Asymptotic High Frequency Methods

Ching-Min Lien 2005.04.18

Introduction
Asymptotic High Frequency Methods

Geometrical Optics (GO) Physical Optics (PO) Geometrical Theory of Diffraction (GTD)
Uniform Geometrical Theory of Diffraction (UTD)

Physical Theory of Diffraction (PTD)

Geometrical Ray Optical Field

Geometrical Optics (GO) employs ray concepts to describe EM wave propagation mechanisms. Fermats Principle

The length of propagation ray path is stationary and minimized. In particular, the ray path in a homogeneous medium is a straight line.

Geometrical Ray Optical Field (cont.)

An infinitesimally narrow ray tube

Geometrical Ray Optical Field (cont.)

Luneberg-Kline Series

For high frequency EM waves, only the leading term is retained.

To satisfy the zeroth-order transport equation the leading term above can be solved as

Geometrical Ray Optical Field (cont.)

Because of the energy conservation,

Geometrical Ray Optical Field (cont.)

The variation of the cross section

Geometrical Ray Optical Field (cont.)

How to describe a curvature?

Geometrical Ray Optical Field (cont.)

The surface can be described as

Geometrical Ray Optical Field (cont.)

Gaussian Curvatures

: radii of principal and orthogonal wavefront curvatures. The Gaussian curvatures can be represented as

It is the determinant of the curvature matrix.

Furthermore,

Geometrical Ray Optical Field (cont.)

Hence, the field can be approximated as

where The EM field will satisfy the following relations since a GO fields are locally plane.

Geometrical Ray Optical Field (cont.)

Note

are perpendicular to one another. : ray caustics.

GO predicts infinite field strength there, and hence is not valid.

Three special cases:

Either Both : Spherical wave. or is infinite : Cylindrical wave. and are infinite : Plane wave.

Geometrical Ray Optical Field (cont.)

Spherical wave

finite,
Cylindrical wave Plane wave

Geometrical Ray Optical Field (cont.)

Moreover,

A phase jump of occurs while the propagation path crosses a ray caustic.

Geometrical Ray Optical Field (cont.)

The GO expression can be simplified as

where

which satisfies (1)

(1) can work with any two orthogonal directions transversal to In general, (1) is not diagonal unless planes of principal wavefront curvature are selected.

Transformation of Source Radiation into GO Field

In most practical applications, EM fields radiated from current sources need to be transformed into GO ray field. Far field of a radiation source can be expressed as

With a point source assumption, where can be chosen arbitrarily, and thus one may simply set

Transformation of Source Radiation into GO Field (cont.)

E-field can be solved as

where

Similarly,

Transformation of Source Radiation into GO Field (cont.)

Hence, the far field can be approximated as

where H-field can be founded with similar procedure. The power density of the EM wave is .

Congruent with the conservation of the energy.

Ray Tracing Techniques and GO Reflections

GO causes discontinuities at incident and reflection shadow boundaries (ISB/RSB).

Null field exists in the shadow region.

Ray Tracing Techniques and GO Reflections (cont.)

Fermats principle or Snells law are used to determine the propagation ray paths and the reflection point hit by the incident ray.

Ray Tracing Techniques and GO Reflections (cont.)

The relationship between the incident direction , reflection direction , and the normal unit vector of the surface. The pair defines the plane of incidence.

Ray Tracing Techniques and GO Reflections (cont.)

The direction of the reflection GO ray is In the neighborhood of the axial ray Oz, we can represent the surface by where is the curvature matrix.

Ray Tracing Techniques and GO Reflections (cont.)

The reflected and incident tube should coincides at the reflection point

The phases are matches in both linear and quadratic terms. The linear condition is Snells law applied to the axial ray. The quadratic condition is
When , it can be reduced.

Ray Tracing Techniques and GO Reflections (cont.)

After the ray tube is bounced from the surface at the curvature will be changed as for a homogeneous and isotropic medium, where

are curvature matrix of incident, reflection GO rays and surface, respectively.

are the transformation matrix between the coordinate system of incident/reflection GO rays and that of the local surface.

Ray Tracing Techniques and GO Reflections (cont.)

