Welcome to the AMTA paper archive. Select a category, publication date or search by author.
(Note: Papers will always be listed by categories. To see ALL of the papers meeting your search criteria select the "AMTA Paper Archive" category after performing your search.)
The planar wide-mesh scanning (PWMS) methodology is based on Non-redundant scanning schemes allowing faster measurements than classical Nyquist-compliant acquisitions based on denser, regular, equally spaced Near Field (NF) sampling. The methodology has no accuracy loss and has been validated at different bands and with different antennas [1]. The effectiveness of the PWMS technique has always been proven in error-free (or quasi-error-free) scenarios, assuming that possible errors introduced by the technique itself are independent of the typical source of measurement uncertainty. In this paper, we investigate for the first time the sensibility of the method wrt one of this error source included in the 18-terms lists [2], considered by the measurement community as an exhaustive list of the NF errors: X and Y probe positioning errors. Such errors are unknown and random and are associated to the mechanical vibrations and/or backlash of the system. The investigation has been done considering actual measurements of a multi-beam reflector antenna with approximately 35 dBi gain (MVG SR40 fed by two MVG SH5000 dual ridge horn). The AUT has been measured in planar geometry emulated by a 6-axis Staubli robot. The test was performed at 22-33 GHz. A set of measurements has been performed introducing a uniformly distributed random error in the range [0-1] mm, corresponding to ?/10 at 30 GHz. Errors are considered unknown. In the paper it will be shown that both in the classical and PWMS approaches the main beam is basically not affected by the introduced errors. The sidelobes are instead affected by such errors especially in the pattern cut where the beam is tilted. Such error levels obtained with the classical approach are comparable to those obtained with the PWMS approach, meaning that the latter is stable and against such type of perturbations.
NF-FF transformations have proven to be a convenient tool to accurately reconstruct the antenna pattern from NF measurements. In this framework, a very hot issue is the reduction of the time required to perform the measurements. To obtain a remarkable reduction of this time, nonredundant (NR) NF-FF transformations with planar spiral scannings have been developed in [1], by applying the NR representations of electromagnetic fields [2]. Optimal sampling interpolation (OSI) formulas have been used to efficiently reconstruct the massive NF data for the classical plane-rectangular (PR) NF-FF transformation from the NR spiral samples. The drastic measurement time-saving is due to the reduced number of needed NF samples acquired on fly, by adopting continuous and synchronized motions of the linear positioner of the probe and of the turntable of the AUT. However, such a time-saving is obtained at the expense of a nonuniform step of the spiral. Therefore, the linear positioner velocity is not constant, but must vary according to a not trivial law to trace the spiral, and this implies a complex control of the linear positioner. This work aims to develop an effective NF-FF transformation with planar spiral scanning for volumetric AUTs, wherein the spiral step is uniform and, hence, the linear positioner velocity becomes constant. To this end, the AUT is considered as enclosed in a sphere, the spiral is chosen in such a way that its step coincides with the sampling spacing needed to interpolate along a radial line according to the spatial band-limitation properties, and the NR representation along such a spiral is determined. Then, an OSI algorithm is developed to recover the NF data needed by the PR NF-FF transformation from the spiral samples. Numerical simulations assessing the accuracy of the developed NF-FF transformation will be shown.
Spherical near-field systems installed in shielded anechoic chambers are typically involved in modern automotive antenna measurements [1-3]. Such systems are often truncated at or close to the horizon to host the vehicle under test while limiting the size/cost of the chamber. The vehicle is usually placed on a metallic floor [4] or on a floor covered by absorbers [5]. The latter solution is intended to emulate a free space environment and is a key factor to perform accurate measurements down to 70 MHz. The availability of the free-space response also enables easy emulation of the car's behaviour over realistic grounds [6-7] while such emulations are more complex when a conductive ground is considered [8]. Conductive ground measurements also suffer from a strong interaction between the conductive floor and the measurement system and only in a limited number of situations such types of floor are a good approximation of realistic grounds (such as asphalts). However, the main advantage of conductive floor systems is the ease of accommodation of the vehicle under test which is simply parked in the center of the system. In absorber-based systems, instead, more time is generally needed to remove/place the absorber around the vehicle. Moreover, at low frequencies (70-400 MHz), large and bulky absorbers are normally used to ensure good reflectivity levels and the vehicle needs to be raised to avoid shadowing effect of absorbers. In this paper we investigate whether the measurement setup phase in absorber-based systems can be simplified by using smaller absorbers at low frequencies and/or not using them at all but considering conductive floors. The loss of accuracy in such scenarios will be studied considering a scaled vehicle and an implemented scaled automotive system where it is possible to access the full-spherical, real free-space scenario which is used as reference. The analysis is carried out considering (scaled) frequencies relevant to automotive applications in the 84-1500 MHz range. Two types of scaled absorbers, of different size and reflectivity, are considered to emulate the behaviour of the realistic full-scale 48-inch and 18-inch height absorbers. Measurements over metallic floor are included also in the analysis.
