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.)
Prior literature in the subject area of far-field antenna measurements has demonstrated an extrapolation technique to isolate and correct the errors associated with nearzone proximity effects, specifically multiple reflections between the probe and the antenna under test (AUT), thus allowing measurements to be acquired at separation distances much shorter than the conventionally defined far-field criteria. A recent paper on this topic described a modern, indoor, far-field antenna measurement range specifically designed to support the traditional extrapolation technique while also incorporating high-speed RF instrumentation and advanced software control of a mobile probe tower. The automation of the traditional technique was emphasized, and the application focused primarily on X-band performance. Herein presented is an updated and more broadband approach which utilizes both amplitude and phase data to extend the implementation to frequencies in the UHF-, L-, and S-band. Optimized correction factors are generated for additional extraneous signals, most notably the effects of multi-path interference. Using the generalized three antenna measurement approach as highlighted in the original technique, measurement examples are provided for broadband antenna range horns, and the resultant far-field gain calculations are again compared to similar data extracted using traditional near-field scanning techniques.
Reconfigurable antennas are very widely useful antennas, but they require extended measurement periods to characterize the range of specified beams. Time-saving measures typically come at the cost of measurement quality. The goal of this effort was twofold: 1) to investigate ways to improve all antenna measurements, including analyzing antenna positions within range spaces, absorber configurations, and mounting structures and 2) to investigate the procedure by which reconfigurable antennas are optimized and determine efficient measurement quality and time-cost tradeoffs.
Reconfigurable antennas are very widely useful antennas, but they require extended measurement periods to characterize the range of specified beams. Time-saving measures typically come at the cost of measurement quality. The goal of this effort was twofold: 1) to investigate ways to improve all antenna measurements, including analyzing antenna positions within range spaces, absorber configurations, and mounting structures and 2) to investigate the procedure by which reconfigurable antennas are optimized and determine efficient measurement quality and time-cost tradeoffs.
To characterize the radiation characteristics of an antenna, determining the power pattern of the antenna is often sufficient. In some cases, however, both the amplitude and phase response are important. For instance, for accurate channel modeling, the antenna has to be de-embedded, requiring knowledge of the complex radiation pattern of the antenna. A vector network analyzer typically measures complex S-parameters, hence, determining the complex radiation pattern seems like a straightforward task. When measuring at higher frequencies, as the wavelength becomes shorter, antenna phase measurements are very sensitive to positioning and alignment errors. Using sophisticated measurement tools, the position and orientation of the antennas can be determined, and this information can be used to correct the measurement data. The stringent requirements on positioning and alignment at millimeter-wave frequencies, however, makes correcting the data based on physical insight, in some cases, a more practical solution. The results of a radiation pattern measurement of a WR-28 rectangular open-ended waveguide will be shown in the full paper. The magnitude of the radiation pattern is symmetric in its two principal planes, which is to be expected, but the phase of the radiation pattern is not symmetric. To explain this lack of symmetry, a two-parameter misalignment model will be presented. It will be shown that the measured phase is much more sensitive to the misalignment than the measured magnitude, explaining why the symmetry is only lacking in the measured phase. Based on the 1,708 available planar cuts, the two parameters in the misalignment model are determined with great confidence. Subsequently, the parameters are used to correct the phase of the measured radiation pattern, restoring the expected symmetry in the phase measurement.
