Edward E. Altshuler

Wire-antenna designs using genetic algorithms

There is a large class of electromagnetic radiators designated as wire antennas. As a rule, an inductive process is used to design these antennas. Either an integral equation is formulated or a simulator is used that gives the current distributions on the wires of the antenna, from which the electromagnetic properties of the antenna can then be determined. Once the antenna properties are known, the parameters are optimized, using guides such as intuition, experience, simplified equations, or empirical studies. However, using an electromagnetics simulator in conjunction with a genetic algorithm (GA), it is possible to design an antenna using a completely deductive approach: the desired electromagnetic properties of the antenna are specified, and the wire configuration that most closely produces these results is then synthesized by the algorithm. In this paper, we describe four antennas designed using GAs. The first is a monopole, loaded with a modified folded dipole that was designed to radiate uniform power over the hemisphere at a frequency of 1.6 GHz. The second antenna consists of seven wires, the locations and lengths of which are determined by the GA alone, that radiates waves with right-hand-circular polarization at elevation angles above 10/spl deg/, also at 1.6 GHz. The last two antennas are modified Yagis. One is designed for a broad frequency band and very low sidelobes at a center frequency of 235 MHz. The other is designed for high gain at a single frequency of 432 MHz. We have built and tested these antennas.

Electrically small supergain end‐fire arrays

The theory, computer simulations, and experimental measurements are presented for electrically small two-element supergain arrays with near optimal endfire gains of 7 dB. We show how the difficulties of narrow tolerances, large mismatches, low radiation efficiencies, and reduced scattering of electrically small parasitic elements are overcome by using electrically small resonant antennas as the elements in both separately driven and singly driven (parasitic) two-element electrically small supergain endfire arrays. Although rapidly increasing narrow tolerances prevent the practical realization of the maximum theoretically possible endfire gain of electrically small arrays with many elements, the theory and preliminary numerical simulations indicate that near maximum supergains are also achievable in practice for electrically small arrays with three (and possibly more) resonant elements if the decreasing bandwidth with increasing number of elements can be tolerated.

A monopole superdirective array

In principle, the end-fire directivity of a linear periodic array of N isotropic radiators can approach N/sup 2/ as the spacing between elements decreases, provided the magnitude and phase of the input excitations are properly chosen. Thus, the directivity of a two-element array of isotropic radiators would approach a value of four, that is, 6 dB higher than that of a single isotropic radiator. We have conducted a theoretical, computational, and experimental study for a two-element superdirective array of resonant monopoles. In agreement with the theoretical and computational curves, the measured gain of the monopole array does indeed continually increase with decreasing spacing of the monopoles, provided the relative magnitudes and phases are maintained. However, for very small separation, maximum achievable gain is not reached due to the presence of ohmic loss.

Electrically small self-resonant wire antennas optimized using a genetic algorithm

One of the major limitations of electrically small antennas is that as the size of the antenna is decreased its radiation resistance approaches zero and its reactance approaches plus or minus infinity. Most small antennas are inefficient, nonresonant and, thus, require matching networks. In this investigation, we use a genetic algorithm (GA) in conjunction with the numerical electromagnetics code to search for resonant wire shapes that best utilize the volume within which the antenna is confined. Antenna configurations, over a ground plane, having from two to ten wire segments, were optimized near 400 MHz and then built and tested. As the cube size deceased from a side length of 0.096/spl lambda/ to 0.026/spl lambda/, the computed Qs increased from 15.8 to 590. The measured Qs increased from 16.0 to 134 for cubes of 0.093 to 0.037/spl lambda/ on edge. This process for designing small antennas using a GA produced new self-resonant antenna configurations.

Design of a loaded monopole having hemispherical coverage using a genetic algorithm

A genetic algorithm is used to design a monopole loaded with a modified folded dipole so that it radiates uniform power over the hemisphere. Each of the wires that make up the antenna are given a range of possible lengths. The genetic algorithm randomly selects a sample population of possible antenna configurations from the total population of all configurations. The radiation pattern of each sample configuration is computed using the numerical electromagnetics code (NEC). The solutions are compared with the desired pattern and ranked in terms of performance. The best solutions are retained and mated with one another and the process is repeated until an optimal solution is obtained. The genetic algorithm quickly produced an antenna that has a nearly uniform power over the hemisphere. Although the antenna was designed to operate at a frequency of 1.6 GHz, it performed satisfactorily over the frequency range from 1.4 to 1.8 GHz. The antenna was fabricated and the computational results were verified experimentally. We have shown that the genetic algorithm is a very powerful tool for designing wire antennas; it is expected that this process can be used to design any antenna that can be analyzed using an electromagnetic code.

