Search

Phased-array radar systems using wide-band linear-FM pulses suffer from array dispersion. This causes a loss in signal-to-noise (S/N) ratio and in radar range resolution (a broadening of the compressed-pulse width). Depending on the choice of the receiver matched filter (MF) and weighting filter (WF) one or the other of these effects may be minimized. The effects of dispersion in a square planar array with either a parallel feed or a center-fed series feed are studied. Loss in S/N ratio and in range resolution are determined for typical cases, and general curves are given for arrays with arbitrary size, scan range, and signal bandwidth. It is shown that there is an optimum or maximum-useful signal bandwidth for each array, for which the minimum S/N-loss MF-WF design is the same as the ideal-compressed-pulse-shape MF-WF design. For pulses having greater signal bandwidths, time delay steering should be used in the array.

The identification and classification of radar targets is facilitated by a newly developed technique based on measurements of the differential phase shift of the target scattering obtained at harmonic phase-locked frequencies. These phase differences are independent of target range and relative motion. It is mathematically proved that for a smooth conducting body in the Rayleigh region the copolarized terms of the scattering matrix have zero phase shift. Based on this, the absolute target phase can be measured. The differential phase shift between two harmonic frequencies has been measured for various types of targets. The experimental results indicate that the differential phase shift is a useful parameter for classifying or identifying targets, and one that can be readily measured in practical radars.

The mechanism by which electromagnetic waves are scattered from the sea surface is studied herein. Principal applications for such studies occur in over-water communication channels and radar. Predictions based on rigorous formulations for diffraction from a (moving) time-space periodic surface are correlated with the experimental results of other authors. Such predictions explain the existence of strong coherent scattering from the sea surface in nonspecular directions (grating lobes) and also anticipate the appearance of well-defined, harmonically related, frequency shifts (time-space harmonics) in scattered signals observed in these experiments. On the basis of such manifestations of the periodic nature of the sea surface, a model is formulated for use in calculation of the angular power spectrum of scattered electromagnetic waves. This model, accounting for spatial coherence effects, assumes for the fluctuation of heights on the water surface a Gaussian distribution with a narrow band-pass (nearly sinusoidal) spatial spectrum. Calculations of the angular, scattered spectrum due to this model show strong predominantly coherent wave summation in directions corresponding precisely to the grating lobes of a purely periodic structure of period equal to the center period of the band-pass spectrum. Spectral (angular) spreading increases with increased bandwidth of the spatial spectrum of the surface. Calculations for the present study are made on the basis of a perfectly reflecting boundary and hence do not display the proper distinction of polarization expected for a completely realistic model. The results, however, display the spectral charactral characteristics necessary to explain strong scattering in nonspecular directions by a random surface.

By carefully accounting for the effects of Faraday rotation, certain of the polarization properties of radar echoes from the moon at 25 MHz could be estimated even though only a single-fixed linearly polarized antenna was used. The combination of Faraday rotation and antenna response sinusoidally modulates the linear FM signals used. This "Faraday modulation" does not appreciably affect the range resolution (which was about 20 \mu S) but strongly affects the proportions of the cross-polarized components of the echoes. The measured frequency and phase of the Faraday modulation on each echo allows the determination of the proportions and, hence, the separate estimation of the components. Because the data reduction primarily involved digital spectral analysis, the effects of the modulation are interpreted in terms appropriate to power spectral estimation. The techniques used, but not the results, are discussed here.

Electromagnetic waves are launched and collected by the same directive radar antenna located on a reentry vehicle. The scattering obstacle is the fireball resulting from a nuclear detonation. The objective of this communication is to determine approximately the incoherent power scattered from the rough surface of the sphere available at the source at a specified time. The received power depends on many factors including the scattering cross section per unit area of the sphere and the area illuminated. The area illuminated is determined by the intersection of the sphere with a volume of space that is a function of the radar pulsewidth and the time of interest.

This paper discusses the design and construction of a branch-fed, variable-phase, 20 to 55 Mc array of 48 log-periodic elements. Phasing to control-beam direction is described and the correction necessary for ionospheric refraction is presented in terms of vertical incidence critical frequency. The array gain as measured by moon-bounce techniques is found to be in good agreement with the theoretical value.