We present an antenna measurement methodology requiring a radically lower number of field samples than the standard Nyquist-based theory maintaining a comparable accuracy. simulations and partial knowledge of the geometry of the Antenna Under Test are combined to build a set of numerically defined expansion functions: the method uses basic knowledge of the antenna and the assumption that scattering from large surfaces can be predicted accurately by numerical tools; areas of the antenna such as feeding structures are treated as unknown and represented by equivalent electric and magnetic currents on a conformal surface. In this way, the complexity, and thus the number of unknowns, is dramatically reduced wrt the full problem for most antennas. The basis functions representing the full antenna are used to interpolate a radically reduced set of measured samples to a fine regular grid of Near Field (NF) samples in standard geometries. Regular NF to Far Field (FF) transformation techniques are then employed to determine the FF. The sampling reduction is evaluated compared to a regular sampling on standard Nyquist-complaint grids. The method can be employed in standard sampling ranges. In [1] asymptotic simulation tools were used to build the numerical basis. In this paper, methods based on Surface Integral Equations (SIEs) are used to compute currents and fields. The currents induced on the antenna structure by each elementary source are computed and used to evaluate the radiated field. Both electric and magnetic elementary sources are placed around the antenna and the SIE problems use a fast algorithm to evaluate matrix-vector products. The methodology is validated with planar and spherical acquisitions on a reflector antenna (MVG SR40) fed by a dual ridge horn SH4000 and in a multi-feed configurations (using several SH5000) at 18 and 30 GHz. Patterns obtained with down-sampled fast approach are compared to standard measurements. Down-sampling factors up to 8 are achieved maintaining very high correlation levels with standard techniques.
Characterizing the performance of automobile-mounted antennas has been an ongoing and evolving challenge for the antenna measurement community. Today, the automotive test environment poses unique challenges with its diversity and complexity of wireless on-board systems and the large electrical size of the test article. The evolution of cellular technologies over the past decade means that the basic mobile handset has now become a smartphone with significantly increased capability; this exact same trend has been mirrored by the automotive industry where we have witnessed the basic car radio and cassette player evolve into a multi-function infotainment unit. Modern vehicles include a multitude of wireless technologies, including cellular (2G, 3G, LTE), Bluetooth, WiFi, Global Navigation Satellite System (GNSS), collision avoidance radar, and more. Testing the complete vehicle is currently the only method available that certifies the correct mode of operation for each technology (including co-existence and interference) and also assures the manufacturer that the various sub-systems are performing as expected in the presence of all other sub-systems and the vehicle itself. While modern vehicles now function like large mobile devices, the conventional Over-the-Air (OTA) measurement systems and techniques available for small form factor devices (e.g. mobile phones) are ill-suited to testing such large devices. In this paper, we will highlight some of the unique challenges encountered in the automotive test environment. We will start by looking into existing methods of measuring radiation patterns of automobile-mounted antennas and providing a qualitative assessment of the various techniques with a focus on near-field solutions. A brief description of OTA testing will follow, coupled with an in-depth look into how techniques that are proven for handset type OTA measurements are being translated to automotive measurements. This section will provide a breakdown of key OTA test metrics, the measurement hardware typically required and key assumptions about the device under test. Finally, some performance tradeoffs and challenges associated with designing a multi-purpose antenna/OTA measurement system will be described.
Anechoic chamber performance is largely dependent on the chamber layout and effectiveness of the RF absorbing material. Over time, absorbers begin to degrade and need to be replaced. Also, new health and safety standards, combined with more modern building fire code regulations, can necessitate updates to a chamber's absorber layout. These factors led to the complete restoration of the Canadian Space Agency's largest anechoic chamber, which is located at the David Florida Laboratory in Ottawa, Canada. Ideally, the internal walls and ceiling of an anechoic chamber should be free of intrusions to facilitate the installation of RF absorbers. Unfortunately, because of the chamber's coupled structure to the building, this was not possible and an extensive sprinkler system and large air circulation vents were installed within the chamber. In this paper, their respective impacts on reflectivity performance is studied and novel mitigation techniques are introduced. Based on practical considerations, these techniques were used to conceal the infrastructure of the fire suppression system and HVAC ductwork inside the anechoic chamber; initial measurements appear to indicate their validity. The techniques and lessons learned from this exercise may be applied by others considering a similar endeavor.
