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.)
Anechoic ranges require constant temperature and humidity, proper lighting to be able to work inside the range and closed-circuit television (CCTV) cameras to monitor the system while the measurement is being done. In addition, anechoic chambers require fire detection and suppression. Traditionally these penetrations are minimized and placed in non-critical areas. But the true effect of them has not been fully investigated. In this paper, antenna measurements as simulated in an indoor far field range. The approach to model the measurement is like the one the author presented in [1] and [2]. Thus, a range antenna (or near- field probe) and an antenna under test (AUT) are placed in free space and the AUT is rotated at discrete angles as it was done in [1]. Then a second model includes CCTV cameras, HVAC vents, light fixtures and both air sampling tubes and fire suppression nozzles and placed around. The simulation with these disruptions is repeated at the given discrete angles. The model does not include the absorber on the range. The model assumes a perfect absorber and the results of the simulated antenna measurement are compared to an ideal case with no disruptions. The results, while being approximations, provide a worst-case error for those disruptions of the RF-absorber layout. The results can be used to estimate the potential uncertainty on the measurement caused by the different systems that must be part of the anechoic enclosure. The technique is applied here to indoor far field measurements, and for near-field systems. Results show that for your typical roll over azimuth positioner, the effects of the penetrations on the ceiling are very small with differences in the -35 to -40 dB levels.
This paper proposes a novel approach based on a Multi-Probe Near-Field Scanner system utilizing bipolar and phased array antenna technology to measure antenna arrays with highly reliable precision. The innovative system incorporates a phased array antenna with full amplitude and phase control, enabling complete manipulation of the polarization state. This advanced capability allows the system to accurately measure and characterize antenna arrays while addressing and correcting polarization distortions caused by polar motion. By using the multi-probe configuration, the system significantly enhances measurement efficiency by capturing near-field data simultaneously across multiple probes. The integration of bipolar technology [1] ensures robust signal processing, while the phased array design enables the electronic rotation of the polarization of the probe array antenna. This feature is critical for mitigating errors introduced by polarization mismatches or distortions, ensuring high-fidelity measurements even in complex scenarios. The proposed system demonstrates superiority over traditional single-probe scanners by reducing measurement time, providing comprehensive polarization control and size reduction of the controlled environment. The paper discusses the system’s design, implementation, and performance.
In recent times, we have become familiar with the use of commercial software for designing our antennas and microwave devices. This is very common since it is easy to find high-performance desktop computers at affordable prices in our daily lives. The use of general-purpose commercial software is widespread because it allows for the simulation of any arbitrary configuration. However, many of us have experienced, given the ease of using commercial software, trying to simulate electrically large electromagnetic devices which take days or, in some cases, cannot be completed at all. While it is true that we now have very powerful simulation tools, by making a few simple assumptions, we can significantly reduce computational time without sacrificing accuracy. In this talk, I will introduce a simple ray-tracing technique that can be used, in combination with physical optics, to calculate the radiation pattern of antennas, as well as directivity, gain, mutual coupling, and even early-time response in complex configurations. The results are not only faster than those produced by conventional commercial software, but also more accurate, as they avoid many of the numerical errors that typically arise when computing electrically large structures.
Monolithic radomes for aircraft nosecones are designed for maximum transmission, minimum sidelobe increase, and minimum beam deflection at the desired radar frequency. In some cases, a radome may have a position dependent thickness to account for radar incidence angle. These radomes are manually tuned with an iterative process that uses radio frequency test range measurements to guide placement of tuning tape on to the radome. Generally, the radomes are thinner than desired, and the tuning adds an appropriate number of tape layers to center the optimum transmission window in frequency. This trial-and-error method is time consuming and range measurements are costly. Instead, this paper discusses an alternative method that is faster and less resource intensive. A microwave spot probe is raster scanned around the radome surface, and an algorithm inverts the electrical thickness at each measurement location from the measured reflection amplitude and phase. This thickness map is compared to the desired thickness and the appropriate number of tuning tape layers are calculated. This paper demonstrates the spot-probe method with a canonical wedge-shaped radome, where one side of the radome was purposely too thin. In addition, a lens-based far- field range is used to measure transmission efficiency and apparent beam deflection error of the wedge radome before and after it is tuned with fiberglass tape.