Placing a Direction Finding (DF) array onto an existing aircraft is typically a difficult endeavor due to the limitations placed by existing antennas or structures which mandate keep-out areas, the additional infrastructure required for potentially dispersed DF antennas, and getting all of the modifications for the DF antennas flight certified. Because of these challenges, along with the basic expense of modifying an aircraft for external antennas, the ability to optimize the antenna lay down for peak DF performance is absolutely essential. This paper will describe the use of a Genetic Algorithm (GA) in the application of defining antenna locations on a platform to form an optimized broadband Direction Finding (DF) array. For this optimization study, the Correlation Interferometry Direction Finding (CIDF) algorithm, will be used to assess candidate array solutions generated by the genetic algorithm. CIDF Correlation domain statistics such as main correlation beamwidth (proportional to DF accuracy), and correlation sidelobe levels (average relates to array robustness, maximum relates to potential for wild bearings) were used to assess each candidate array over the entire frequency band of interest. This paper will show how that the use of a genetic algorithm, with an optimization function based on CIDF correlation statistics, and a fitness function adjusted in population size and mutation rate, will yield the derivation of a robust DF antenna array configuration. This paper will derive the critical optimization and fitness functions, and then use examples of a large jet aircraft, a medium size business jet, and a small Unmanned Aerial Vehicle (UAV) to demonstrate the genetic algorithm capability to solve the DF antenna placement problem. As the reader may not be familiar with the theory of interferometric direction finding, or genetic algorithms, a brief tutorial will be provided in Sections I and III respectively.
This paper presents a novel method using multiple compact antenna test range (CATR) reflectors to simulate the Radio Resource Management (RRM) measurements required for 5G devices capable of beam-forming in the millimeter wave frequency range (i.e. FR2). Four CATR reflectors are arranged on a semi-circle with the device under test (DUT) on a dual axis positioner in the center of the intersection of four planar waves in order to generate five sets of two Angles of Arrival (AoA), thereby capable of simulating multiple basestations from different directions for the 5G device to monitor and perform handovers. The reflectors create far-field conditions at the device under test (DUT) such that quiet zones of up to 20-30cm in size can be achieved. Absorber baffles are strategically placed as to reduce scattering from adjacent reflectors. In addition to RRM measurements, one reflector can be used to also perform in-band RF beam characterization[JMFL2] while additional reflectors can measure out of band emissions at the same time, thereby decreasing total measurement times by a factor of 2-3 times.
Antenna placement or antenna in-situ performance analysis on large and complex platforms such as ships, airplanes, satellites, space shuttles, or cars has become even more and more important over the years. We present a systematic investigation of different antenna types for space applications in G- and S-band on an experimental aircraft. In this process, the individual antennas are measured with the help of a dual reflector compact antenna test range (CATR) under far-field conditions in various configurations. These results are validated and compared utilizing a finite element method (FEM) based solver simulation model. At first, the antennas are simulated and measured alone without any supporting or mounting structure. Subsequently, the effect of mounting structures on the overall radiation performance is added by analyzing the antennas over a large conducting ground plane, on top and the side of winglets, and on top of a cylinder body with dimensions on the order of the actual aircraft. For the detailed in-situ investigations, a second method of moments (MoM) based simulation tool is employed which works on measured sources. These measured sources are obtained from the CATR measurements of the isolated antennas. By means of a spherical wave expansion (SWE), they are transformed into a near-field source for the simulation model. These measured data based results are again compared and validated with the full FEM simulation for the complete aircraft setup and the simplified cylinder body. By this means, the expensive design and measurement of a full-scale electromagnetically equivalent mock-up of the aircraft could be saved. Furthermore, the pure simulation of the installed antenna performance often suffers from the limited availability of exact antenna design parameters. In some cases, the antenna design data remains undisclosed deliberately due to IP reasons. The presented results reveal the influence of physical structure on the radiation characteristics and demonstrate the benefits of working with measured data in simulation tools.