With these parameters, GO rays can be expressed as

where and reflection coefficient

are related by the surface as

Ray Tracing Techniques and GO Reflections (cont.)

The dyadic reflection coefficient, , is in general a 2x2 matrix. However, if the incident and reflection fields are resolved into components parallel and perpendicular to the plane of incidence, it can be simplified and diagonalized as

where

and

are

The Fresnel reflection coefficients of the TE and TM modes. The hard and soft reflection coefficients. 1 and -1 (?)

Ray Tracing Techniques and GO Reflections (cont.)

Finally, the total GO field is given by where

is the Heaviside step function.

is determined by the ISB is determined by the RSB It shows discontinuities across the incident and reflection shadow boundaries.

Shooting and Bouncing Rays Method

Shooting and bouncing rays method (SBR) simplifies the ray tracing procedure.

Represent the source radiation in terms of discrete ray tubes. Then trace each ray tube independently. Each ray tube represent a pulse function that has uniform amplitude and quadratic phase variation over its cross section.

Shooting and Bouncing Rays Method (cont.)

The phase in the incident tube is with

Hence, the GO ray can be described as

Shooting and Bouncing Rays Method (cont.)

The SBR method is outlined as

Represent the source radiation in terms of ray tubes, and the directions can be determined. Determine the shape of ray cross section.
Circular, triangular or quadratic shapes are polular.

Trace each ray propagating path and its associated shape of cross section.

Shooting and Bouncing Rays Method (cont.)

Trace the central ray only.

Does not concern the shape change during propagation.

Corner rays as well as the central ray are traced simultaneously.

The shapes of ray tubes are determined by corner rays, power conservation is assumed to be satisfied.

Shooting and Bouncing Rays Method (cont.)

SBR in the analysis of a cavity scattering problem.

Shooting and Bouncing Rays Method (cont.)

Incident plane wave is discretized in terms of a few ray tubes.

Only central rays are traced. The shapes of tubes are assumed to be circular.

Each ray is traced inside the cavity until it hits the open-end aperture again.

A set of equivalent currents are defined on the aperture. The scattering field are radiated from equivalent currents.

Shooting and Bouncing Rays Method (cont.)

The expression remains valid as long as the field point is not located extremely close to the aperture. Ray caustics of the reflected ray will make errors.

Errors begin to increase with wider angles. Errors increase in longer cavity in which more bounces of rays occur.

Shooting and Bouncing Rays Method (cont.)

The radar cross section (RCS) of a circular cavity with radius of and length of .

Hybrid model. SBR method.

Generalized Ray Expansion Method

Two-step ray expansion

A set of equivalent currents are defined by the EM fields radiated from original sources. Ray expansion is performed on the field radiated from

The ray paths launched from the locations of remain the same regardless of the variation of original sources.

Generalized Ray Expansion Method (cont.)

GRE method needs to trace the ray paths only once while the ray tracing procedure needs to be repeated. However, the number of rays is much greater

All possible ray directions need to be considered.

Generalized Ray Expansion Method (cont.)

Aperture of the open end is dubdiveded into subapertures Rays are launched radially from the phase center of each subaperture. The same set of rays are employed in the expansion of radiation from the aperture regardless of the directions of incident waves.

Physical Optics
Physical optics (PO) is an equivalent current method based on GO approximation and extend its applications, only in the lit regions.

Lit region: Illuminated by a given field. Shadow region: Not illuminated by the field.

The total EM fields on the object surface obtained by GO approximation

where

Physical Optics (cont.)

For a PEC

Null tangential electric field on its surface. Magnetic current vanishes. Electrical current With image theory on a local planar surface, the scattering field outside the object

which is valid for a large surface radii of curvature compared with wavelength.

Section 3
Image Theory

The PEC can be assumed as an infinite plane. To calculate the field generated according to the monopole. (a), (b) (c), (d)

Physical Optics (cont.)

The surface used in the integral for PO is limited to the lit region as that for GO. The interfaces between lit and shadow regions form edge type of boundaries.

Edge diffraction

The current distributions in PO approximation do not include any effect caused by edges of region surfaces.

Physical Optics (cont.)

If the scattering field integral is evaluated asymptotically by stationary phase method, the saddle point contribution will give rise to the reflected field as predicted by GO.