The problem of modelling a radiator or a scatterer using an equivalent radiator is of interest in a large number of applications as, for example, antenna synthesis [1], electromagnetic compatibility [2] and the design of echo generators [3]. Such a modelling problem requires determining the shape and the dimensions of a radiating surface capable to generate, in a certain region of space, an electromagnetic field close to that produced by the radiator/scatterer of interest. The purpose of this contribution is that of proposing, for a fixed equivalent radiator's shape, an approach for the solution of such dimensioning issue. The solution is new and, at the best of our knowledge, the dealt with problem has not been yet addressed throughout the literature. The proposed approach relies on the use of the Singular Value Decomposition (SVDs) of the operators linking the radiator/scatterer to the field on the region of interest, say D, and the equivalent radiating panel to the field on D, again. The singular functions of such operators corresponding to the most significant singular values represent the spaces to which the fields radiated by the primary radiator/scatterer and that radiated by the equivalent one essentially belong to, respectively. The approach consists into determining the dimensions of the equivalent radiator minimizing the error by which the field radiated on D by the equivalent radiator approximates the primary radiated/scattered one. The error is expressed as a hermitian, definite positive quadratic form so that the problem amounts to the maximization of its minimum eigenvalue. Numerical results will be presented for an equivalent, planar radiator of rectangular shape.
Spherical Near-Field (SNF) systems using a swing arm gantry configuration have been the go to solution for automotive measurement systems. Recent advances in the automotive industry have warranted a need for SNF systems with high mechanical positioning accuracy supporting measurements up to 40 GHz and beyond. This paper presents the design and implementation of a new swing arm gantry positioner having an 8-meter radius and a radial axis to support high frequency SNF measurements. We first define the relation of the gantry axis to the global coordinate system and discuss primary sources of errors. Next, a robust mechanical design is presented including design considerations and implementation. We then present errors measured using a tracking laser interferometer for probe position through the range of gantry axis travel. Static corrections for probe positioning errors are implemented in the control system using the radial axis. The resultant residual error for the swing arm gantry is then shown to have the accuracy required for high frequency SNF measurements.
The Austin RCS Benchmark Suite has recently been introduced to enable quantitative and objective comparison of computational systems for solving electromagnetic scattering problems, particularly, those relevant to aerospace applications. In the last year, five sets of problems were added to it: dielectric almonds (problem set III-B), mixed material almonds (III-C, III-D), perfectly electrically conducting (PEC) aircraft models (IV-A), and dielectric aircraft models (IV-B). For each problem set, a range of lengths and frequencies of interests are identified, interesting features are highlighted, and datasets containing reference results (from measurements, analytical methods, or numerical methods) are shared online. Although data from several radar cross section (RCS) measurement campaigns of non-metallic targets are available in the literature, these lack the information necessary to precisely model the materials, target geometries, and measurement setups, to quantify uncertainties in the data, and to identify appropriate directions for improving computational methods' performance. This limits their utility for benchmarking computational systems. This article presents an expansion of the Suite to include problems with more complex materials and reference results from a measurement campaign that attempted to ameliorate the deficiencies of existing datasets. Specifically, a set of thin-plate problems are added to the Austin RCS Benchmark Suite to increase material diversity. These consist of problem sets II-B: thin perfect-electrically conducting (PEC) plates, II-C: thin dielectric plates, II-D: thin magnetic radar absorbing material (MagRAM) plates, and II-E: thin MagRAM-coated PEC plates. Reference RCS data that enables validation, RCS measurement and material property uncertainty quantification, and benchmarking are also provided by conducting a simulation-supported measurement campaign in a compact range. To facilitate reproducibility, a popular low-loss dielectric material and a commercially available MagRAM were chosen for these problems: The dielectric material for problem set II-C is PolymaxTM polylactic acid (PLA). For problem sets II-D/E, the ARC Technologies' DD-13490 material is used. Thin plates were manufactured and their RCS were measured at Lockheed Martin's Rye Canyon Facility. The monostatic RCS measurement results and supporting simulation results are available online. Performance data for simulations as well as RCS measurement results with accompanying uncertainty will be presented for problem sets II-B/C/D/E at the conference.
