to improve the receiver noise figure. Mixer MXR2 is dc-bi-
ased in the square-law region, which ensures that the mixing
process is efficient for low LO drive levels. In general, the RF
input signal to mixer MXR2 consists of transmitted noise at
1–2 GHz scattered and reflected from various targets. However,
since the LO signal has a unique delay associated with it, only
870 IEEE TRANSACTIONS ON ANTENNAS AND PROPAGATION, VOL. 48, NO. 6, JUNE 2000
(a) (b)
Fig. 3. Simulated Doppler spectra of rotating target at (a) 40 rpm and (b) 75 rpm.
the signal scattered from the appropriate range bin will mix with
the LO to yield an IF signal at 160 MHz. Signals scattered or
reflected from other range bins, will not be correlated with the
delayed replica. The output of the mixer MXR2 is connected
to a narrow-band bandpass filter FL1 of center frequency 160
MHz and bandwidth 5 MHz, ensuring that only 160-MHz sig-
nals get through. Theoutput of filterFL1 at160 MHz is split into
two equal outputs by power divider PD5. One of these outputs
is amplified and detected in a 70-dB dynamic range 160-MHz
logarithmic amplifier AMP6 of 20-MHz bandwidth. The other
output of power divider PD5 is connected to one of the inputs
of I/Q detector IQD1, whose reference frequency input is one
of the outputs from PD4. Both of the signals are centered at
160 MHz; thus, the
detector provides the in-phase and
quadrature
components of the two signals. Since frequency
translation preserves phase differences, the
and outputs can
be used to extract the Doppler shift produced due to the motion
of the target.
If the radar system is configured in both copolarized and
cross-polarized modes, it will produce the following outputs at
various ranges as set by the delay lines: 1) copolarized ampli-
tude; 2) copolarized phase angle; 3) cross-polarized amplitude;
and 4) cross-polarized phase angle. The system outputs can,
therefore, be related to the polarimetric scattering characteris-
tics of the target besides Doppler estimation.
III. T
HEORY OF DOPPLER ESTIMATION USING COHERENT
RANDOM NOISE RADAR
Since the transmitted amplitude has a Gaussian amplitude
distribution and uniform power spectral density, it can be mod-
eled as
(1)
where
represents the Gaussian amplitude distribution,
(
) is the center frequency at 1.5 GHz, ( ) is uni-
formly distributed over the
0.5-GHz frequency range, and
is the arbitrary transmitter phase.
The time-delayed version of the transmitted signal
is
mixed in MXR1 with the reference frequency (
) at 160
MHz to produce the lower side-band output
given by
(2)
where
is some constant and is the delay.
The echo from the target is expressed as
(3)
where
is the velocity of light, and are the amplitude and
phase of the target reflectivity, and the term
represents the
time taken by the transmitted wave to return to the receiver from
the target at range
.
The instantaneous phase of the echo voltage can be defined
as
(4)
where
(5)
is the instantaneous wavelength.
If the target is in motion,
will change with time and (3)
can be written as
(6)
where
is the target velocity given by , and is a con-
stant.
This received echo is mixed with the output of MXR1 at a
delay time set equal to
, yielding
(7)
where
is some constant.
NARAYANAN AND DAWOOD: DOPPLER ESTIMATION USING COHERENT RANDOM NOISE RADAR 871
(a) (b)
(c) (d)
(e) (f)
Fig. 4. Measured Doppler spectra of linear motion at short range at 1-GHz fixed frequency for velocities of (a) 1.1 m/s; (b) 1.8 m/s; and (c) 2.3 m/s; also, at 1–2
GHz random frequency for velocities of (d) 1.1 m/s; (e) 1.8 m/s; and (f) 2.3 m/s.
The output of MXR2 and are fed to the detector
producing inphase and quadrature components that are propor-
tional to the cosine and sine of the phase difference, respectively
(8)
(9)
where
and represent the amplitudes of the and com-
ponents, respectively. We note that the
and outputs are
time-varying functions depending upon the target velocity
.
The Doppler frequency
is given by ) times the total
phase and can be shown to be equal to
(10)
