{"@attributes":{"version":"2.0"},"channel":{"title":"Resources","link":"https:\/\/www.remcom.com\/resources","description":"Discover Cutting Edge Insights On Electromagnetic Simulation & Wireless Engineering By Stefanie Lucas!","language":"en","pubDate":"Wed, 12 Aug 2026 18:15:36 GMT","item":[{"title":"RF Wireless System Digital Twins and Automation for AI | Remcom","link":"https:\/\/www.remcom.com\/resources\/video\/rf-wireless-system-digital-twins-and-automation-for-ai","description":"
\n \"RF <\/a> \n<\/div> \n

RF system validation relies heavily on prototyping, chamber, and over-the-air measurements, yet physical testing cannot feasibly capture the vast majority of practical edge cases such as complex device orientation, dynamic blockage, and multipath environments. <\/span><\/p>","pubDate":"Wed, 01 Jul 2026 17:47:20 GMT","author":"kaitlyn.brickley@remcom.com (Kaitlyn Brickley)","guid":"https:\/\/www.remcom.com\/resources\/video\/rf-wireless-system-digital-twins-and-automation-for-ai"},{"title":"CUDA Implementation of Moving Window Finite-Difference Time-Domain | Remcom","link":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/cuda-implementation-of-moving-window-finite-difference-time-domain","description":"

\n \"CUDA <\/a> \n<\/div> \n

With the need for real time propagation growing each day, a new way to simulate propagation models is needed that is closer to real time while maintaining the same amount of accuracy. The way to get the most accurate results is to use a high fidelity model such as a Finite Difference Time Domain (FDTD)<\/a> model. This model is limited in space and computational resources. With the way the algorithm is calculated, the computational space can become massive very quickly. Therefore modeling long distances can be impossible. One way around that is to employ a method called Moving Window Finite Difference Time Domain (MWFDTD). This method only takes into account the area around the pulse. However, this is not close to real time due to computational intensity. One way to make this faster would be to make the calculations faster by using a Graphics Processor Unit (GPU). The GPU can be used to speed up the calculations like those found within MWFDTD. <\/span> <\/p> \n

 <\/p> \n

Introduction<\/h2> \n

This paper discusses the challenges and techniques involved in converting the MWFDTD algorithm from a C++ implementation to an appropriate form for leveraging graphics processor units (GPUs) through NVIDIA\u2019s CUDA framework. The GPU approach employs thousands of threads simultaneously which requires special design considerations in order to achieve maximum speedups. Previous work  [1,2,3]  has  concentrated on the challenges of  implementing GPU-targeted software through the use of the OpenGL API or the Cg language. With proper understanding of CUDA, it is possible to reach speedups of beyond two orders of magnitude over traditional CPUs and increase performance by a factor of  two or more over previous GPU implementations.<\/p> \n

Moving Window Finite Difference Time Domain<\/h2> \n

In order to model long range propagation using traditional 2D FDTD, a vertical plane containing the entire irregular terrain profile is projected onto a rectangular grid consisting of evenly spaced grid points in the x-y plane. In addition, time is divided into evenly spaced intervals. To begin the simulation, an electromagnetic pulse is excited at the transmitting antenna and at each time step, the electromagnetic fields at each grid point is determined by solving Maxwell\u2019s equations using the second-order finite differencing method of Yee[4]. The MWFDTD propagation model is based on a modified FDTD to model radio wave propagation[5],[6].  <\/p> \n

One fact of a traditional FDTD method is the propagation radio pulse only occupies a small part of the computational space.  MWFDTD takes advantage of this by limiting the computational space to the area surrounding the pulse. The window is only as wide and high as the pulse will be. All the other space in a traditional FDTD calculation could be considered as waste. This allows longer runs because the memory and time limitations are not as great. As the pulse propagates along the terrain, the FDTD mesh is moved forward to track the pulse as depicted in Figure  1. The window moves at the speed of light to follow the pulse that is also moving at the speed of light.<\/p> \n