In modern systems where RF front ends are tightly integrated, the parameters of passive aperture gain and active electronics noise figure become difficult to obtain, and, in many cases, impossible to measure directly. Instead, the parameter referred to as Gain over Noise Temperature, or G/T, becomes the performance metric of interest. Recently, the antenna measurement community has seen an increased demand to use near-field measurement systems for determining G/T values. Papers presented at AMTA over the past few years have shown that it is possible to determine G/T values using measurements taken in planar near-field antenna ranges. The CW-Ambient technique was one of the techniques proposed for computing G/T values by utilizing planar near-field measurements [1,2]. In this paper, we show how the CW-Ambient technique can also be applied to calculate G/T values in spherical near-field antenna measurement systems. This paper provides a brief summary of the CW-Ambient technique, and then presents the procedure and equations required for computing G/T using a spherical near-field system. To validate the recommended procedure, we compare predicted and measured G/T values for a separable unit under test (UUT). Since the passive aperture for this UUT is separable from the back-end active electronics, we measure the aperture gain of the UUT and the noise figure of the back-end electronics individually, and then compute the composite G/T value for this assembly. We then compare these composite values against G/T measurements from a spherical near-field antenna measurement system. We summarize these comparisons and provide conclusions regarding the validity of using a spherical near-field system to measure G/T.
This paper investigates on the use of under-sampling over the azimuthal dimension to reduce measurement time on spherical near-field scanning. This means that the number of angular phi samples is reduced, which allows to reduce the number of positioner steps, obtaining measurement time savings virtually proportional to the number of samples reduction. Of course this under-sampling introduces an error, which can be interpreted as an aliasing term over the retrieved Spherical Wave Expansion of the Antenna Under Test (AUT). The axial symmetry of the vast majority of antennas allows the application of significant under-sampling ratios with little aliasing errors. However, this information is not a priori known due to the lack of a reference AUT radiation pattern, or in the case of malfunctioning antennas with degraded symmetry. Here we proposed a measurement procedure for the exploitation of the AUT axial symmetry. The procedure consists on an iterative AUT measurement with increasing number of azimuthal cuts. As the number of cuts increases, the aliasing error decreases, thus obtaining the final radiation pattern with a lower uncertainty. We will introduce an aliasing error estimator, which estimates the error caused by the under-sampling without any a priori knowledge of the AUT. This estimator can be used as a stopping criterion of the iterative measurement procedure when the desired accuracy is achieved. The proposed technique will be demonstrated using different antennas, showing considerable reductions in measurement time with low errors in the transformed far-field pattern, and with the guarantee that the error is below a given threshold thanks to the derived estimator.
Direct far-field (DFF) testing has become the standard test methodology for sub-6 GHz over the air (OTA) testing of the physical layer of radio access networks with the far-field multi-probe anechoic chamber (FF-MPAC) being widely utilized for the test and verification of massive multiple input multiple output (Massive MIMO) antennas when operating in the presence of several users. The utilization of mm-wave bands within the 5th generation new radio (5G NR) specification has necessitated that since the user equipment should, preferably, be placed in the far-field of the base transceiver station (BTS) antenna, excessively large FF-MPAC test ranges are required or, the user equipment is paced at range-lengths shorter than that suggested by the classical Rayleigh criteria or, a modified compact antenna test range geometry must be developed and utilized. This paper presents a novel design for a new compact antenna test range (CATR) design that uses a parabolic toroid as the main reflector. The folded optics utilized within this design possesses superior pseudo-plane wave scanning capabilities than those available from equivalent, classical, point source offset parabolic reflector CATR designs. This wide-angle scanning capability is a crucial feature for successful over-the-air testing and measurement of mm-wave 5G NR Massive Multiple Input Multiple Output (MIMO) antenna systems within multi-user applications providing 60 degrees of azimuth scan and 15 degrees of elevation scan to the incoming plane wave at the AUT. CATR quiet-zone results are presented, compared and contrasted with more classical designs before results of the effects of the CATR test channel on a number of commonly encountered communication system figures of merit in wide scanning cases are presented.