A set of new 3:1 Dual-Polarized Antennas has been developed for use in near-field ranges as the probe or range antenna and for use as a Compact Antenna Test Range (CATR) feed [1]. Key development parameters of the antenna are: a wideband impedance match to the coaxial feed line, E and H-plane 1 dB beam widths in excess of 30 degrees, -30 dB on axis cross-polarization, minimum polarization tilt and a phase center that varies over a small region near the aperture. To accomplish these design parameters, a family of range antennas has been developed and previously introduced [1]. Two versions of the antenna have been manufactured and tested for performance. A 2-6 GHz version has been developed using traditional machining techniques and a 6-18 GHz version has been produced using additive manufacturing (3D printing) techniques. In this paper, we focus on the performance of the 6-18 GHz antenna produced using additive manufacturing. The measured performance of the antenna will be presented and compared to previous simulation. The advantages of additive manufacturing for this type antenna will be discussed. Finally, the applicability of the antenna as a CATR feed and its use in near-field scanning will be discussed.
Tapered Chambers are best suited for antenna pattern measurements at low frequencies. The advantage of such chambers over rectangular shaped chambers would be achieving a desired performance level in terms of field uniformity and ripple at the quiet zone due to the low reflectivity of the chamber. To achieve such performance using a rectangular shaped chamber could lead to design of larger rooms and associated significant cost. Hence, this paper tries to analyze the characteristics of the tapered chambers using the novel Fourier analysis characterization method. The Fourier analysis method was applied on the transverse and longitudinal planar scan data at the quiet zone of a tapered chamber. The analysis yield the performance of the chamber at different frequencies depicting the plane wave behavior at the low frequencies and breakpoint of the plane wave behavior with increase in frequency. It also shows the images or reflection hotspots formed at the throat of the tapered section at the higher frequencies. In addition, the longitudinal scan analysis portray the reflections from the back wall of the chamber. In conclusion, the known concepts and ideologies of the tapered chamber design are reexamined from a different perspective based on the analysis results.
There is a growing need for wearable RF modules integrated into clothing for medical sensing applications. Also, there is a concurrent interest for RF devices to collect energy and power body-worn devices, such as biosensors used to monitor vital signs. To do so, an approach is to integrate into garments, near-field wireless power transfer and harvesting RF circuits that can collect ambient RF energy from nearby Wi-Fi sources to charge biosensors. In this regard, resonant RF harvesting structures for near-field power transfer have been explored before and demonstrated close to 100% of power transfer efficiency (PTE) at close distances. However, this impressive efficiency can only be realized when there is perfect polarization and special alignment between transmitter and receiver. Herewith, we provide an approach that mitigates the efficiency challenges due to misalignment. With the above in mind, we propose a new class of resonators referred to as corrugated X resonators that are resilient to lateral, angular, and diagonal misalignments. The resonators operate at 500 MHz. Using these resonators, their PTE was found to range from 40% to 60% across 1 to 10-cm distances for broadside direction. In case of only lateral misalignment in the direction perpendicular to the feedline, the PTE is 70% across the same distances. Also, a PTE of up to 85% was achieved when the misalignment was only in the direction of the feedline. At the meeting, we will present the design and performance of the developed low-profile resonators RF surfaces. The evaluation is done when integrated into wearable textile surfaces under near-field illumination for RF harvesting.
In the 5G communication, antenna array has been widely used for high-speed wireless communication. For reliable antenna array system, the failure diagnosis of antenna array is one of the most important problems that has been studied for a long time. The back-projection method using near-field measurements is a one of the failure diagnosis technique based on the plane-wave expansion. However, when antenna elements are densely placed, it is difficult to estimate the excitation coefficients of the antenna elements with the back-projection method, because the obtained images from the conventional back-projection method has only a resolution of one wavelength. In addition, since there is usually a trade-off between measurement accuracy and measurement time. Therefore, it is difficult to satisfy the both requirements of accuracy and short measurement time. We have reported the element failure detection algorithm using a 2-layer shallow neural network with planar near-field measurement last year. In this report, the element failure detection of antenna array is performed with a minimum number of measurement points while maintaining enough accuracy by learning the relationship between excitation coefficients of antenna array and the electric far-field distribution by a shallow neural network. In the case of 64-elements short dipole antenna arrays, the estimation error of excitation coefficients of antenna array less than 1% are achieved by our trained neural network with a minimum number of far-field measurements with 50 dB SNR. The detailed algorithm and simulation results will be reported in the full-paper and the presentation.
