Multiple Target, Dynamic RF Scene Generator

David J. Wayne, Scott T. McBride, John T. McKenna

NSI-MI Technologies
Suwanee, GA, USA

[email protected], [email protected], [email protected]

Abstract- The evaluation of RF Sensors often requires a test

capability where various RF scenes are presented to the Unit
Under Test (UUT). These scenes may need to be dynamic,
represent multiple targets and/or decoys, emulate dynamic
motion, and simulate real world RF environmental conditions.
An RF Scene Generator can be employed to perform these
functions and is the focus of this paper. The total test system is
usually called Hardware in the Loop (HITL) involving the
sensor mounted on a Flight Motion Simulator (FMS), the RF
Scene Generator presenting the RF Scene, and a Simulation
Computer that dynamically controls everything in real time.
This paper describes the system concept for an RF Scene
Generator that simultaneously represents 4 targets, in highly
dynamic motion, with no occlusion, over a wide range of power,
frequency, and Field of View (FOV). It presents the test results
Figure 1. Block Diagram showing RFSG within the HITL
from a prototype that was built and tested over a limited FOV,
while being scalable to the total FOV and full system capability.
The primary component of the system is the RF Scene
The RF Scene Generator employs a wall populated with an
array of emitters that enables virtually unlimited velocity and
Generator (RFSG). Its purpose is to present an RF scene that
acceleration of targets and employs beam steering to provide simulates the motion of multiple RF emitting targets by
high angular resolution and accuracy of the presented target controlling the az-el Angle of Arrival (AoA) of each target’s
positions across the FOV. approximated plane wave. Each simulation frame can have
independently commanded azimuth, elevation, RF carrier
Keywords: RF Target Simulator, HITL, RF Scene Generator, frequency, and polarization for each of the targets, enabling
Multiple Targets, Beam Steering Wall of Emitters, Steering Array apparent motion with high dynamics.
Calibration, Plane-Wave Generator, Radar Environment
Simulator.
An important requirement is that multiple targets must be
able to share an az-el aspect without occlusion, such that all
targets would always be visible from the quiet zone. This
I. OBJECTIVE requirement discourages the use of a single moving antenna
This paper describes the system concept for an RF Scene per target.
Generator (RFSG) intended to provide significant upgraded
capability for an existing HITL facility and presents the results II. SYSTEM DESCRIPTION
of proof-of-principal testing of a prototype. The objective of
the prototyping effort was to demonstrate the RF Scene The requirement for very fast multiple-target motion
Generator (RFSG) performance against demanding system without occlusion over the FOV drove the system solution to
requirements, utilizing a scalable design representative of the be a wall of emitters rather than the more conventional
final system, to reduce the risk of building the ultimate system. mechanical motion of emitters [1] and [2]. The wall of
As a result of the successful prototype tests, the full system is emitters has no moving parts. The RF scene generator is
currently under development and scheduled for completion in essentially a very large array capable of subarray generation
2018. Figure 1 shows the block diagram of the HITL system. of multiple targets. Figure 2 shows a representative full-FOV
system. It shows several hundred emitters, arrayed on a
cylindrical wall. It shows the system concept including racks
of associated RF modules, operational computer,
communication network and wall structure on which the
emitters are mounted.
 Figure 2. Full RFSG System Concept Figure 3. RFSG Beam Forming and Switching Matrix