Spherical wavefunction expansions (SWEs) form the basis for spherical near field measurements and are also very useful for the analytical assessment of antenna performance, e.g. achievable quality factor and directivity/gain. On the other hand multipole source expansions also have great utility. The cartesian multipole expansion in particular provides much physical insight to the behavior of compact sources of electromagnetic radiation. The relationship between cartesian multipoles and spherical wave function expansions is not one-to-one and is quite complicated as is evidenced by the equivalent multipole source configurations for low-order (up to radial index of 3) spherical wavefunctions previously given by W. V. T. Rusch. Here we provide a formalized approach to the determination of the spherical wave function expansion for a particular cartesian multipole source and also the cartesian multipole which would excite a particular spherical wavefunction. The impetus here is the modeling of couplers for wireless power transfer. One example of this is the so-called DD coupler for automotive wireless charging described in the SAE J2954 standard which can be modeled as a TE-to-z cartesian multipole consisting of four displaced magnetic dipoles. In terms of Harrington's cartesian multipole nomenclature, this source has an order of n=2 and m=1 for the TE-to-z electric vector potential. It excites two TE-to-R spherical wavefunctions, one with radial index of 1 and azimuthal index of 1 and the other with radial index of 3 and azimuthal index of 1. More complex multipole sources are required to describe more elaborate WPT couplers such as some associated with the Qi wireless charging standard. WPT systems necessarily operate with switched-mode sources and hard-switched rectifiers and therefore produce spectral components of current over a broad frequency range. Thus the behavior of the extraneous electromagnetic field at frequencies well above the fundamental is of great interest. The cartesian multipole representation of the WPT couplers facilitates the understanding of this extraneous electromagnetic field. The knowledge provided by such a multipole model would be useful in determining that measurements over a full sphere or hemisphere are absolutely necessary; that is, that typical emissions measurements are inadequate.
An automotive radar is known as providing the vehicle drivers with useful information such as the positions of the surrounding vehicles and the obstacles on the front, side, and rear of the driver. This information can beapplied to control the vehicle in an autonomous mode and helpthe vehicle to drive safely at night or in bad weather conditions. For example, the adaptive cruise control (ACC) radar system in a medium range operates to transmit an electromagnetic wave signal and receive the wave signal reflected from neighboring vehicles or obstacles and estimates the distances between the radar and the neighboring vehicles and their relative speeds by using the time difference between the two signals and the Doppler frequency variation. However, there are so many traffic situations on a road, and it is almost impossible to experiment the ACC radar system in all the real traffic scenario. Therefore, it is required to implement a simulation setup to cover all development process from designing a radar antenna to collecting radarsignal to identify any traffic situations. In this work, automatic ACC radar antennas are designed and applied to simulate a real traffic scenario. The radar antennas are designed to satisfy the ACC radar specification and loaded on a vehicle to transmit and receive RF signals in simulation environment. The radar antennas are designed with Altair Feko and the relative distances and speeds of the vehicles surrounding the radar built-in vehicle in a virtual traffic situation are obtained with Altair WinProp. In addition, a trihedral corner reflector (TCR) is also included in this study to calibrate the radar antenna system by comparing the reflected characteristics from a TCR target with those from a vehicle. Finally, the ACC antenna specification (beam-width) is evaluated in terms of how useful arethe data collected by a radar vehicle which overtakes a front vehicle.These datashould also provide a good basis to validate the results of experimental data for automotive radar system implementation.
With trending topics such as car-to-car communication or automotive radar, the evaluation of antenna performance, including the impact of the larger structure in which it is integrated, is a topic of rapidly growing interest. Implementing large anechoic chambers, supporting e.g. full-vehicle testing, is definitely an approach, which can serve as a reference. Nevertheless, the cost and complexity associated to such antenna test ranges are quite prohibitive. Some measurements eventually get close to impossible, as they involve unreasonable testing times and efforts, due, for instance, to the size to wavelength ratio of the DUT. This paper introduces an economic alternative, based on the convergence of measurements and full-wave simulation. The described method relies on a three-step approach: (i) measure the phasor electric field radiated by a smaller-size antenna module over a surface enclosing the test sample; (ii) use an algorithm to calculate equivalent electric and magnetic currents over a surface closely surrounding the DUT; (iii) inject these currents, as a Huygens source, into a full-wave solver, where the entire supporting structure is then taken into account. The relevance of the technique is demonstrated with a complex example, where a multilayer 8x8 antenna array frontend module operating at Ka-band is both evaluated numerically and experimentally. Two sets of simulations are created, where either the detailed numerical antenna model or the equivalent current surface is integrated under a radome at the nose of an aircraft. Key radiation parameters (peak gain, side lobe levels, etc.) are compared, showing the relevance of the approach. Its limitations and uncertainties are also discussed.