PO can only predict accurate fields in the specular regions where reflection contributions dominate. However, PO provides uniform field distributions and transitions at any place, potentially extending the applicability of GO.

Geometrical Theory of Diffraction and Its Uniform Version

GO

Discontinuous fields across the ISB and RSB. Null fields in the shadow regions. First introduced by Keller. Modified by Pathak and Kouyoumjian later (UTD) Correct the GO based on interpretation of diffraction mechanisms.

Geometrical Theory of Diffraction (GTD)

Geometrical Theory of Diffraction and Its Uniform Version (cont.)

Law of diffraction

A diffracted ray and the corresponding incident ray make equal angles with the edge at the point of diffraction

Cone of diffracted rays.

Geometrical Theory of Diffraction and Its Uniform Version (cont.)

Law of diffraction

If the incident ray is perpendicular to the edge, the diffraction cone degenerate to a disk. In two-dimensional case, all rays are by definition perpendicular to the edge.

Geometrical Theory of Diffraction and Its Uniform Version (cont.)

The overall propagation mechanism

Geometrical Theory of Diffraction and Its Uniform Version (cont.)

The total field is the superposition of GO field and diffracted field

The diffracted field not only provides contribution in the shadow region, but also provides uniform field distribution across ISB/RSB.

The diffracted fields can be expressed as where

: Diffraction coefficient, depending on the local nature. : Divergence factor due to ray propagation.

Geometrical Theory of Diffraction and Its Uniform Version (cont.)

Diffracted field:

The dyadic edge diffraction coefficient can be determined as and the divergence factor is hence

Geometrical Theory of Diffraction and Its Uniform Version (cont.)

Geometrical Theory of Diffraction and Its Uniform Version (cont.)

Choose and orthogonal to

are chosen on the plane

For a PEC wedge,

Geometrical Theory of Diffraction and Its Uniform Version (cont.)

F is a transition function

Renders uniform total field distribution while the field point moves across ISB/RSB.

where is an integer that most closely satisfies the relation

Geometrical Theory of Diffraction and Its Uniform Version (cont.)

are used to determine the size of transition region.

If the incident field is a spherical wave,

and

if the wedge has a straight edge.

Diffracted field in the vicinity of ISB/RSB is

Upper/Lower sign applies on the lit/shadow region. The diffracted field is added to the incident/reflected field to result in , and ensure the continuity of fields at ISB/RSB.

Geometrical Theory of Diffraction and Its Uniform Version (cont.)

The transition function

Geometrical Theory of Diffraction and Its Uniform Version (cont.)

The effect of the diffracted field.

Practical Applications of HighFrequency Techniques

Many antennas or structures are electrically large and impractical or computationally inefficient to be solved by numerical methods.

Pattern analysis of large aperture antennas, like pyramidal horn and reflector antennas. Antennas near complicated structures like ships, satellites or buildings.

Far-Field Pattern of Pyramidal Horns

One approach is to integrate the electric and magnetic currents of a rectangular waveguide TE10 mode.

Only for main beam.

UTD can be used to calculate the full E and Hplane patterns

Includes front and back regions of the antenna.

Far-Field Pattern of Pyramidal Horns (cont.)

Use a spherical wave with TE10 mode

In the E-plane,

Total field:

Far-Field Pattern of Pyramidal Horns (cont.)

Observation points:

Region I : GO and two diffracted fields. Region II, IV and VI : Only two diffracted fields. Region III and V : Only one diffracted field.

H-plane pattern can be resolved with similar approach.

Far-Field Pattern of Pyramidal Horns (cont.)

Luneberg-Kline Series

For GO, only the first term is retained. Hence, the wavelength could not be too large. The wave front is locally plane.
For small x, higher order terms are neglected.

Furthermore, the assumption of the ray tube

The curvature could not be too large.

Pattern Analysis of Reflector Antennas

Conventional reflector antennas

Paraboloids - direct-fed systems and as the main reflector in dual-reflector systems. Hyperboloids - subreflector in the Cassegrain systems. Ellipsoids - subreflector in the Gregorian systems. Main beam and the first few sidelobes
PO and aperture integration (AI) (Close to or within reflection caustics)

Comparison between different methods.