An investigation has been made of the vertical-plane dual-beam properties of a doubly curved shaped-beam reflector to provide design information relative to possible modifications for a large class of surveillance antennas (in particular, the cosecant-squared family) in order to improve their performance when working in the presence of serious ground clutter problems. Experimental results obtained with an X -band model reflector utilizing various dual-feedhorn geometries are presented. Measured secondary patterns demonstrating several important performance characteristics are shown for a linearly polarized situation. Measured primary patterns for the feeds which generate these results are included, Two-way radar coverage diagrams for one set of dual-beam patterns are examined in detail. These plots make it clear that for many ground clutter minimizing applications it appears desirable to mix at RF the signals received by both feedhorns.

Some preliminary work relating to the pulse radar return from ionized media, such as fireballs resulting from nuclear explosions and the wakes of high-speed reentry vehicles, is reported. In all cases the objective is to retrieve the envelope of the backscattered signal correctly related in time to the transmitted pulse.

In many monostatic configurations, the effective scattering volume is defined longitudinally by the radar gating and transversely by either the antenna beamwidth or the extent of the scattering region. Often the radar signal must traverse fluctuating transmission regions on its way to and from the range cell. Consequently, the return signal is perturbed not only by events within the range cell but contains statistical characteristics ascribable to the fluctuating transmission path as well. The limit of large-scale fluctuations (which is of interest in the current context) is dominated by strong incoherent small angle scattering. A systematic analytic sequence resembling that pursued by Bremmer [1] (though significantly complicated by the three dimensionality and temporal dependence of the fluctuations) constitutes the main subject matter of this study. The proposed iterative scheme, while properly accounting for multiple small angle scatter, orders successive iterates by the number of large angle scattering events they represent. While the discussion is carried within the physical framework of an electromagnetic (radar) application, an extension to the acoustical (sonar) analog is simple. Furthermore, the extension of the presently investigated monostatic configuration to bistatic ones is possible but is by no means trivial. The reported work has initially been motivated by an attempt to interpret on board radar cross section and spectre measurements associated with an ionized wake.

A mathematical model for microwave backscatter from the sea at near-vertical incidence is presented. The model, which is formulated in terms of Huygens principle, expresses the backscatter in terms of an "expected transfer function" of the round-trip path between the antenna and the surface of the sea. In its present form, the model is restricted to describing the case in which the microwave-illuminated area of the sea is large enough for the distribution of illuminated sea-surface area in elevation to be accurately known. It differs from other models in that it does not make the assumption that the incident electromagnetic field is plane or that the antenna is located in the Fraunhofer zone of the scattering surface. The analysis is carried out in terms of a general probability-density function of sea-surface elevation. However, an approximately Gaussian probability-density function is assumed for the purpose of working out a detailed example. The most interesting feature of the model is that it indicates that values of average radar cross sections per unit area of the sea ( \sigma \deg ), when measured at near-vertical incidence, cease to be independent of range and of antenna gain. Needed extensions of the model should treat the case in which the illuminated area size is small, and the problem of mathematically describing the boundaries between the large and small illuminated area cases. A number of experiments are recommended for verifying predictions of the model.

Users of extended range HF (3 to 30 MHz) radar and communication systems employing the ionosphere desire signal reception at incidence angles near-grazing to the local earth tangent. For vertical polarization, the vanishing received fields at low incidence angles over dielectric earth may be increased by using large ground screens. In this paper a ground-screen formulation based on scattering techniques is developed. The ground screen is viewed as a scatterer in free space, excited by a plane wave. A Fresnel image wave is added to establish the air-earth interface. Formulations are developed for the semi-infinite screen and for the circular-cylinder surface segment screen. The semi-infinite screen is representative in performance to a round screen of radius equal to the distance from the edge of the semi-infinite screen at which the field is computed or measured. For a ground screen on a hill the cylindrical segment is appropriate. Computations for a simulated earth were made and corroborated by experimental simulation with scale models. Improvements in field strength of 7 to 14 dB or more can be achieved with large screens over no screen. "Relatively small" tilted or raised flat screens, and cylindrical segment screens, can give improvements equal to very large flat screens on the earth.

A technique is developed for measuring the rms wave height averaged over an area of the sea that is much greater than any horizontal scale of the surface waves. The method involves a nadir looking radar which transmits and receives two monochromatic signals simultaneously. Signal processing at the receiver involves the computation of the correlation between the two returning signals as a function of theft variable frequency separation. The cross correlation between the amplitude and phase functions of the individual returning carriers depends on the distribution of discrete scatterers along the direction of propagation. This information can be used to determine the rms surface elevation (about the mean); it does not depend on the temporal or spatial frequency spectrum of the wave height or slope. Under conditions which are typical for a microwave signal being normally incident and reflected by the sea, the two frequency correlation function R(\Delta k) is seen to be equal to the characteristic function of the surface elevation of specular points. Laboratory measurements have been conducted on wind driven waves, and the measured correlation function compares favorably with the theory developed here.