The characterization of 3D radiation pattern of antennas using spherical measurements techniques is nowadays a common testing procedure. However, the well-established technique using Spherical Harmonics (SH) can be time consuming since the required number of field samples is proportional to the square of the electrical length of the antenna. The main solution is to reduce as much as possible the number of field samples to reconstruct the antenna pattern. This can be achieved by either exploiting prior knowledge about the antenna or leveraging general properties common to the fields radiated by antennas field. Our approach uses the sparsity of the spherical harmonic expansion of the field radiated by antennas and only requires the maximum electrical dimension of the antenna and its frequency to set the truncation order of the harmonic series. It nonetheless requires the proper tuning of specific parameters to ensure that the antenna radiation patterns, reconstructed from a small number of sampling points, is accurate enough. First, the minimum number of field samples to ensure a proper interpolation must be set, the reconstruction accuracy (for undersampled data) is linked to the number of significant modes involved in the expansion: roughly speaking, the sparser, the better. Second, the data fidelity term and more specifically the error tolerance, common to all regularization scheme such as the sparse one at hand, must be carefully chosen. Finally, an efficient post-processing procedure, based on the rotation of the antenna, is proposed for improving the sparsity of its SH spectrum, and thus enables extracting as much information as possible from a given fast measurement data set. All these points are validated numerically and experimentally on spherical near and far-field data.
A method to solve the far-field distance problem is the usage of reflectors in order to transform the spherical wave from the feed antenna into a plane wave which ultimately leads to the same far-field condition but in a more compact way. Therefore, these facilities are called compact antenna test ranges (CATRs). However, the finite size of the reflector(s), despite edge treatment, is the main cause of erroneous signals impinging into the test zone in which an antenna under test (AUT) is characterized. Especially at lower frequencies, every further structure inside the test chamber, although covered with absorbers, is an additional source of scattered signals. One method of correcting stray signals and to improve the AUT measurement accuracy is to compensate the non-ideal test-zone field (TZF) via spherical wave expansion (SWE). For this technique, a complete description of the TZF, i.e. measurements on a closed surface around the test zone, is required. The most convenient approach is to use a spherical scanning surface. In ranges with a roll-over-azimuth positioning system, the spherical scanning can be realized by an additional measurement arm. Using the chamber positioning system, thus, strongly reduces the required additional hardware and makes spherical scanning of the test zone a practical approach. With the knowledge of the incident unwanted field components, AUT measurements carried out in the corresponding test zone are correctable. Spherical near-field measurements of the test zone created by the CATR installed at the Institute of High Frequency Technology, RWTH Aachen University are performed at a frequency of 2.4 GHz using a simple scanning arm. Reliability of the results is ensured by comparing the measurements to full-wave simulations of the CATR. The radiation pattern of a base transceiver station antenna serves as a test case and is subsequently corrected for the erroneous signals using the SWE.