Four-by-four subarrays are used to create one emitting IV. CREATING AND STEERING THE RF SCENE
target per subarray anywhere on the wall. With four targets,
this means that no more than 4*16=64 elements are radiating A subarray centered near the commanded az-el
at any one time. The elements on the wall represent a regular coordinates is fed the RF signal. Complex weights are applied
az-el grid of aspects with fixed angular spacings Az and El. to the subarray’s elements in order to control the signal
For a spherical quiet zone, Az=El. strength, direction of arrival, and flatness of the phase. The
number of elements needed in that subarray, and also the
Each subarray provides an area of coverage equal to spacing among the elements, depend at least on the range
Az*El. Continuous target motion across the wall’s FOV length, the quiet-zone size, the phase flatness required, and the
involves energizing the appropriate subarray, steering the maximum frequency to be tested. For the geometries we have
beam across its area of coverage, then dropping off a row so far addressed, a 16-element (4x4) subarray has been
and/or column of the subarray and energizing the next appropriate. The height and width of the wall need to be such
appropriate row and/or column. This process is followed that a subarray exists at each corner of the desired field of
across the FOV. view.
The subarrays are capable of emitting up to 4 targets When shown a 4x4 subarray of elements, it may seem
simultaneously. Each target signal can have independent reasonable for one familiar with antennas to think of this
frequency, power and modulation characteristics. subarray in the paradigm of a phased array or beam-forming
array. While it is those things, an important difference is that
There are two options for controlling polarization. The
the beam it is forming is always pointing at the quiet zone.
first, switched H-V, is to merely permit switching between the
Another important difference is that rather than forming a
two element ports, for the azimuth (H) or elevation (V)
beam in a particular direction with respect to the subarray, it
polarization components. The second option, generalized
is forming a ‘beam’ with a particular phase slope (and thus a
polarization, doubles the number of vector modulators (and
desired AoA) in the quiet zone. Because the element spacings
RF cables, amplifiers, and summing junctions) and provides
are several wavelengths in each dimension, there will also be
independent complex weights to the two ports of each
grating lobes in directions away from the quiet zone. Those
energized element. The generalized-polarization option costs
grating lobes, however, are of no interest other than the
significantly more than the switched H-V approach. The H-V
potential for stray signals that scatter back into the quiet zone,
approach was used in the RFSG prototype.
and that potential is no worse than if radiating from just one
of the elements. The potential to steer the subarray’s beam via
III. OPERATION planar phase tilts across the elements is also not relevant,
The HITL Simulation Computer determines and sends the because we always want the power directed to the quiet zone.
desired real-time AOA commands and RF signal to the RF Rather than viewing the subarray within the paradigm of a
Scene Generator. The function of the RFSG’s Operational phased array, it may be more informative to think of the quiet-
Computer, in response to the AOA commands, is to calculate zone field as a linear superposition of the contributions from
the appropriate element weighting functions for the subarray the subarray’s elements. If the subarray is in the far field, and
and, by commanding a switching matrix, route the RF signal an element exists at the commanded az-el aspect, then that
to the appropriate emitters on the wall. The RFSG uses a single energized element can produce the desired plane wave
combination of digitally controlled vector modulators and and the other subarray elements would be fully attenuated. If
attenuators to shape the signal, and strategically placed RF the elements are closer together than can be resolved by an
amplification and attenuation to achieve the desired dynamic antenna the size of the quiet zone, then an approximated plane
range while maintaining a low noise floor. A Layer II Gigabit wave can be produced between two elements as a weighted
Ethernet network is used to communicate with the digitally sum of those two elements’ plane waves. The desired phase
controlled components throughout the RFSG. of such a plane wave from a 20o aspect is shown in Figure 4.
Note that we do not care about the field outside the spherical target can be routed to as many as 64 elements (1024 elements
bounds of the quiet zone. Provided that each element points max). In order to support multiple targets that can potentially
toward the quiet zone and that the element weights do not steer overlap, a summing junction is needed between the VM
the beam outside the quiet zone, this general approach ensures Module output and the element input. Those summing
that the resulting quiet-zone field will be at least as strong as junctions, together with part of the switching network, are
the field from any other direction. located on a Summer Module. Amplifiers are also
strategically placed in the chain to provide the desired
combination of signal power, noise power, and spurious-free
dynamic range in the quiet zone.

V. PROTOTYPE TEST HARDWARE

The objective of the prototyping effort was to demonstrate
the RF Scene Generator (RFSG) performance against
demanding system requirements, utilizing a scalable
prototype, to reduce the risk of building the ultimate system.
The Prototype Test Hardware is shown in Figure 6. It involves
an array of 16 broadband emitters mounted on tripods, a rack
containing electronics for beam forming, matrix switching,
and communication, plus a control computer and associated
Figure 4. Top view of desired plane-wave phase RF cabling. The hardware is as representative as possible and
scalable to the full system, including 77 ft. long RF cables
from the equipment rack electronics to the emitter array. The
In the general case, each element contributes a spherical wave tripods incorporate adjustable mechanisms, providing
to the quiet zone. A set of complex element weights is needed flexibility to configure the emitter array in various ways. The
that sums these spherical waves to approximate the one emitters could be re-positioned in height, spacing and angle
desired plane wave. We use a least-squares fit of the elements’ with respect to each other and the UUT. Tests were performed
QZ fields to the desired complex plane wave at the locations in 4 x4 as well as 1x16 array configurations.
where the element patterns are sampled. In this general case,
the plane wave will be made to come from the inner square of
elements represented by the green box in Figure 5. When the
commanded aspect corresponds to one of these four element
locations, nearly all of the power will come from that one
element. The contributions from the other elements will serve
to compensate for the primary element’s spherical phase front.