High-resolution radar becomes limited by the feeding mechanism in traveling wave antennas, forcing the design to some form of corporate feed. A spectral method of solving this problem is introduced which conveniently handles any transmitter pulse shape or receiver bandpass. The equivalence between this approach and an older transient solution is established. The method is applied to a uniform and a cosine illuminated aperture for both one- and two-way transmission through the antenna. It is shown that as a rule of thumb, the traveling wave feed should be less than half the spacial length of the transmitter pulse in the feed ( L/V_{g}T ). Antennas exceeding this length must be divided into several sections, each of which must be corporately fed.

In attempting to estimate the radar cross section of airborne vehicles, it is often only necessary to consider average values of the radar returns. A method of providing a quick estimate of the average bistatic radar cross section of the vehicle components would be useful. Ray optics provides a method of predicting the radar cross section of electrically large, perfectly conducting, simply curved, convex bodies such as spheres, ogives, ellipsoids, etc. This paper extends the method of ray optics to the case of an arbitrary body, which may be concave and/or convex, on which doable reflection and depolarization can occur. The incident radiation on the scattering body is represented by a large number (10^{5}-10^{6}) of rays. The rays reflected in a given direction with a given polarization are collected at infinity and combined by phasor addition. For the bodies investigated, this method yields results within \pm2 dB of measured results except in small angular regions where trapped waves contribute significantly to the radar return.

A method for obtaining a desired radar scattering pattern by reactively loading a conducting body is given. The theory uses the concept of characteristic modes of a loaded body. Any desired real current can be resonated by reactive loads to make it the dominant mode current of that body. If no other mode is near resonance, the radar scattering pattern becomes nearly the same as the radiation pattern of the resonated current. A quality factor Q is defined as a measure of the broadband behavior of a scatterer. Procedures for computing the real currents having minimum Q and maximum gain-quality ratio are given. A pattern synthesis procedure is developed for obtaining the real current whose radiation field pattern is the least mean-square approximation to a desired field pattern. Numerical examples are given for each procedure discussed.

A method is proposed for improving the azimuthal resolution of HF radar several times beyond the limitation due to ionospheric scattering. Azimuthal resolution improves directly with array size, in accordance with standard array theory, until the beamwidth shrinks to a few degrees. Further increase in array size does not produce a dependable resolution improvement. In radio parlance, the signal is said to be decorrelated across the aperture. This "decorrelation" is due to scattering by the nonhomogeneous ionosphere. The scattered energy accounts for only a portion of the received wave. The remainder, an almost specular component, not only has a very much narrower angular distribution of its energy, but appears to have a much longer (perhaps 100 times) time constant of fluctuation. This difference in fluctuation rate permits separation of the two components by filtering the received radar echo. The decorrelating component is thus reduced, and the available azimuthal resolution, i.e., the resolution limitation due to scatter, correspondingly increases. When literature data are applied to the theory, a several-fold increase in available azimuthal resolution is predicted.

The standing-wave method of measuring radar cross sections has been adapted for use at 26 MHz. A target in the middle of a large flat field was illuminated by a transmitter 300 feet away, and a probe was moved along the transmitter-target line to measure the amplitude and position of the standing wave due to the target. Aluminum dipoles of different lengths were first used as targets. Both the target cross section and the phase of the scattered radiation varied with the target length in the same way as predicted from previous microwave measurements. The results demonstrate the validity of extrapolating microwave measurements to the high-frequency part of the spectrum. Calibration of the system was achieved by using the known cross section of a half-wave dipole. The technique was used to measure cross sections of trees, street lights, and vehicles. With vertical polarization, oak trees had cross sections between 30 m2and 70 m2, indicating that trees can be an important source of ground backscatter. A street light had a cross section of 240 m2, and vehicles had cross sections ranging from 0.7 m2to 28 m2, depending on aspect.