In this work, a novel two-steps technique to correct 3D probe positioning errors affecting the non-redundant (NR) spherical near-field/far-field transformation (NF/FFT) technique for quasi-planar antennas, modeled with a double bowl, is presented. The developed approach benefits from the synergy of two complementary correction steps. In the first one, the so-called NR correction compensates for the phase errors due to the deviations of the sampling points from the nominal scanning sphere. This is achieved through an ad hoc phase correction scheme, whereby an optimal phase factor is extracted from the acquired samples, accounting for their actual (i.e., erroneous) position. In the second step, an iterative approach is exploited to manage the residual 2-D errors affecting the data attained at the end of the previous step, allowing for an accurate and efficient restoration of the correctly located samples. Ultimately, the large number of input data for the classical spherical NF/FFT, uniformly spaced on the scanning spherical grid, are obtained via an optimal sampling interpolation formula. Numerical results showing the effectiveness of the proposed method in compensating even severe 3-D positioning errors are reported.
Hypersonic platforms develop skin temperatures exceeding 1500°C. Materials applied to these skins often require specific electric and magnetic properties which must be validated at temperature, requiring tailored furnaces, refractory metals, specialized ceramics, and procedures to protect RF hardware and personnel. Free space measurement configurations maximize the distance between high temperature components and measurement hardware, thus limiting errors and damage due to radiant heat. Free space material property inversions assume planewave illumination on an infinite sample, which, in practice, is approximated by a finite sample with attenuated fields at its boundaries. Far field generation using Gaussian shaping dielectric lenses, Hogg horn reflectometers, and lens covered antennas have all been previously employed. Given the need to maximize sample- to-hardware distance, an alternative room temperature measurement approach has been developed for the 8-18 GHz band as a baseline architecture for testing up to 1500°C and beyond. This baseline system utilizes two opposing ellipsoidal reflectors with a shared focus to create a localized sample plane. First published in 1971 for use with high temperature dielectric measurements, the simplicity of the elliptical shape, frequency scalability, low-cost material, and straightforward fabrication, makes this approach a good candidate for high temperature sample illumination. The baseline system extends the initial work to include magnetic materials and a discussion of primary error sources, including incident field performance versus feed displacement, reflector conductivity, feed-reflector interactions, and off-normal illumination. System performance is demonstrated through complex permittivity and permeability estimates for Neoprene, reticulated absorber, one-pound density Polystyrene, and commercial magnetic absorber. Planned modifications to the baseline system for high temperature applications are also presented.
For any measurement of spatially distributed electromagnetic fields, the use of signal averaging, to suppress influence of noise and thus increase dynamic range, in combination with on-the-fly sampling, to reduce measurement time, give rise to smearing of the measured signal since the samples are made at different spatial points and thus for different field values due to the continuous relative motion between the source of the field and the probe. This work demonstrates how the smearing error due to signal averaging with on-the-fly sampling can be exactly quantified and also exactly corrected. This correction eliminates the need for the practical approaches of reducing the smearing error – either by reducing the speed of the mechanical motion of the measurement system, thus compromising on measurement time, or by reducing the number of samples to be averaged, thus compromising on dynamic range. Instead, the correction allows for maximum speed and maximum signal averaging within the relevant sampling criterion.
This paper explores a Phase Retrieval (PR) method for automotive Over-The-Air (OTA) spherical Near-Field (NF) measurements. Two signals are acquired phase-coherently: one from a fixed reference antenna attached to the vehicle, and the other from a conventional NF probe. The phase information is retrieved from the relationship between the two signals. An off-the-shelf Software-Defined Radio (SDR) two-channel receiver module is used. This setup enables one-step NF OTA measurements, potentially improving time- and cost-efficiency. Moreover, it solves the inaccessibility of antenna connectors. This approach has been tested on a vehicle in an actual automotive OTA measurement setup with a modulated signal centered at 810 MHz, which presents a challenging and realistic scenario. This work demonstrates that phase retrieval is feasible with the proposed setup. Moreover, good agreement with a reference measurement taken using a Vector Network Analyzer (VNA) is achieved at high Signal-to-Noise Ratio (SNR) levels. E.g., phase errors with respect to the reference measurement in the NF of less than 10° at 90% of the angular sampling points can be obtained with an SNR of at least 27 dB. The corresponding transformed Far-Field (FF) EIRP shows also good agreement between the PR and VNA measurements, with an Equivalent Noise Level (ENL) of -32 dB. In the evaluated example, the direct OTA measurement in the NF yields a much less similar pattern to the reference FF than with the proposed method.
