Nothing Special   »   [go: up one dir, main page]

Low-Cost and High Resolution X-Band Laboratory Radar System For Synthetic Aperture Radar Applications

Download as pdf or txt
Download as pdf or txt
You are on page 1of 6
At a glance
Powered by AI
A low-cost high resolution X-band radar system for SAR applications was developed using inexpensive consumer electronics for around $240. It is capable of range resolution of 1.4 inches and dynamic range of 60 dB.

The system was developed to provide an affordable entry point for universities and small businesses to get involved with radar cross section measurements and SAR algorithm development.

It works by frequency modulating a signal and measuring the beat frequency between the transmitted and received signals, where the beat frequency corresponds to the range of targets. A linear ramp modulation is used to achieve this.

LOW-COST, HIGH RESOLUTION X-BAND LABORATORY RADAR SYSTEM FOR

SYNTHETIC APERTURE RADAR APPLICATIONS

Gregory L. Charvat, MSEE


Department of Electrical and Computer Engineering
Michigan State University
2120 Engineering Building
East Lansing, MI 48825

ABSTRACT An explanation of the radar system design is presented in


Using a discarded garage door opener, an old cordless section 3. Section 4 presents range profile measurement
drill, and a collection of surplus microwave parts, a high results. High resolution SAR imaging results are
resolution X-band linear rail synthetic aperture radar presented in section 5. Section 6 will discuss conclusions
(SAR) imaging system was developed for approximately and future work. References are provided in section 7.
$240 material cost. Entry into the field of radar cross
section measurements or SAR algorithm development is 2. Background
often difficult due to the cost of high-end precision pulsed FMCW radar was first widely used in radio altimeters,
IF or other precision radar test instruments. The low cost starting in the mid 1930’s [1]. FMCW has a number of
system presented in this paper is a frequency modulated design advantages, including a high average power and
continuous wave radar utilizing a homodyne radar short range capabilities. FMCW is unique in its ability to
architecture. Transmit chirp covers 8 GHz to 12.4 GHz range targets extremely close to the radar transmit and
with 15 dBm of transmit power. Due to the fairly wide receive antennas. The major disadvantage of FMCW
transmit bandwidth of 4.4 GHz, this radar is capable of radar (or any CW radar system) is antenna coupling. The
approximately 1.4 inches of range resolution. The transmit to receive antenna coupling limits dynamic range
dynamic range of this system was measured to be 60 dB in a FMCW radar system.
thus providing high sensitivity. The radar system
When a CW radar system is FM modulated, the range to
traverses a 96 inch automated linear rail, acquiring range
target information provided is in the form of beat
profiles at any user defined spacing. SAR imaging results
frequencies. This is known as FMCW radar. The beat
prove that this system could easily image objects as small
frequencies on the video output of an FMCW radar
as pushpins and 4.37 mm diameter steel spheres.
system correspond to multiple targets and their
corresponding ranges. FMCW radar systems are also
Keywords: Radar Imaging, Synthetic Aperture Radar,
capable of measuring the Doppler shift of a moving target.
Measurement Systems, Low Cost Radar, Low Cost Rail
The block diagram of a basic FMCW radar system is
SAR.
shown in figure 1.
1. Introduction
A low cost, high resolution, X-band synthetic aperture
radar (SAR) imaging system was developed by the
Michigan State University Electromagnetics Research
Group. The purpose behind this research effort was to
develop a low cost entry level system for use by
universities or small businesses looking to enter the field
of radar cross section (RCS) measurements or SAR Figure 1: Block diagram of a basic FMCW radar
algorithm development. A high performance but low cost system.
rail SAR was developed using a discarded garage door
opener, an old cordless drill, and some surplus microwave Looking at figure 1, OSC1 is FM modulated with a
parts for a very low total material cost. In this paper the triangular ramp input with a period of 2 ms. The
implementation, range profile data results, and radar triangular ramp is an alternating linear ramp with both
imagery from this system will be presented. positive and negative slopes as shown in figure 2.