The limited computational space does not compromise the accuracy of MWFDTD. MWFDTD is just as accurate as a regular FDTD calculation.  \"Error! Reference source not found.\"<\/span> is the path loss calculated at a frequency of 410MHz, with a FDTD computational domain of 1000 cells wide by 4600 cells high, a cell spacing of 7.31 cm by 7.31 cm, which corresponds to 10 cells per wavelength at 410 MHz. The time step is chosen to be 0.181 ns. The results were compared with data obtained by ITS. The difference between MWFDTD and the measurements is at most 2-3 dB.  <\/p> \n

CUDA GPU<\/h2> \n

Once used for the sole purpose of driving graphical displays, the GPU has evolved into a powerful computational device. The Tesla C1060 offers a peak performance of  933 GFLOPS and a memory bandwidth of 102 GB\/s. This can yield significant performance gains over even a 2.66 GHz Intel Core 2 Quad processor which has a theoretical peak performance of 42.56 GFLOPS (using all four cores and full SSE2 optimizations) and a memory bandwidth of 10.7 GB\/s. The recently released Fermi C2050 and C2070 GPUs extend these differences even further with a peak performance of 1.03 Tflops and a memory bandwidth of 144 GB\/s. <\/p> \n

CUDA allows software developers to leverage this power without the need for special computer graphics knowledge [7][8].  The GPU is organized into a series of multiprocessors. Each of these multiprocessors contains a set of stream processors and a shared memory cache to facilitate thread cooperation. <\/p> \n

CUDA compliant GPUs follow a single instruction multiple thread (SIMT) architecture [9]. In a SIMT model, threads are launched simultaneously in groups termed warps. Each thread in the warp can execute concurrently as long as they are performing the same instruction on different pieces of data. If threads of a warp diverge through a conditional branch, thread execution must be serialized. The greatest speed gains are achieved by designing software that minimizes thread divergence within a warp. Algorithms such as FDTD perform small amounts of work on large amounts of data, so memory bandwidth tends to be a critical factor in application performance. The GPU offers another significant advantage in this area. GPU threads can switch contexts nearly instantaneously - without the need to store and restore thread state. This fact allows the GPU to hide memory latency by launching thousands of threads simultaneously and performing rapid context swapping while waiting for input data. <\/p> \n

Instantaneous context switching is important for hiding memory latency. CUDA devices offer several types of memory. The three most important types for these purposes are shared, constant, and device (global) memory. Shared memory is cached on chip memory with low latency that can be accessed cooperatively by a group of threads.  It is particularly useful for implementing user defined caches. Constant memory is a cached read-only section of memory currently limited to 64 KB. Device or global memory is relatively high latency but is currently available in amounts up to 6 GB per device. <\/p> \n

Implementation<\/h2> \n

Functions targeted for the GPU are implemented as kernels in CUDA.  Kernels are very similar to C functions except there are certain extensions needed.  This creates a very easy transition for a new user to implement any function using a kernel.  The MWFDTD GPU library was written using a number of kernels to utilize the GPU.  <\/p> \n

The GPU kernels were used in the main update equations, boundary conditions and shifting the arrays as the window moves.  These are the core FDTD functions within the MWFDTD function.  Each of these functions required special consideration with regards to the MWFDTD implementation.<\/p> \n

The first step was to convert the update equations from C to CUDA.  This was done by converting the update functions to update kernel functions.  The main task here was to eliminate the loop that was used within the equations.  The kernel functions will loop over the data available without having the user implement a loop.  The rest of the functions were then converted using the same idea. <\/p> \n

The next step was to define all of the constant data like material identifiers, material constants, and all other constant arrays.   This data will be needed at each time step and each time the window will move.  There are two methods that CUDA capable GPU can read or write from\/to global memory; coalesced and uncoalesced.  Coalesced reads or writes allowed the GPU to perform memory transactions of 32, 64, or 128 bytes simultaneously.  This is the most optimized what to read and write to a GPU due to the fact uncoalesced reads have to be serialized which could require between 400 and 600 clock cycles.  The MWFDTD model utilized this fact by allocating the update coefficients in constant memory instead of global memory.  This is due to the fact these needed to be read at every time step.<\/p> \n