Compact range techniques for measuring the gain patterns of full-size microwave antennas and for making radar reflectivity measurements are described. The basic principle of this technique is the use of a large collimating device to generate a uniform plane wave across the aperture of a target or antenna without requiring the normal far-field separation. Two different collimating devices were used in the investigation, a paraboloid with a point-source feed and a parabolic cylinder with a line-source feed generated by a large hoghorn. Pattern and gain measurements were made on both compact ranges using a 30-inch paraboloidal test antenna, and the measurements were compared with similar ones made on conventional outdoor ranges. Radar cross-section patterns as a function of aspect angle were measured for various size standard targets and compared with theoretically calculable radar cross-section patterns. The results which have been achieved are very encouraging. They demonstrate that the performance of compact ranges at the X -band is comparable to that of outdoor ranges.

When array antennas are used for wide-frequency operation a difficulty with grating lobes is encountered at the upper frequency end of the operating band. This paper points out that the use of arrays for transmission and reception with different interelement spacings can be used to extend the operating frequency band in a radar system. The principle of transmitting-receiving arrays with different illuminations was used by Mattingly [2] to lower the sidelobe level of a radar array antenna. Birge [3] used it to reduce the number of elements in an array while preserving the resolution. The same principle is used in this paper to extend the operating frequency bandwidth of a radar antenna system. Relationships for an optimized radar array antenna system are derived. The criterion for optimization is the maximum bandwidth with a given total number of elements. It is shown that in an optimized system the number of elements in one array antenna is one greater than in the other, and the ratio of interelement spacings is inversely proportional to the square of the number of elements in the respective arrays. Based on these conditions, a function k is computed and tabulated. This function determines the number of grating lobes permissible in the specified angular sector, without overlap, as a function of the number of elements in the two arrays. An application to a high frequency (HF) ionospheric radar antenna system, now under construction, is used as an illustration of the principles outlined in the paper.

An antenna capable of high resolution can be synthesized by moving a side-looking radar a long distance and treating positions along the path as the positions of elements in an actual physical array. Important characteristics of such synthetic-aperture radar systems and the effect of target motion on their resolution have been considered. If the motion of the target is known, a transformation can he made in which the target appears stationary, and the relative motion between target and radar determines the effective synthetic aperture. However, if the assumed velocity of the target differs from its actual velocity, the velocity error will cause an apparent displacement of the target and a degradation of the performance of the radar. The effect of velocity errors was considered.

Previous work indicates that although accurate predictions are generally obtained from the geometrical theory of diffraction for cylinders and frustums, scattering from the rear edges of these bodies is sometimes poorly predicted. A new formulation provides satisfactory results for a cylinder, but conical bodies require further work.

An expression is derived for the average backscattering cross section of an obstacle for arbitrary transmitter and receiver antenna polarizations. Results are given for straight wires, circular loops, and one-turn helices.

The bistatic scattering cross section of a reentry vehicle with wake of nonuniform axial conductivity is investigated using a simple model. For some typical physical parameters normalized radar cross sections are presented in tabular form for a reentry vehicle with wake of exponentially decreasing axial conductivity.

An effort recently undertaken to compute radar cross sections of thin metallic disks for frequencies near resonance is described. All calculations are based upon an ideal disk scattering problem in which each disk is assumed to be perfectly conducting and infinitely thin and the incident field is assumed to be an electromagnetic plane wave. A good agreement between the values of the backscattered fields for the ideal disk scattering problem and the thin metallic disk scattering problem for large values of ka is found, but may become poor for ka less than or equal to 1.0.

Introduces computer programs (FORTRAN IV) designed for field calculation from conducting bodies of revolution. A set of four programs which calculate the generalised impedance matrix and admittance matrix for triangular expansion functions, bistatic radar scattering for an axially incident plane wave, radiation from rotationally symmetric aperture antennas, and radar backscattering for an obliquely incident plane wave, are presented. Results for a number of examples are summarised and discussed. The accuracy of computation depends on the smoothness of the body and of the excitation

High-frequency scattering by a finite perfectly conducting right-circular cone is analyzed by modifying geometrical diffraction theory for application at and near nose-on aspects. Formulas appropriate for predicting principal-polarization radar cross sections as functions of frequency and aspect angle are presented. Calculations on a cone are in very good agreement with monostatic measurement data in the limited aspect interval 0 \leq \phi \leq \gamma , where \gamma , is the half-angle of the cone.

Radar and dc pulses of Gaussian shape are reflected from dry, damp, and wet earth for both vertical and horizontal polarization. The envelope of the radar pulse is retrieved following reflection and examined for distortion. The shape of the reflected dc pulse is also determined. Pulse distortion resulting from multiple reflections is easily formulated by constructing a composite transfer function consisting of the product of reflection coefficients and retardation factors.