With the development of the Internet of Things (IoT) technology, the antenna becomes increasingly integrated and miniaturized. Vector Network Analyzer (VNA) is a standard instrument for characterizing antennas. However, IoT devices are small and scattered installed, which makes it costly to carry out in-field characterizing on IoT devices. Integrating an antenna characterizing system into IoT devices can release the difficulty significantly, but characterizing antennas usually requires a high-performance Analog to Digital Conversion system, which is expensive and power-consuming. However, the antennas are not required to be tested frequently, which means this is not a real-time application. Therefore, this paper proposes a method that can realize time-domain reflectometry based VNA with just a comparator or differential receiver. In this system, impulses are sent into the antenna, and the frequency response is captured by analyzing the time-domain reflection. Analog to Probability Conversion (APC), Probability Density Modulation (PDM), and Equivalent time sampling (ETS) concepts are used to reduce the real-time performance requirements of the ADC system, so that the cost and power consumption can be tolerable on a tiny IoT device. With this technology, the antennas could be characterized in-field and remotely, making the maintenance easier. This technology is implemented with Xilinx ZYNQ Ultrascale+ series FPGA. Two antennas are tested, and the experimental results show that the system can successfully measure the radiofrequency. The sampling rate is set to 89.6Ghz, and a micro-volt level voltage resolution is achieved. Since the essence of technology is a trade-off between time and performance, the sampling rate and resolution can be further increased theoretically according to specific applications.
The interest for mm- and sub-mm-wave systems has grown in the last years, mostly driven by communications and radar industries. In particular, the new generations of communications systems, Beyond 5G, put the focus in the use of these higher frequency bands, exploiting the large bandwidth availability to enable the transmission of high data rates. In this context, not only new wideband high-gain antenna concepts are needed, but also advances in the applied antenna measurement procedures. In this work, a D-band lens antenna design with gain larger than 30 dB is presented, which can be applied in point-to-multipoint communication scenarios. The elliptical lens, fabricated in low-cost plastic material, is fed by an aluminium split-block, with an interface to a standard WR-6 waveguide. The antenna covers 42% bandwidth with an aperture efficiency higher than 80%, and radiation efficiency higher than 86%. Wide-band radiation patterns are achieved thanks to a leaky-wave air cavity placed between the waveguide feeder and the lens, which enhances the feeder directivity inside the lens, bringing high illumination efficiency. The lens radiation pattern can be steered by displacing the feeder along the focal plane. The lens radiation patterns and gain have been characterized by means of a 50 cm range length spherical distributed-axis scanner, built in a compact anechoic chamber (absorber tip-to-tip dimension: 0.64 m x 1.25 m x 0.93 m). The elevation arm of the scanner is equipped with a new active dual-polarized probe, with integrated down-converters and diplexers. The introduced technology allows measurements in the 110 to 170 GHz range with a dynamic range better than 65 dB (1 kHz RBW, 0 dBi DUT gain assumptions). The agreement with full-wave simulations is excellent over the whole frequency band for the broadside beam and for a beam steering at 9?. The antenna S-parameters have been as well measured, validating full-wave simulation results. Overall, the design and modelling, manufacturing and measurement accuracy meet the challenging requirements in these high frequency bands.
Internet of things (IoT) has impacted global supply chains in terms of improving performance and increasing productivity. Modern IoT devices such as Radio Frequency Identification (RFID) tags play a pivotal role in packaging surveillance and monitoring of products as it moves along the supply chain. Additionally, recently these tags are equipped with sensing elements that convert the traceability-centric supply chain to value-centric by enhancing visibility of the nature of product as it moves along the supply chain. To utilize the full potential of such IoT technologies, an intelligent network planning is a pre-requisite that ensures good and reliable communication between the RFID tags, the reader, and the cloud. Traditionally, optimal positioning of these RFID devices to obtain a complete coverage is a difficult task that requires careful planning and physical experimentation which involves investing significant time and financial resources. To avoid such limitations, computer aided simulation of positioning the devices allows engineers to explore different scenarios to ensure complete coverage within a given area within a short time frame. The design and coverage planning of various entities of the RFID infrastructure is pivotal in realizing value-centric approach towards supply chain management. In this paper, two different case studies are presented that utilizes a grid-based optimization approach for coverage planning of passive and active RFID tags. For this purpose, first, a high-frequency solver FEKO is used for designing the RFID reader and tag antenna. Second, for the coverage analysis, a wave propagation tool WinProp is used for optimizing the position of the RFID readers and tags. The antennas are designed to operate at a UHF RFID frequency (915 MHz) and a semi-industrial warehouse setting is used for network planning. The details of the design and simulation of the individual entities of the RFID infrastructure along with two different scenarios for coverage analysis of active (battery-powered) RFID tags and passive (battery-less) RFID tags are presented in the paper. These simulations are the first step towards realizing the value-centric supply chain management approach.