The last step was to implement the functions that will move the window. Each time the window moves new data has to be read into the problem, old data deleted, and all arrays have to be shifted. This was done utilizing various aspects of CUDA\u2019s capability of shifting arrays within memory. The key was to minimize the number of reads from main memory.  The reads were done while the results from the previous window were being written back to main memory. Also the number of reads were minimized by only reading in the new column of data and shifting the data that was already on the card.<\/p> \n

To document the speedsup within the new MWFDTD model, simple test cases were developed to utilize the new GPU implementation of the model.  Each test case was run first in the released MWFDTD 2.5 version and then rerun using the same setup file in the new GPU Implementation in 2.6 version.  Results were then compared.  <\/p> \n


As with other FDTD methods, terrain type impacts scenario set up thus affecting runtimes. The terrain type impacts the number of cells per wavelength. The more cells present within the window, the longer the runtimes. Dry sand terrain scenarios run with 20 cells per wavelength while sea water terrain scenarios run with 100 cells per wavelength. 10 cells per wavelength in the material is needed. To find this value, Equation 1 is used. Table 2 shows a summary of the cells per wavelength per test.<\/p> \n

 <\/p> \n


Each test case consists of a terrain defined by the terrain type at the set range. The waveform used a sinusoid with a carrier frequency defined in the Table 1. The antenna was an isotropic antenna with the defined waveform. There was one transmitter for each project. The number of receivers for each project is defined in the table.  <\/p> \n

The computer configuration for the tests is listed in Table 3.<\/p> \n


Timing for a complete simulation was the benchmark used to evaluate the performance of the new MWFDTD model. Each of the tests described in Section 5 where first run using the CPU implementation then run on the same computer using the GPU implementation.  <\/p> \n


The speedup comparisons are depicted in Figure 3.<\/span><\/p> \n

Test 3 has the largest speedup due to the material that was used. This material required a denser grid. This denser grid creates the need for more calculations to be performed each time step. The GPU runtimes for this particular case are closer to the other cases. This is due to the fact the GPU can process the calculations simultaneously thus processing more data in a shorter amount of time. This shows the true power of the GPU.<\/span><\/p> \n

The speedups achieved are between 38X \u2013 62X. The results show the more complex problems benefit the most from the new GPU implementation. The Wet Earth test cases shows the highest speedup due to the materials used, the range of the problem, and the frequency of the antenna. The sea water case has a lot of cells in each window which results in a long runtime. The majority of the GPU runtimes are very similar as opposed to the CPU runtimes which are all over the board. This is due to the fact the GPU can simultaneously calculate hundreds of grid points while the CPU is limited on the amount of calculations it can perform. This allows most if not all of the cells in one vertical column to be calculated at once instead of one or maybe two at a time like on the CPU thus resulting in the same runtimes for each column.  Due to the ability to only calculate one column at a time, the GPU has enough memory to calculate the denser grids like in the wet earth test case in similar time as the less dense grid test case.<\/span><\/p> \n

MWFDTD was an ideal program to port over to the GPU. The calculations are able to be performed in parallel to completely utilize the power of the GPU. <\/span><\/p> \n

References<\/p> \n

    \n

  1. M. J. Inman, A. Z. Elsherbeni, and C. E. Smith \u201cGPU Programming for FDTD Calculations,\u201d The Applied Computational Electromagnetics Society (ACES) Conference, Honolulu, Hawaii, 2005. <\/span><\/p> <\/li> \n

  2. M. J. Inman and A. Z. Elsherbeni, \u201c3D FDTD Acceleration Using Graphical Processing Units,\u201d The Applied Computational Electromagnetics Society (ACES) Conference, Miami, Florida, 2006. <\/span><\/p> <\/li> \n