A background discussion on frequency modulated


continuous wave (FMCW) radar is presented in section 2.
When a negatively sloped linear ramp FM modulates
OSC1, the beat frequency at the IF port of MXR1 is
represented by:
8∆ ff m R
f b = f b− = + fd (2)
c

Where: f b = the beat frequency at the IF port of
MXR1 when OSC1 is modulated with a
negative linear ramp

From the equations above, the range to target is found


using:
c
R= fb (3)
8∆ ff m
Figure 2: Linear ramp used to FM modulate the Where: f b = the average frequency difference
FMCW radar system.
If the target is moving, the velocity of the target can be
The FM output of OSC1 is fed into PA1. PA1 amplifies
found using:
OSC1 to an appropriate transmit level. The output of
λ
OSC1 is radiated out of the transmit antenna. The FM
modulated carrier is reflected off of the target at a range
v=
4
( f b− − f b+ ) (4)

of R meters. The reflected signal is delayed in time on its


way to and from the target. The reflected signal is The amplitude of the return signals can be approximated
received by the receive antenna, and amplified by LNA1. using the radar range equation [2].
The output of LNA1 is fed into the RF port of MXR1.
Some power from PA1 is coupled into the LO port of The most important concept explained here is that a shift
MXR1. When the LO and the RF are multiplied together in time corresponds to a shift in frequency. This is
in MXR1, the IF output of MXR1 is the range to target in because the radar is frequency modulated in time. The
frequency. This range to target in terms of frequency is current value of the transmitted frequency is different than
known as the beat frequency. The greater the beat what was transmitted 2 ns ago. These small and subtle
frequency on the IF output port of MXR1, the greater the frequency differences make up the beat frequencies on the
range to target. If the target is moving, then the Doppler IF output of MXR1, and hence the range to target
shift of the moving target is added onto the beat frequency information in the form of low frequency beats.
present on the IF port of MXR1. The relationship
between frequency, FM chirp bandwidth, range to target, The Electromagnetics Group at Michigan State University
and Doppler frequency shift can be found using the has previously presented a low cost FMCW rail SAR
equations for both a positive and a negative linear ramp imaging system, known as the unique approach to FMCW
modulation waveform [2]. When a positively sloped [3, 4]. This system was capable of SAR imaging,
linear ramp FM modulates OSC1, the beat frequency at however it lacked the sensitivity and range resolution
the IF port of MXR1 is represented by: required to research highly advanced SAR imaging
8∆ ff m R algorithms, and for that reason it was decided to develop a
f b = f b+ = − fd (1) more sophisticated high resolution FMCW radar system
c [5]. Based on promising results from [5] it was then
Where: f b = the beat frequency at the IF port of MXR1 decided to utilize this system as a high resolution linear
rail SAR imaging system.
f b+ = the beat frequency at the IF port of
MXR1 when OSC1 is modulated with a 3. System Implementation
positive linear ramp. The low cost high resolution X-band laboratory radar
∆ f = chirp frequency deviation system discussed in this paper is a homodyne FMCW
f m = FM modulation rate system like that shown in section 2. The radar system
front end is shown in Figure 3. Figure 4 shows the data
R = range to target acquisition, power supply, and motion control chassis.
Figure 5 shows the complete system in operation.
generated by DAC1 and OP3, thus producing the 8 GHz
to 12.4 GHz transmit chirp. The output of OSC1 feeds
into the directional coupler CLPR1. The coupled output
of CLPR1 is fed through circulator Circ2 and feeds the
LO port of the double balanced mixer MXR1. The
through port of CLPR1 is fed through the circulator Circ1
to the transmit horn antenna Ant1. Ant1 is a standard gain
X-band horn that is fed by a WR90 waveguide transition.
The transmit power of the chirp signal is 15 dBm. The
chirp signal from Ant1 is then radiated out toward the
target scene.
Figure 3: The low cost high resolution X-band
laboratory radar system front end.

Figure 4: Data acquisition, power supply, and motion


control chassis.

Figure 6: Block diagram of radar system.