With the advent of small-scale satellite technologies, a significant challenge faced by antenna engineers today is the design of high-gain and low-profile lens antennas. The conventional approach to lens design has been to shape the surfaces of dielectric materials to form homogeneous dielectric lenses. Alternatively, the refractive index of the lens can be varied throughout the lens body to form Gradient-Refractive Index (GRIN) lenses, which are now conceivable because of the advent of metamaterials. Optical ray paths are controlled inside the GRIN meta-lens rather than relying only on refraction at the air-lens interface. In particular, spherically symmetric lenses (such as the Luneburg lens) can achieve much greater control over the ray trajectories than homogeneous lenses since incoming electromagnetic waves can travel in a curved path within the volume of the lens. Due to spherical symmetry, however, the material that fills the volume of the spherically symmetric GRIN lens increases dramatically, and the weight becomes impractical for applications that require highly directive antennas. In order to overcome this disadvantage, the concepts of flat meta-lenses resulting into lightweight and slimmed alternatives to the spherically symmetric lenses have recently gained attention. In this work, we first present a methodology for the synthesis of the refractive index of flat GRIN lens antennas. Previous methods to obtain the material inhomogeneity relied on the assumption that the ray path is straight within the volume of the lens. These methods are approximate and applicable only for on-axis fed lenses. Contrary to the previous methods, the applied numerical synthesis algorithm based on Geometrical Optics and Particle Swarm Optimization is used to synthesize both on-axis and off-axis fed flat lenses with circularly symmetric aperture phase thus providing conically scanned beams for the off-axis designs. Then, metamaterial elements of variable sizes distributed on planar dielectric substrates are synthesized to form a multi-layered flat metamaterial lens and satisfy the required refractive index distribution. Simulation and near-field/far-field measured results of a proof-of-concept prototype of a multi-layered flat lens operating at 13.4 GHz are also presented. Additionally, microwave holographic approach is used to evaluate the goodness of the exit aperture amplitude and phase.
5th generation mobile network will use 28 GHz band and 38 GHz band in Japan. We developed compact type antenna measurement system using optical fiber link and vertical articulated robot for 5G antenna measurement. Our developed optical fiber system consists of 850 nm multi-mode VCSEL (Vertical cavity surface emitting laser diode) and PD-TIA (Photo diode with transimpedance amplifier). Our optical fiber link can use for microwave measurement from 10 MHZ to 30 GHz with 90 dB dynamic range and up to 40 GHz with 70 dB dynamic range. Our using vertical articulated robot has 6-axis 1 m length arm that can hold WR-28 open ended waveguide probe (OEWG).
In the last few years, MIMO (Multiple-Input-Multiple-Output) Cellular Antenna system functionality for call, text and data streaming in Automobile has been developed and integrated into vehicles lines, by several Automotive OEMs. However, the performance of the MIMO Cellular Communication system in term of throughput on the uplink and downlink fluctuates, and is influenced by several dynamic network factors such as, Network traffic, time of day, network size and location/environment and which is in contrast to an (Over-The-Air) OTA chamber Environment. This paper compares the MIMO Cellular Antenna performance in Field Test with MIMO Cellular Antenna performance in an OTA Chamber Environments. The parameters that will be compared in the study measurements are throughput, SINR (Signal-to-Interference-Plus-Noise Ratio) and RSSI (Received Signal Strength Indicator).