  3. Adams,  Samuel,  Jason  Payne,  and  Rajendra  Boppana,  \u201cFinite  Difference  Time  Domain  (FDTD) Simulations Using Graphics Processors\u201d, HPCMP Users Group Conference, 2007 <\/span><\/p> <\/li> \n

  4. K. S. Yee, \u201cNumerical solution of initial boundary value problems involving Maxwell\u2019s equations in isotropic media,\u201d IEEE Trans. Antennas Propagat., vol. 14, pp. 302-307, 1966.<\/span><\/p> <\/li> \n

  5. Luebbers, R.; Schuster, J.; Wu, K., \"Full wave propagation model based on moving window FDTD,\" Military Communications Conference, 2003. MILCOM 2003. IEEE , vol.2, no., pp. 1397-1401 Vol.2, 13-16 Oct. 2003<\/span><\/p> <\/li> \n

  6. M. F. Hadi and M. Piket-May, \u201cA modified FDTD (2,4) scheme for modeling electrically large structures with high-phase accuracy,\u201d IEEE Trans. Antennas Propagat. vol. 45, pp. 254-264, 1997.<\/span><\/p> <\/li> \n

  7. Lindholm, E., Nickolls, J., Oberman, S., and Montrym, J.  \u201cNVIDIA Tesla: A Unified Graphics and Computing Architecture,\u201d IEEE Micro 28, pp. 39-55, March 2008.<\/span><\/p> <\/li> \n

  8. Nickolls, J., Buck, I., Garland, M., and Skadron, K., \u201cScalable Parallel Programming with CUDA,\u201d Queue 6, pp. 40-53, March 2008.<\/span><\/p> <\/li> \n

  9. \u201cCUDA Programming Guide, 2.1,\u201d NVIDIA
    <\/span><\/p> <\/li> \n<\/ol> \n

     <\/p>","category":["Wireless Propagation","Wireless InSite"],"pubDate":"Fri, 19 Jun 2026 17:31:05 GMT","guid":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/cuda-implementation-of-moving-window-finite-difference-time-domain"},{"title":"Full Wave Matching Circuit Optimization Shortens Design Iterations | Remcom","link":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/full-wave-matching-circuit-optimization-shortens-design-iterations","description":"

    \n \"Full <\/a> \n<\/div> \n

    Full wave matching circuit optimization (FW-MCO) is a new technology that combines full wave, 3D electromagnetic (EM) simulation with circuit optimization into a novel approach for solving an age-old RF problem: determining which component values provide the desired match for a given matching network layout. Gone are the days of soldering components in and out of a prototype, trying to achieve the desired performance. This article describes the design process using the design of a matching circuit for a GPS-Bluetooth  antenna.<\/p>","category":["Mobile Device","Matching Network Design","CEO","Antenna Design","XFdtd"],"pubDate":"Fri, 12 Jun 2026 15:00:36 GMT","guid":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/full-wave-matching-circuit-optimization-shortens-design-iterations"},{"title":"Using Simulation to Optimize Safety, Performance, and Cost Savings When Integrating an Antenna Onto a Platform | Remcom","link":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/using-simulation-to-optimize-safety-performance-and-cost-savings-when-integrating-an-antenna-onto-a-platform","description":"

    \n \"Using <\/a> \n<\/div> \n

    Introduction<\/h2> \n

    Successful integration of an antenna onto a vehicle platform poses many challenges.  Vehicle features impact antenna performance by blocking, reflecting or reradiating energy, and co-site interference can impair the effectiveness of multi-antenna configurations.  Platform motion and environmental factors such as terrain and buildings may reduce system effectiveness in actual operational conditions.  Furthermore, radiation hazards may pose risks to nearby personnel.  Modeling and simulation provides a powerful tool to aid in understanding these issues and developing solutions.  This article provides a variety of examples of simulation-based assessments used to analyze antenna performance, identify problems, and evaluate potential solutions.<\/p>","category":["XGtd","Antenna Placement","Wireless InSite","XFdtd"],"pubDate":"Fri, 12 Jun 2026 14:44:00 GMT","guid":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/using-simulation-to-optimize-safety-performance-and-cost-savings-when-integrating-an-antenna-onto-a-platform"},{"title":"EM Modeling & Providing Narrowband IoT Coverage with Low Earth Orbit Satellites | Remcom","link":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/providing-narrowband-iot-coverage-with-low-earth-orbit-satellites","description":"