Wide-angle and backlobes

GO and UTD

Physical Optics Analysis

Reflector antennas

Electrically large The radiation directions of interest are near and within the specular reflection region.

Incident megnetic fields can be found by ray tracing. High accuracy in the main beam and near sidelobes. Poor accuracy at wide angles.

UTD Analysis
Accurate sidelobes and wide angles and back lobes. Scattered field For parabolic reflector,

Main beam : PO Side lobes at wide angles : Edge diffracted fields only.

UTD Analysis (cont.)

Circularly symmetric parabolic reflector

Diameter: 24 Focal length: 8 PO : UTD : Rest of the patterns.

UTD Analysis (cont.)

Parabolic reflector antenna

PO and AI do not predict correct sidelobes at wide angles and back lobes.
PO is more effective in predicting the pattern where the reflection dominates.

UTD is used to predict accurate values at these regions. UTD has been used to predict full pattern.

Cassegrain and Gregorian dual-reflector systems

Scattering Analysis in a Complicated Structure

Interaction between antennas and their surrounding structures. The only feasible approach

Use UTD to include all possible ray paths. Direct line-of-sight, reflection, diffraction and multiple bounces.

Hybridization of UTD with Numerical Methods

Hybrid integration with conventional numerical methods

MoM, FDTD, FEM Reduce the number of unknown required in MoM. Near the edges,
Regular MoM basis, such as pulse functions, are not valid. UTD-Type of solutions as basis.

Hybrid UTD-MoM

Hybridization of UTD with Numerical Methods (cont.)

Consider a two-dimensional finite PEC strip illuminated by a TM incident plane wave. Conventional MoM may employ pulse functions to represent the induced current as

where the pulse function

is defined as

Hybridization of UTD with Numerical Methods (cont.)

By solving the electric field integral equation, where

: unknown coefficients. : related to excitation or incident fields matrix consisting of mutual impedances.

Hybridization of UTD with Numerical Methods (cont.)

Divide the basis function into two groups.

Interior region
HF solutions are valid. Currents can be represented by the UTD-type of basis.

Exterior region
Regular MoM basis.

In the UTD region,

can be approximated

(a) where are the distances measured from the nth pulse basis to the left and right edges, respectively.

Hybridization of UTD with Numerical Methods (cont.)

The alternative expression

: deduced from (a). N linear equations to be solved. M<<N unknowns. One may randomly select M equations or apply least squares method to solve for the unknowns.

Hybridization of UTD with Numerical Methods (cont.)

Comparison of two difference approaches.

Width of the strip: Regular MoM: 100 basis. Hybrid method: 14 basis
5 pulses in each edge. 4 GTD basis.

Physical Theory of Diffraction

The prediction of PO becomes erroneous away from the specular reflection region.

Diffraction effects are not concerned.

PTD is introduced to correct PO with diffraction contribution

Similar fashion as GTD/UTD correct the GO.

Induced Current aspect (Edge effects) Field aspect (Diffraction fields)

Physical Theory of Diffraction (cont.)

PTD remains valid even in the ray caustic regions of GTD/UTD is the currents are predicted correctly. A simple approach for the current of PTD correction is

Use the results conducted by GTD to calculate the error made by PO. Hence, an equivalent current can be generated. Equivalent current method (ECM)

Future Developments
Most techniques based on GTD/UTD. Other GTD/UTD type of solutions

Typical diffraction mechanisms in radiation Mutual coupling Problems as well as their time domain versions Uniform asymptotic theory (UAT) Spectral theory of diffraction (STD) Incremental theory of diffraction (ITD)

Other solutions

Reference
G. A. Deschamps, Ray Techniques in Electromagnetics, Proc. IEEE, vol.60, pp.1022-1035, 1972. R. Kouyoumjian and P. Pathak, A uniform geometrical theory of diffraction for an edge of a perfectly conducting surface, Proc. IEEE, vol.62, pp.1448-1461, 1974. D. A. McNamara, C. W. I. Pistorius and J. A. G. Malherbe, Introduction to the uniform geometrical theory of diffraction, Artech House, 1990.