Linear motion components are widely used in antenna metrology positioning systems, but long motion ranges can reduce positioning accuracy due to the difficulty of eliminating rail component non-linearity. Further, offset device-under-test (DUT) mounting can magnify rotation errors arising from slight variations of rail parallelism not captured by a strictly linear motion model. We present methods to reduce position and rotation inaccuracy through kinematic calibration using a Legendre polynomial model. The motion range of the rail system is conditioned for use over an appropriate evaluation range. The calculation of calibrated coefficients is detailed for an individual rail calibration and is further integrated in a pose error minimization calibration framework for general robotic implementation. This approach is validated using experimental 6 degree-of-freedom (DoF) measurements collected using the Large Antenna Positioning System robotic antenna range. Compared to a strictly linear rail model, the Legendre polynomial model reduced the RMS position error by 67.7 % to 27.90 μm and RMS rotation error by 29.9 % to 84.92 μrad.
With the increasing number of channels in integrated radar MMICs, radar modules and networks, beamformer calibration techniques must adapt to the physical dimensions of these sensors. Typical far-field calibration requires measurements at the Fraunhofer distance. This ensures a maximum phase error of 22.5° over the aperture. However, literature shows that phase errors below 5° are required for acceptable side-lobe suppression. Compact Antenna Test Ranges (CATR) create virtual far-field conditions in limited space but unfortunately they introduce magnitude and phase errors in their measurement. A method for calibrating MIMO radars in CATR settings is presented using spatial averaging to reduce these errors systematically. Simulations with a 16-channel FMCW radar show maximum errors of below 0.25dB for magnitude and less than 2◦ for the phase with a single-digit number of spatial averages. Calibration with a 77-GHz MIMO radar sensor in the CATR confirms the technique’s ability to mitigate test zone non-idealities, improving radar imaging quality.
A low-cost drone kit is demonstrated for measurement of HF antenna radiation patterns. An RTL-SDR V3 serves as the receiver on the drone and is controlled by a Raspberry Pi 4. The PixHawk 6X flight controller navigates the drone to programmed waypoints and triggers data collection by the Raspberry Pi and RTL-SDR. A GPS real-time kinematic (RTK) system is used to increase positioning accuracy to less than 10 cm. The antenna-under-test (AUT) is a cylindrical folded helix, derived from the spherical folded helix antenna. It has three arms with 1.5 turns each and is integrated into a spring-loaded collapsible leaf-basket. Dimensions are 0.64 and 0.54 meters tall and diameter, respectively, for a ka of 0.41 at 28 MHz. 36 5.1-meter radial ground wires are attached to the antenna when deployed. Measurements of the radiation pattern are taken between 45 and 50 meters away and show good agreement with simulations. Horizontal position is accurately achieved and the source of increased error in vertical position is described.
This paper presents a study aimed at developing guidelines for generating accurate Huygens Boxes from low-frequency antenna measurements, particularly in the VHF/UHF range, for antenna placement analysis. In flush-mounted scenarios, it is standard practice to measure the source antenna on a finite ground plane and apply a pre-processing step, known as the Infinite Plane Boundary Condition (IPBC), to emulate the response over an infinite ground plane. For the first time, a simulation-based approach is used to quantify far-field reconstruction errors arising from three key limitations in applying IPBC at low frequencies, namely: the size of the Huygens Box, the dimensions of the ground plane, and the truncation of the scanning area. Among these, scanning area truncation is particularly critical, as IPBC requires radiation pattern data from both the upper and the lower hemispheres to effectively mitigate edge diffraction effects. While a ground plane of 5–7 wavelengths is typically recommended, such dimensions are often impractical at VHF due to physical constraints. This study investigates the impact of using a reduced ground plane down to one wavelength and less. Additionally, the influence of varying Huygens box sizes is examined to determine the necessary margin between the antenna and the box boundary. The overall analysis is conducted using two RF sources: a single blade antenna and a 2-element blade antenna array. The accuracy of the IPBC method is evaluated in both free-space conditions and in a realistic aircraft model scenario.