Figure 6. Prototype Test Hardware

Figure 5. 4x4 subarray The Prototype was installed in a HITL anechoic chamber
which is the full system’s eventual destination. The 16 emitter
array was placed 40 ft from the QZ consistent with the
Each active element’s weight is effected by a combination of geometry of the final system. To calibrate, the amplitude and
a PIN attenuator and a vector modulator (VM) located in a phase response through each RF path, as well as the I and Q
VM Module. Provided that each component has been response of each vector modulator, was characterized in 100
adequately characterized, it is then straightforward to find the MHz increments. Based on the results, a calibration table
commands that will yield the desired complex weights. The relating element weighting values as a function of AoA and
primary remaining task is to route the weighted RF signals to frequency was constructed. An X-Y scanner was placed in the
the appropriate set of elements. In the geometries we have so QZ and used for calibration and performance (verification)
far addressed, each of the 16 VM modules’ outputs for each testing.
 Figure 7 shows the block diagram of equipment used to columns represent the change of a particular dot from its level
create the calibration information for the array and to verify in the first test at that polarization.
performance. An X-Y scanner and the MI-350 instrumen-
Figure 8 shows the measured azimuth AoA error while
tation with an active-antenna interface[3] allow efficient
Figure 9 shows its repeatability over the 15-hour span.
measurement of the field in the QZ for calibration and
verification purposes. The RFSG was designed for test at the
outset, and that combined with this measurement efficiency
allowed extensive diagnoses and a thorough checkout of
performance.

Figure 8. AoA Accuracy in Elevation

Figure 7. Calibration and Verification Equipment

VI. TEST RESULTS Figure 9. AoA Repeatability in Elevation Commands

A primary function of the RFSG is to present plane waves
from specified aspects. The accuracy and repeatability of Figure 10 shows the RFSG’s consistency in gain with
those angles of arrival (AOA) is often critical. It is also aspect, frequency, and polarization. Figure 11 shows the
important that the RFSG’s gain from its input to the quiet zone repeatability of the data in Figure 10.
is insensitive to AoA, frequency, and polarization.
A comprehensive set of tests was performed exercising
the Prototype RFSG across its operating parameters and
ensuring repeatability. In one overnight acquisition, AoA,
received-power, and phase-ripple data were collected over a
15-hour period, continually measuring across 6 frequencies
and 36 target positions. Each column of 216 dots represents
the result for each beam-frequency state for the acquisition
that started at the indicated time. Commanded polarization
was toggled with each acquisition. Each acquisition moved
the X-Y scanner through a raster covering the central slice
through the quiet zone, cycling through the 216 beam-
frequency states at each position increment. Points acquired
outside the circular quiet-zone boundary were ignored. The
Figure 10. Amplitude Flatness With Time, Frequency,
AoA was determined from the corresponding slope of a best-
Polarization, and AoA
fit plane through the unwrapped phases. Phase ripple was
calculated as peak-to-peak deviations from that best-fit plane.
The following plots are a sample of the measured data.
Each of the 4320 dots in each plot represents a measurement
for a specific beam-frequency state. In the repeatability plots,
the first two columns are identically zero, and the other
 N’0 is Noise in a 1-Hz bandwidth at input to RFSG Tx
antenna terminals (each element)
GT is RFSG Tx antenna gain
GR is Gain of receive antenna
is wavelength
R is Chamber path length
E is Effective number of “hot” elements when 64
elements are radiating noise from nearly the same aspect,
after attenuation by the test article’s pattern
Figure 11. Repeatability of RFSG gain Bw is Receiver Bandwidth
MDS is test article’s Minimum Detectable Signal
Figure 12 shows the quiet-zone phase ripple about the best-
fit plane wave for each beam-frequency state. Figure 13 shows Margin is SNR for MDS
the repeatability of the data in Figure 12.