    \n \"EM <\/a> \n<\/div> \n

    Kenneth M. O\u2019Hara and Gregory J. Skidmore, Remcom Inc., State College, Pa. 

    This article describes the modeling of a SATCOM link, specifically the use case of using a satellite overlay to extend service continuity to IoT devices in a poorly covered rural area.<\/p>","category":["5G MIMO","Wireless InSite"],"pubDate":"Tue, 09 Jun 2026 13:57:38 GMT","author":"kaitlyn.brickley@remcom.com (Kaitlyn Brickley)","guid":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/providing-narrowband-iot-coverage-with-low-earth-orbit-satellites"},{"title":"Time Domain Simulation of Electrostatic Discharge Testing | Remcom","link":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/time-domain-simulation-of-electrostatic-discharge-testing","description":"

    \n \"Time <\/a> \n<\/div> \n

    An electrostatic discharge (ESD)<\/a> is the sudden flow of current between two electrically charged objects, caused by the breakdown of the dielectrics separating them, i.e., dielectric breakdown. In the case of electronic devices, the resulting current flow and possible spark can permanently damage the device (see Figure 1). An often recited yet unsubstantiated quote is \u201c\u2026losses associated with ESD in the electronics industry are estimated at between half a billion and $5 billion annually.\u201d In reality, estimating the exact cost of ESD loss is extremely difficult; nonetheless, ESD forces the development and testing of many hardware prototypes during design and manufacturing and contributes to a high number of warranty claims with loss of consumer confidence if a failure occurs in the hands of the consumer. Therefore, electronics manufacturers go to great lengths to properly shield sensitive components and design systems to reduce, dissipate and neutralize static charge.<\/p>","category":["ESD","XFdtd"],"pubDate":"Fri, 29 May 2026 17:01:56 GMT","author":"kaitlyn.brickley@remcom.com (Kaitlyn Brickley)","guid":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/time-domain-simulation-of-electrostatic-discharge-testing"},{"title":"XFdtd Analyzes Complex Beam Steering with Antenna Array Simulation | Remcom","link":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/xfdtd-analyzes-complex-beam-steering-antenna-arrays","description":"

    \n \"XFdtd <\/a> \n<\/div> \n


    As telecommunications standards have shifted toward higher frequencies such as millimeter waves, it has become necessary to incorporate arrays of antennas into communication systems. These more sophisticated arrays enable higher gain, smaller beam width patterns that can overcome the high path losses inherent to these frequencies and help deliver the higher throughput promised by these new standards. Even for small arrays with just a few elements,it can be difficult to fully understand the coverage possible with different power and phasing combinations. When arrays grow larger and contain tens or even hundreds of elements, advanced tools are required to fully describe the operation and effectiveness of the design.     XFdtd\u00ae Electromagnetic Simulation Software<\/a> contains analysis tools for rapidly characterizing the performance of arrays of antennas for beamforming and beam steering applications. Superposition allows for rapid synthesis of complex problems from basic data obtained from an electromagnetic simulation by quickly overlaying data in a linear space.<\/p> \n

    For example, consider multiple antennas in a problem space. Requisite data, such as S-Parameters, efficiency, and far-zone are computed once in the EM simulation for each antenna of interest. These may be quickly combined with complex scaling to yield the radiation pattern and associated active S-Parameters and efficiency for any desired antenna power and phasing combination. For situations where specific port power and phasing options are known ahead of time, the resulting far-zone pattern is computed immediately in post-processing.<\/p> \n