Design of a vehicular antenna for GPS/Iridium using a genetic algorithm

In this paper, we describe a vehicular wire antenna, designed using a genetic algorithm that may be used for both the GPS and Iridium systems. It has right-hand circular polarization, near hemispherical coverage, and operates over the frequency band from 1225 to 1625 MHz. This antenna was simulated using the numerical electromagnetics code (NEC) and then fabricated and tested. The antenna consists of five copper tubing segments connected in series, has an unusually odd shape, and is very inexpensive. It fits in a volume approximately 10 cm/spl times/10 cm/spl times/15 cm, The input voltage standing wave ratio (VSWR) and circular polarization radiation patterns were computed and measured. The VSWR was under 2.2 at the design frequencies of 1225, 1575, and 1625 MHz. The gain varied by less than 12 dB for a 170/spl deg/ sector; it generally fell off near the horizon so the variation was less for 150/spl deg/ and 160/spl deg/ sectors. This new design process, which uses a genetic algorithm in conjunction with an electromagnetics code, produces configurations that are unique and seem to outperform more conventional designs.

An Impedance-Matched 2-Element Superdirective Array

With proper adjustment of the amplitude and phase of the element excitations, the directivity of an N-element end-fire array of isotropic radiators can approach N2 as the spacing between the elements approaches zero. To achieve this end-fire superdirectivity for a 2-element array, the excitation phase difference between the elements must closely approach 180deg. When closely spaced elements are driven nearly 180deg out-of-phase, the element radiation resistances approach zero, resulting in a high input Voltage Standing Wave Ratio (VSWR) and reduced radiation efficiency. At the same time, there is also a significant decrease in the operating bandwidth. In this letter we present the design of a 2-element, superdirective multiple arm folded monopole array that achieves a near 50 Omega input radiation resistance at each element, resulting in a matched input VSWR, higher radiation efficiency and therefore, a substantial increase in realized or overall efficiency.

Evolving wire antennas using genetic algorithms: a review

Communication, radar and remote sensing systems employ thousands of different types of wire antennas, and there is an increasing need for high-performance, customized antennas. Current methods of designing and optimizing them by hand using simulation or analysis are time- and labor-intensive, limit complexity, increase the cost and time expended, and require that antenna engineers have significant knowledge of the universe of antenna designs. Local optimization methods are not much better, since an initial guess that is close to the final design must be provided. Using a genetic algorithm (GA), it is possible to prescribe the desired performance of an antenna and allow the computer to find the parameters for the design. The GA does not require an initial guess, and the amount of design information the engineer must supply can be very minimal. This paper presents a review of a few wire antennas from previous publications designed by GA unconventional purposes. This approach has potential to revolutionize antenna design.

Tropospheric range-error corrections for the Global Positioning System

The Global Positioning System (GPS) is a highly accurate navigation system that has a broad spectrum of military, civilian, and commercial applications. It uses a triangulation scheme based on the time delays of signals from the satellites to the user; these time delays are then equated to distances. However, as the timing signal passes through the Earths atmosphere it undergoes an additional time delay due to the index of refraction. The time delay produced by the troposphere approaches a value corresponding to a range error of about 25 m for an elevation angle of 5/spl deg/ and decreases with increasing elevation angle to less than a few meters at zenith. It has been shown that there is a good correlation between the range error and the surface index of refraction. Worldwide statistics of surface refractivity have been analyzed and shown to be correlated with site latitude, height above sea level, and time of year. Regression lines for range-error corrections based on these parameters are derived. Range-error accuracies vary from about 8% down to 3.7% of the total range error, depending on the amount of information that is available.

Military antenna design using simple and competent genetic algorithms

Over the past decade, the Air Force Research Laboratory (AFRL) Antenna Technology Branch at Hanscom AFB has employed the simple genetic algorithm (SGA) as an optimization tool for a wide variety of antenna applications. Over roughly the same period, researchers at the Illinois Genetic Algorithm Laboratory (IlliGAL) at the University of Illinois at Urbana Champaign have developed GA design theory and advanced GA techniques called competent genetic algorithms-GAs that solve hard problems quickly, reliably, and accurately. Recently, under the guidance and direction of the Air Force Office of Scientific Research (AFOSR), the two laboratories have formed a collaboration, the common goal of which is to apply simple, competent, and hybrid GA techniques to challenging antenna problems. This paper is composed of two parts. The first part of this paper summarizes previous research conducted by AFRL at Hanscom for which SGAs were implemented to obtain acceptable solutions to several antenna problems. This research covers diverse areas of interest, including array pattern synthesis, antenna test-bed design, gain enhancement, electrically small single bent wire elements, and wideband antenna elements. The second part of this paper starts by briefly reviewing the design theory and design principles necessary for the invention and implementation of fast, scalable genetic algorithms. A particular procedure, the hierarchical Bayesian optimization algorithm (hBOA) is then briefly outlined, and the remainder of the paper describes collaborative efforts of AFRL and IlliGAL to solve more difficult antenna problems. In particular, recent results of using hBOA to optimize a novel, wideband overlapped subarray system to achieve -35 dB sidelobes over a 20% bandwidth. The problem was sufficiently difficult that acceptable solutions were not obtained using SGAs. The case study demonstrates the utility of using more advanced GA techniques to obtain acceptable solution quality as problem difficulty increases.