For features of very weak scattering while masked by background and clutter, care must be taken in the measurement design as well as data processing, in order to extract the true RCS values. A good example of flush-mounted fasteners for a low-observable (LO) aircraft, arranged in an array, was reported by Lutz, Mensa, and Vaccaro [1]. In the Boeing 9-77 range, retro-reflectors (called retros) were routinely taped on a test-body for monitoring its 3-D locations and angles during measurements. Though the retro’s RCS may be several orders below that of a test-body, a challenge was to discover their exact values. In the millimeter wave range (MMWR), we measured 2-D arrays of retros arranged in both square and hexagonal lattices taped onto a flat metal surface pitched 20o down. RCS measurements were made as a function of frequency and aspect angle. From the 2-D FFT images, the nominal RCS for a retro in VV polarization was found to be -75 dBsm, independent of the geometry and number of retros even down to one unit-cell [2]. But for the HH polarization, there is no backscattering from such a flat metal surface. References [1]. J. Lutz, D. Mensa, and K. Vaccaro, "RCS measurements of LO features on a test body," Proc. 21st AMTA, pp. 320-325 (1999). [2]. P. S. P. Wei and J. P. Rupp, " RCS measurements on arrays of weak scatterers enhanced by diffraction," Proc. 26th AMTA, pp. 263-268 (2004). ---------------------------------------------- ** Sam Wei is at: 4123 - 205th Ave. SE, Sammamish, WA 98075-9600. Email: paxwei3@gmail.com, Tel. (425) 392-0175
One of the keys to developing new science and technologies is to have sound metrology tools and techniques. Whenever possible, we would like these metrology techniques to make absolute measurements of the physical quantity. Furthermore, we would like to make measurements directly traceable to the International System of Units (SI). Measurements based on atoms provide such a direct SI traceability path and enable absolute measurements of physical quantities. Atom-based measurements have been used for several years; most notable are time (s), frequency (Hz), and length (m). There is a need to extend these atom-based techniques to other physical quantities, such as electric (E) fields. We are developing a fundamentally new atom-based approach for that will lead to a self-calibrated, SI traceable E-field measurement and has the capability to perform measurements on a fine spatial resolution in both the far-field and near-field. This new approach is significantly different from currently used field measurement techniques in that it is based on the interaction of radio-frequency (RF) E-fields with Rydberg atoms (alkali atoms placed in a glass vapor-cell that are excited optically to Rydberg states). The Rydberg atoms act like an RF-to-optical transducer, converting an RF E-field strength to an optical-frequency response. In this new approach, we employ the phenomena of electromagnetically induced transparency (EIT) and Autler-Townes splitting. This splitting is easily measured and is directly proportional to the applied RF E-field amplitude and results in an absolute SI traceable measurement. The technique is very broadband allowing self-calibrated measurements over a large frequency band including 500 MHz to 500 GHz (and possibly up to 1 THz and down to 10's of megahertz). We will report on the development of this new metrology approach, including the first fiber-coupled vapor-cell for E-field measurements. We also discuss key applications, including self-calibrated measurements, millimeter-wave and sub-THz measurements, field mapping, and sub-wavelength and near-field imaging. We show results for mapping the fields inside vapor cells, for measuring the E-field distribution along the surface of a circuit board, and for measuring the near-field at the aperture in a cavity.
Understanding absorber performance in the W band (75-100 GHz) has become increasingly important, especially with the popular use of W band radars for automotive range detections. Commercial absorber performance data is typically available only to 40 GHz. Measurements performed in the W band in anechoic chambers are often under the assumptions that high frequency absorber data can be extrapolated from the data below 40 GHz. In this paper, we provide a survey of common microwave absorbers in the W band. It shows that the extrapolated data from the lower frequencies are not accurate. Absorber analysis models for low frequencies such using homogenization concept are no longer valid. This is because, for the millimeter wave, microstructures of the foam substrate become important, and the dimensions of the pyramids are much greater than the wavelengths. We examine performance variations due to parameters, such as carbon loading, shape, and thickness of the absorbers. We will also show how paint on the absorber surface might affect the absorber reflectivity, and if the common practice of black-tipping (leaving the tip of the absorbers unpainted) is an effective technique to alleviate paint effects.
One of the sources of the measurement errors in outdoor antenna test ranges, when testing from VHF through C-Band, is the ground reflected signal between probe antenna and the antenna under test (AUT). Those errors are due to antenna(s) relatively large beam width(s) at these frequencies, especially when AUT is placed on the large platform such as an aircraft. If reflected wave is not eliminated by the use of absorbers at the reflection point or redirection by the use of diffraction fences, then the range operates as a ground reflection range (GRR), where the reflected signal creates a lobbing pattern when the direct and reflected signals are overlaying in- and out-of-phase as a function of position and frequency, causing undesirable amplitude variations at the test point. Ground reflections may be a major cause of error for GRR measurements when testing large antennas or antennas mounted on large structures which require a large displacement of the AUT during the antenna pattern collection process. A concept of using vertically positioned two-element array probe antenna (source antenna) to suppress ground-reflected signals in GRR-s is presented in this article. Suppression is achieved by pointing first null of the probes gain pattern towards the reflection point on the ground. All analytical evaluations are based on geometrical optics approach. Comparison of the proposed approach to a traditional single-element probe (source) antenna approach, demonstrates a significant improvement in measurement accuracy. Estimates and verifications of analytical evaluations are based on Computational Electromagnetics (CEM) modeling tool such as WIPL-D code. Simulations are performed in the VHF frequency band (200 MHz).