    A second situation is where the specific beam directions are known in advance and the power and phasing of the ports is desired to maximize the Effective Isotropic Radiated Power (EIRP) in those directions. Using statistical analysis, the coverage provided by the array for all possible directions may be determined by the calculation of the cumulative distribution function of the EIRP which shows the percent of the spherical far-zone volume with positive gain for a given input power.<\/p> \n

    We can also examine a 1x8 array of 28 GHz patch antenna elements where the phasing between elements is swept using the superposition feature from -90 to 90 degrees in 30-degree steps. Seven unique beams will be created which focus a fan beam between -30 and 30 degrees, as shown in Figure 1a. If the array is used in a configuration with two adjacent 1x4 subarrays, the superposition feature can be used to generate separate patterns for each subarray as well, such as shown in Figure 1b.<\/p> \n

    Alternatively, for situations where the desired beam directions are known and the port settings are desired, the array optimization feature is available. A large, two-dimensional array, such as an 8x8 patch array at 28 GHz, will create narrow beams which can be swept over a broad range of directions above the array. Figure 2 shows the Max Hold pattern created with the array optimization tool for a case where the beams are desired at every 15 degrees in the azimuthal direction and an elevation angle down 30 degrees from normal. Following the calculation, the necessary port power and phasing requirements of the elements to form each beam are available.<\/p> \n

    A large array may also be operated as a set of smaller sub-arrays which function independently. In Figure 3 the same 8x8 array is shown functioning as a set of four 4x4 arrays which each have their own unique power and phasing arrangements to communicate with unique receivers. Using the array analysis tools, the sub-arrays may be defined and the operating characteristics explored.<\/p> \n

    Multiple arrays on a single device are used for spatial diversity to provide coverage over a wider range of angles than a single array could produce. A possible use case for this would be a mobile telephone for 5G with multiple arrays located around the edges of the phone. Four arrays located on the sides, top, and bottom of a phone case, as shown in Figure 4, are each four element patch arrays that can produce a steerable beam that covers a wide region normal to the array. When used in combination, the arrays can provide coverage in multiple directions, which is best demonstrated with the CDF of EIRP calculation. This one-dimensional function describes the percentage of the far-field sphere that is covered by an array for a given input power. In Figure 5, CDF of EIRP plots are shown for different combinations of the arrays on the edges of the mobile phone, and it can be seen that the single arrays alone provide positive gain for about 66% (1 \u2013 0.34 from the CDF plot) of the directions for a typical 23 dBmW input power. Coverage improves to about 90% when two arrays in one of the corners are used in combination, and full coverage results when arrays on opposite sides of the phone are used together.<\/p> \n

    As device performance is pushed to new levels by the increasing demands of 5G high frequency systems, the need for more comprehensive analysis tools for complex antenna systems continues to grow. XFdtd simplifies the process for understanding device performance by providing efficient ways to validate array coverage.<\/p>","category":["Mobile Device","Antenna Design","5G MIMO","XFdtd"],"pubDate":"Tue, 05 May 2026 18:43:06 GMT","author":"kaitlyn.brickley@remcom.com (Kaitlyn Brickley)","guid":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/xfdtd-analyzes-complex-beam-steering-antenna-arrays"},{"title":"Assessing 5G Radar Altimeter Interference for Realistic Instrument Landing System Approaches | Remcom","link":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/assessing-5g-radar-altimeter-interference-for-realistic-instrument-landing-system-approaches","description":"