A commonly employed technique in the field of phaseless antenna measurements is the two-sphere method, wherein complex spherical mode coefficients are reconstructed from amplitude-only near-field measurements taken on two spheres that are separated by a specific distance. Recent numerical studies have indicated that the second sphere can also be substituted with a polyhedral surface. However, traditional positioning systems are incapable of measuring such configurations. This work demonstrates that a robotic antenna measurement system can efficiently measure the near-field on the surface of an octahedron. Two approaches for scanning a spiral grid with a robot are presented. It is shown that the additional degrees of freedom offered by the robot can reduce measurement time by up to 15 % compared to conventional spiral measurements. Furthermore, the far-field pattern derived from the complex near-field data on an octahedron produces highly accurate results, achieving an average equivalent error signal of −56.6 dB when compared to standard spherical measurements. A comparison between phaseless reconstruction using two spheres and a combination of a sphere and an octahedron reveals that both methods yield comparable accuracy.
One of the major challenges of radio frequency (RF) propagation at 60 GHz is the high path loss and significant attenuation due to obstructions. These limitations hinder the implementation of wireless communication when the line-of-sight (LOS) path between the transmitter and receiver is blocked. To address this, intentional use of reflective environments is proposed as a potential solution to establish alternative paths between the transmitter and receiver particularly for Industrial Internet of Things (IIoT) applications. We have developed a test chamber to create and alter reflective scenarios at 60 GHz with full traceability to primary standards. The chamber allows device testing and scenario design for wireless devices at 60 GHz. Measurements were conducted in a hybrid anechoic/reflective chamber using open-ended waveguide antennas for transmission and reception. A synthetic aperture at the receiver was employed to emulate large array configurations, by moving a receiver antenna with a small articulated robot to create up to a 35×35-element grid. Spherical reflectors of various sizes and positions were tested to generate intentional reflections. The channel-induced error vector magnitude (EVM) using 64-QAM modulation was measured for each configuration and for different array sizes. Results show that EVM improves with increasing reflector size. The study also identifies the minimum array sizes needed to achieve acceptable EVM performance. These findings are promising for the development of reliable 60 GHz WLAN systems using reflector-assisted non-line-of-sight (NLOS) communication.
Antenna measurements are subject to many sources of error and uncertainty. One of the most common and significant is impedance mismatch error. This error is a result of a difference in impedance looking into the antenna terminals and the circuitry that directly precedes it. Furthermore, this error is difficult to quantify and compensate for when the antenna is embedded in a dense, 3D integrated package. This paper proposes a theoretical methodology to correct this error if the return loss of the antenna is known. Its usefulness is demonstrated by implementing it as a custom NSI2000 script to process data from recent antenna measurements. The results from two test cases show the expected behavior and prove that even in well matched cases, errors can be identified and compensated for. The simplicity and ease of implementation of this method make it attractive for use with any type of antenna measurement for the purpose of increasing gain measurement accuracy.
Plane Wave Generators (PWGs) utilize arrays of radiating elements to approximate plane wavefronts, thereby creating localized far-field-like conditions within a Quiet Zone (QZ). Their compact form factor makes them especially advantageous at low frequencies, such as in the VHF and UHF bands, where traditional Compact Antenna Test Ranges (CATRs) become impractically large. This paper presents results from a comprehensive validation campaign of a 19-element PWG demonstrator, conducted as part of a broader development program aimed at realizing a full-scale system for VHF/UHF testing. The campaign, executed at Pulsaart by AGC, involved both element-level and array-level assessments using a spherical near-field multi-probe system. Key objectives included validating QZ synthesis, calibrating array excitations via digital twin modeling and field expansion methods, and quantifying realized excitation errors. The findings confirm the robustness of the PWG design, the effectiveness of the calibration process, and the minimal impact of mutual coupling and active impedance variations on performance.