Reflected chirp signals from the target scene are then


received by the receiver antenna Ant2. Ant2 is a standard
gain X-band horn that is fed by a WR90 waveguide
transition. The output of Ant2 is fed into LNA1. LNA1
is a 25 dB gain, 8 GHz to 12.4 GHz LNA. The output of
LNA1 feeds the RF port of MXR1. The IF output of
MXR1 feeds into the video amplifier OP1. The output of
OP1 is fed through a 60 KHz active low pass filter OP2.
The output of OP2 is the video output of the radar system.
Figure 5: Compete system in operation. This video output contains the beat frequencies which
provide range to target information.
This particular radar system chirps linearly from 8 GHz to
12.4 GHz with a chirp rate of 440 GHz/sec. The The video output of OP2 feeds into the analog to digital
sensitivity of this system is 25.1 μV, and the dynamic converter ADC1. Data acquisition and ramp modulation
range is 60 dB. A block diagram of the system is shown are performed coherently and synchronized by clock
in Figure 6. generator OSC2. ADC1 is a 16 bit ADC sampling at 200
KSPS. ADC1 samples the video output of OP2
OSC1 is a voltage tuned YIG oscillator that tunes from 8 coherently with the digital to analog converter DAC1.
GHz to 12.4 GHz. OSC1 is modulated by a linear ramp The output of ADC1 is fed into a first in, first out (fifo)
register denoted as FIFO1. The data output of FIFO1 is
then transferred to the data acquisition and analysis
computer. The data acquisition and analysis computer
controls the entire system and its parameters. This
computer also processes the range profile data.

The data acquisition and analysis computer fills the


second fifo, FIFO2, with values for linear ramp
modulation of OSC1. The data from FIFO2 is sampled
into DAC1. DAC1 outputs samples coherently with
ADC1. The ramp output of DAC1 is a stair cased digital
version of a pure ramp. OP3 is a 5 KHz active low pass
filter that filters the stair case effect thus smoothing the
linear ramp waveform which is modulating OSC1.

The data acquisition and analysis computer also controls


the stepper motor controller. The stepper motor controller
is connected to an inexpensive biphasic stepper motor.
The output shaft of this stepper motor is fed into an old
cordless drill planetary gear set transmission in order to
multiply the torque up to that required to move the garage
door opener rail, see figure 7. The output of the
transmission is coupled to the discarded Genie screw type Figure 8: A discarded Genie screw type garage door
opener rail was utilized to move the SAR front end.
[7] garage door opener rail, see figure 8. A custom
machined carrier was made to mount the radar front end
Utilizing what limited resources were available we were
onto the garage door opener rail.
able to successfully implement an FMCW linear rail SAR
with large transmit bandwidth and precision positioning
capabilities on a budget of only $240 total material cost.

4. Range Profile Data


A number of range profiles were acquired using the low
cost high resolution X-band laboratory radar system.
These range profiles were acquired to get a rough idea as
to the SAR imaging possibilities of this system. Due to
precision stepper motor drive assembly the lack of a readily available X-band LNA at the time,
these tests were conducted using a chirp bandwidth of
only 2.5 GHz rather than the full 4.4 GHz. Radar transmit
chirp for these tests spanned 8 GHz to 10.5 GHz. Targets
were placed in front of the radar system at various ranges.
Range profile data was acquired, converted to complex I
and Q in software, and the discrete Fourier Transform was
taken to produce the time domain data. Round trip time
domain data was converted to linear distance from the
radar front end, and shown here. Figure 3 is a picture of
stepper motor with the experimental setup, with seven 0 dBsm cylinders
drive gear mounted placed in the snow (range profile data was acquired
planetary gear set outdoors during the winter). Coherent background
subtraction was used in all range profile experiments.
Figure 7: An old cordless drill planetary gear
transmission was utilized to multiply the torque from a
low cost biphasic stepper motor in order to
automatically position the garage door opener based
linear rail.
Figure 11 shows a range profile of seven 0 dB/sm
cylinders placed in a staggered line spaced every 1 ft,
starting at a range of 7 ft and ending at a range of 13 ft.
From this range profile plot, the position of each of the
seven cylinders is clearly indicated. There is a slight
range error probably due to slant angle of radar and the
staggering of cylinders.