In numerical simulation of antenna problems, accuracy of antenna representations is essential to ensure the reliability of results. Integration of measured Near Field (NF) representation of antenna in Computational Electromagnetic (CEM) solvers opens new perspectives to solve this problem. Moreover, it is possible to replace the simulated model of the antenna by a measured model, which represents the real antenna. No additional information about mechanical and/or electrical design of the antenna is required by the numerical solvers. Indeed, the measured NF model in terms of equivalent currents already provides a complete and detailed representation of the antenna itself. The applicability of this approach has been already studied for complex and/or large scenarios, antenna placement, scattering problems and EMC applications. Another interesting use of the combination between measurement and simulation is to enhance the evaluation of the antenna coupling. Previous investigations have been carried out on an H/V polarized array of three identical cavity-backed cross-dipole antennas. In this study only the radiation pattern of the central element of the array was measured (in a stand-alone configuration). Its representation in terms of equivalent currents was integrated in the simulation, for the calculation of the coupling with other elements. For each element two feeding ports, H/V polarization, have been investigated. In particular, measured patterns at five frequency points were used to determine the antenna coupling over the whole frequency band by simulation. A good agreement was found between the measured mutual S parameters on the real array and results obtained by the combination between measurement and simulations. This investigation demonstrated the validity of this approach. In this paper a continuation of the previous study will be performed, exploring the following topics: Enhancement of the representation of the NF source by inclusion of placement boundary condition. Use of measured NF source models to represent another element of the array, not only the central one. The calculation of the antenna coupling will be determined for these new configurations.
The Benefield Anechoic Facility (BAF) at Edwards Air Force Base is the world's largest known anechoic chamber. Due to its unmatched size and complement of test equipment, the BAF hosts far-field pattern measurements at all azimuth angles and multiple simultaneous elevations of installed antennas on large aircraft across a frequency range of 0.1 - 18 GHz. Antenna tests at the BAF rapidly produce large quantities of data, which often require immediate analysis to allow system owners to make relevant improvements. Historically, the BAF had accomplished quality assurance manually. Analysis was accomplished post-test by customers and the BAF team. Today, the BAF team has developed scripts that use kernel density estimation and basic machine learning to automatically check incoming data for errors and highlight unusual results for review. During a 2019 test of over sixty installed antennas on a B-1B bomber, the BAF team used these scripts to produce calibrated, quality-assured antenna patterns in near real-time. Rapid processing brings deficiencies to the customer's attention fast enough to allow corrections to be applied and re-tested during the same test event ? highly significant and valuable as aircraft and BAF schedule times are limited and may be a one-time opportunity to gather required data. This paper will explore the algorithm used to evaluate antenna patterns, as well as the expected characteristics of patterns that enable the selection of relevant data. Development and application of this algorithm found that using kernel density estimation to calculate the number of maxima in a pattern's distribution of gain values, then performing this recursively over only the main lobe, can identify problems such as incorrect switching, mismatched transmission lines, and multipath. Algorithm optimization was achieved using generated data, then verified by applying the algorithm to previous test data. For the B-1B, the script searched for data that deviated from an expected pattern with clean main and side lobes, minimal frequency dependency, and a low-power noise distribution at all azimuth angles outside the lobes. Finally, this paper will discuss the results of using this algorithm during a live test, and future improvements and applications for this data processing technique.