Figure 14 illustrates the multiple active elements’ radiated

noise into the test article’s receive pattern.
RFSG

R
N’0 R RCVR
R
R UUT

GT
Figure 12. Phase ripple about best-fit plane wave Figure 14. Noise transmission

Figure 15 shows the measured radiated noise N’0 (single

element, 1 Hz bandwidth) of the prototype over frequency.
-120
-125
-130
-135
-140 Measured Noise at Input
dBm/Hz

-145 to RFSG Tx Antenna

Terminals, N'0
-150
Floor of this measurement
-155
-160
-165
-170
Figure 13. Repeatability of phase ripple 5 6 7 8 9 10 11 12 13 14 15 16
GHz

VII. NOISE FLOOR

Figure 15. Measured noise at element input
A key RFSG performance requirement is that any radiated
noise within the receiver’s bandwidth must be less than the
UUT Minimum Detectable Signal level (MDS) by some
margin (Margin). This requirement is described in Equation
(1):

𝜆 2 𝑀𝐷𝑆
(𝑁0′ 𝐺𝑇 )𝐺𝑅 ( ) 𝐸 ≤ 𝑀𝑎𝑟𝑔𝑖𝑛∗𝐵𝑤 
4𝜋𝑅

where
 VIII. DEMONSTRATED FIGURES OF MERIT currently under development and scheduled for completion in
Key figures of merit demonstrated in this effort are 2018. Future papers are planned to report findings of interest.
summarized in TABLE 1.

TABLE 1. DEMONSTRATED PERFORMANCE REFERENCES

Figure of Merit Description Demonstrated [1] Wayne D., Tellakula A, Cawthon G., Langston J., Pinson C., Awadalla
Frequency Range The ability to meet 5 to 15 GHz M. “RF Target & Decoy Simulator”, AMTA Proceedings 2011
performance over [2] Wayne D, Schlegel A., Nichols, S., Bodnar D., Hess D., Cawthon G,
bands of interest Tellakula A., Skinner J., Dam T., Myles B., Gratt H., “Mobile RF
Target Motion The ability to achieve  > 2000/s Target Simulator”, AMTA Proceedings 2009
virtually unlimited  > 4,000,000/s2 [3] McBride, S.T., “An interface between a near-field acquisition system
accelerations and velo- No occlusion and active arrays with digital beamformers”, AMTA Proceedings 2012
cities of motion with no
multi-target occlusion
AoA Accuracy & The ability to produce a ≤ 0.1 Accuracy
Repeatability plane wave from the < 0.025 Repeatability
commanded az-el di-
rection
Number of The ability to 2
Targets simultaneously
simulate multiple
targets without
occlusion
Power The ability to produce a < ±0.6 dB Gain
Consistency power flux density that < ±0.1 dB Repeatability
remains constant with
aspect, frequency, and
polarization
Radiated Noise The ability to keep the N0’ ≤ -150dBm/Hz (per
Floor amplified noise element)
reaching the quiet zone
‘low enough’ that it
does not interfere with
receipt of low signals
Instantaneous Amount the RFSG IDR > 100 dB
Dynamic Range input can be decreased
before the test article is
confused by noise
Instantaneous Amount the RFSG IBW ≥ 60 MHz
Bandwidth input’s frequency can
be varied from the
commanded carrier
before AoA or power
degrades
Update Rate The maximum frame > 1 KHz
rate of the simulation
RF Settling Time The amount time in < 10 s
transition between
simulation frames
Quiet-Zone Size The volume of the 1 ft3
sphere over which the
plane waves are
approximated and
verified

IX. SUMMARY
A system concept for an RF Scene Generator has been
presented, capable of highly dynamic, multiple target motion,
without occlusion, and capable of emitting wide band signal
modulation, changing intensities and polarizations. A Risk
Reduction Prototype was described and test results presented.
The results demonstrated successful performance in numerous
important figures of merit, across a range of operating
conditions, establishing confidence that the prototype is
scalable to full system implementation. The full system is