Figure 9: Range profile experimental setup showing


seven 0 dBsm cylinders spaced every 2 ft placed in the
snow.

Figure 10 shows a range profile of seven 0 dBsm


cylinders placed in a staggered line spaced every 2 ft,
starting at a range of 7 ft and ending at a range of 19 ft. Figure 11: Seven 0 dBsm cylinders spaced every 1 ft.
From this range profile plot, the position of each of the
seven cylinders is clearly indicated. The first two In both figures 10 and 11 it is clear that the first cylinder
cylinders are slightly lower in amplitude than the last five. at 7ft shows up at approximately 13ft on the radar display.
There is a slight range error probably due to slant angle of This is due to a constant delay internal to the radar system
the radar to the ground. In these experiments, the radar is due physical parts layout and cable lengths. This delay is
approximately 2 ft above the line of cylinders. Also, the easily calibrated out later when SAR imagery is made
cylinders are staggered slightly (as shown in Figure 9) so using this system.
that the maximum amplitude return occurs.
5. Synthetic Aperture Radar imagery
SAR imagery was created using 4.4 GHz of chirp
bandwidth from 8 GHz to 12.4 GHz. Aperture spacing on
the 96 inch linear rail was 0.5 inches per range profile, for
a total of 192 range profiles in the data acquisition matrix.
The range migration algorithm (RMA) written directly
from [6] was utilized as the imaging algorithm.

Radar imagery was created using coherent background


subtraction and calibration to an 18 inch tall 3/8 inch
diameter aluminum dowel. Figure 12 shows the radar
image of a 1:72 scale model B52. Figure 13 shows the
radar image of GO STATE written in pushpins. Figure 14
shows the radar image of GO STATE written in 4.37 mm
diameter steel spheres.

Figure 10: Seven 0 dBsm cylinders spaced every 2 ft.


has excellent imaging capabilities, both high resolution
and high sensitivity, for a very low cost.

6. Conclusions and Future Work


From the results presented in this paper it is clear that the
low cost high resolution X-band laboratory radar system
is a capable linear rail SAR imaging system. This system
has high resolution capabilities. This was shown in the
detailed radar image a 1:72 scale model of a B52. This
system is also sensitive, shown capable of imaging small
targets such as pushpins and 4.37 mm steel spheres.
Innovations such as utilizing a discarded garage door
opener and a transmission from an old cordless drill
allowed the MSU Electromagnetics Group to develop a
rail SAR on the budget of approximately $240. Potential
Figure 12: Radar image of a 1:72 scale model B52. future work on this project will include studies on
advanced motion compensation and Autofocus
techniques.

7. REFERENCES

[1] M. P. G. Capelli, “Radio Altimeter,” IRE Transactions on


Aeronautical and Navigational Electronics, vol. 1, pp. 3-7; June
1954.

[2] M. I. Skolnic, Radar Handbook. New York: McGraw-Hill,


1970.

[3] G. L. Charvat, L. C. Kempel. “Synthetic Aperture Radar


Imaging Using a Unique Approach to Frequency-Modulated
Continuous-Wave Radar Design.” IEEE Antennas and
Propagation Magazine, February 2006.

Figure 13: Radar Image of GO STATE made out of [4] G. L. Charvat, “A Unique Approach to Frequency-
pushpins. Modulated Continuous-Wave Radar Design.” East Lansing MI:
A thesis, submitted to Michigan State University, 2003, in
partial fulfillment of the requirements for the degree of Master
of Science.

[5] G. L. Charvat, L. C. Kempel. “Low-Cost, High Resolution


X-Band Laboratory Radar System for Synthetic Aperture Radar
Applications.” East Lansing, MI: IEEE Electro/Information
Technology Conference, May 2006.

[6] W. G. Carrara, R. S. Goodman, R. M. Majewski, Spotlight


Synthetic Aperture Radar Signal Processing Algorithms.
Boston: Artech House, 1995.

[7] http://www.geniecompany.com/

Figure 14: Radar image of GO STATE in 4.37 mm


diameter spheres.

From the results shown in figures 12, 13, and 14 it is clear


that the low cost, high resolution, X-band linear rail SAR

You might also like