    \n \"Assessing <\/a> \n<\/div> \n


    Introduction<\/h2> \n

    In early March of 2020, the Federal Communications Commission (FCC) made 280 MHz of mid-band spectrum available for flexible use in the United States to address the critical need for unoccupied spectrum to deploy 5G wireless services [1].  Another portion of mid-band spectrum, that between 4.2 and 4.4 GHz, is reserved for use by radar altimeters operating on aircraft worldwide.  Radar altimeters are critically important for navigation, particularly during low visibility instrument landing system (ILS) approaches to airport runways, and are relatively vulnerable to interference.  By releasing the portion of bandwidth between 3.7 and 3.98 GHz for flexible usage, the FCC ensured a 220 MHz guard band interval would exist between future 5G emissions and the radar altimeter spectrum.  Though the FCC had determined the 220 MHz guard band sufficient to ameliorate 5G radar altimeter interference, a study prepared by the Radio Technical Commission for Aeronautics (RTCA, Inc.), using empirical data for radar altimeters representative of models currently deployed on civilian aircraft [1], concluded that their results \u201creveal a major risk that 5G telecommunications systems in the 3.7\u20133.98 GHz band will cause harmful interference to radar altimeters on all types of civil aircraft\u201d [2].  While the wireless communication industry and the companies (AT&T and Verizon) who won the auction for the released spectrum have taken issue with certain methodologies and conclusions of the RTCA study [3, 4], all interested parties brokered an agreement in early January 2022 to temporarily restrict and\/or limit emissions of 5G base stations in buffer zones around 50 designated US airports (see Figure 1) to provide time to resolve outstanding issues with radar altimeter equipment [5, 6].<\/p>","category":["Wireless Propagation","5G MIMO","Wireless InSite"],"pubDate":"Thu, 30 Apr 2026 18:30:42 GMT","author":"kaitlyn.brickley@remcom.com (Kaitlyn Brickley)","guid":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/assessing-5g-radar-altimeter-interference-for-realistic-instrument-landing-system-approaches"},{"title":"60 GHz Phased Array Antenna Simulation: WiGig Beamforming Analysis","link":"https:\/\/www.remcom.com\/resources\/examples\/60-ghz-phased-array-antenna-design-and-beamforming-analysis-using-xfdtd","description":"

    \n \"60 <\/a> \n<\/div> \n

    Introduction<\/h3> \n

    Designing a high-gain phased array for the 60 GHz WiGig band requires accurately capturing the full electromagnetic behavior of a multilayer, aperture-coupled structure\u2014including S-parameter coupling between elements, far-field radiation patterns, and beam-steered EIRP coverage. XFdtd\u2019s full-wave FDTD solver<\/a>, combined with its superposition and array optimization post-processing capabilities, addresses each stage of this workflow efficiently.

    This example models a dual-array, 8-element (2\u00d74) aperture-coupled patch antenna operating in the WiGig channel 2 band centered at 60.48 GHz (frequency range of interest: 55\u201365 GHz). Two orthogonally oriented 1\u00d74 arrays are mounted on a shared L-frame structure. The simulation characterizes individual element patterns, mutual coupling (S-parameters), superposed beamforming states, and beam-steered EIRP\u2014culminating in a CDF of EIRP for each array and a composite max-hold plot showing total spatial coverage.<\/p> \n

     <\/p>","pubDate":"Thu, 16 Apr 2026 16:25:50 GMT","author":"kaitlyn.brickley@remcom.com (Kaitlyn Brickley)","guid":"https:\/\/www.remcom.com\/resources\/examples\/60-ghz-phased-array-antenna-design-and-beamforming-analysis-using-xfdtd"},{"title":"Breaking Boundaries In Wireless Communication: Simulating Animated, On-Body RF Propagation","link":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/breaking-boundaries-in-wireless-communication-simulating-animated-on-body-rf-propagation","description":"

    \n \"Breaking <\/a> \n<\/div> \n

    By Tarun Chawla, Remcom<\/span><\/p>","pubDate":"Thu, 22 Jan 2026 17:38:11 GMT","author":"kaitlyn.brickley@remcom.com (Kaitlyn Brickley)","guid":"https:\/\/www.remcom.com\/resources\/articles-and-papers\/breaking-boundaries-in-wireless-communication-simulating-animated-on-body-rf-propagation"}]}}