This paper presents a computational imaging method that uses a metasurface-based transmitting antenna to reconstruct high-resolution scattering signatures of arbitrarily shaped radar targets. Instead of relying on mechanical scanning or bulky antenna arrays, the proposed approach takes advantage of programmable phase distributions on a metasurface to generate diverse illumination patterns. These patterns help encode different scattering responses from the target, which are collected by a single receiving antenna. The scattered electric field is modeled as a linear matrix equation. This model includes the effects of each transmitting unit cell, the distribution of equivalent point scatterers on the target surface, and the propagation paths between the transmitter, the target, and the receiver. The result is a forward model that links randomized metasurface phase patterns to the measured backscattered field. To achieve high-resolution image recovery, a large number of unknowns—representing the complex amplitudes of point scatterers—must be estimated. Although multiple phase masks are used to increase measurement diversity, the number of measurements is still much smaller than the number of unknowns, making the system underdetermined. To solve this, we use a compressive sensing technique known as basis pursuit denoising (BPDN), which finds sparse solutions in the complex domain and enables accurate reconstruction despite the limited data. We verify the proposed method using numerical simulations on canonical radar targets. The reconstructed images show high similarity to the original targets, confirmed by quantitative comparisons such as structural similarity indices and error norms. These results demonstrate that the method can effectively extract scattering profiles using a compact, electronically reconfigurable antenna system. This work shows the potential of combining metasurface technology with compressive sensing to build efficient, lightweight radar imaging systems that do not require complex hardware setups.
We compare three techniques to improve sample efficiency when measuring near-field emissions from embedded computing systems on a planar region close to the device-under- test (DUT). The techniques are based on either an expansion as trigonometric polynomials or on assumptions about the mathematical regularity of functions that describe the measurands. These assumptions lead to a non-adaptive compressed sensing (CS) approach, an adaptive sparse grids (SG) approach, and an adaptive approach through a Gaussian process regression (GPR) surrogate model. We compare the techniques on a simulation of small computing devices that are measured by a near-field magnetic field probe on a planar region at different subsets of Nyquist grids at different heights above the device. The simulations show that observation height and observation area are important parameters for deciding an effective subsampling algorithm: when observation heights are lower, GPR and CS perform similarly in terms of sampling efficiency for the same estimation error. SG sampling outperforms Nyquist-style sampling, but requires more samples than CS or GPR.
Time-varying metasurfaces, also known as reconfigurable intelligent surfaces (RISs) or smart surfaces, offer new opportunities and applications, where fast-time modulation of the biasing conditions of individual elements enables beam- steering, frequency shifting, information encoding, and more. The unexplored potential offered by these modulated surfaces is matched by a number of challenges in the characterization of their performance. This work highlights ongoing efforts at GTRI to develop standard techniques for the characterization of these surfaces, with an emphasis on those designed for frequency conversion. First, the complex reflection response must be obtained for the entire range of static DC biasing conditions in order to compute the phase-voltage relationship required to design modulation waveforms. We present a three-standard offset short calibration method to perform static characterization of surface phase versus DC bias conditions in the free-space focused beam system, avoiding phase errors that are introduced by over- constrained time gating of resonant structures with two-standard calibrations. Second, the metasurface must be characterized over frequency under various modulation conditions.
Larger low-observable targets are being mounted onto RCS pylons. In many cases not only Azimuth rotation of the target, but a degree of movement in elevation is desired. This requires in many cases a large number of positioning cables to run from the base of the pylon to the tip where the rotator is placed. At the same time the low-observable qualities of the target call for pylon ogives with higher ratios to minimize the background RCS of the pylon that supports the target. The higher ratios call for very thin structures that cannot handle the weight of the rotator or have not enough space for the control and power cable to be fed to the rotator. A way of solving this problem is to have a variable ratio pylon, where the ogive at the tip is different from the ogive on the main body of the pylon. To analyze these pylons a higher-order basis-function method of moments (HOBFMoM) approach has been used in the past [1]. To conform the quadrilateral flat patches to the round geometry of the pylon, patches smaller than 0.3λ were used. While this was still an advantage over the typical 0.1to 0.05λ patches it placed limits on the highest frequencies that could be analyzed give the available computational resources. In this paper the authors present an approach to the meshing of the structure that allows for computing the monostatic RCS at frequencies in the x-band for a 2.4 m tall pylon. In addition, the effects of the non- physical absorber terminations are further analyzed.