GB2603160A - Flat panel leaky-wave array antenna with 2D scanning - Google Patents

Abstract

A square, multiple-feed-port, leaky-wave, flat panel resonant cavity array antenna 1 comprises a low loss dielectric substrate 5; a porous dual-polarised periodic wall 2; and a reconfigurable high impedance surface 3; where the antenna 1 is configured to provide independent and simultaneous excitation of forward and backward TE10 and TE01 modes in the cavity to produce a two-dimensional (2D) electronic scanning of a linear or circularly polarised radiation pencil beam. The periodic wall 2 may comprise a double layer periodic surface of orthogonal directed electrically isolated inductive conducting tracks or wires forming a grid for beam shape optimisation. The periodic wall 2 may comprise an uni-polarised single layer of conducting tracks 4 which are diagonal relative to the sides of the square cavity. The antenna 1 may have a non-planar conformal surface or two or more flat panels with coordinated beam scanning. The multiple ports may comprise corner ports 6, 7 or cavity side ports and a central port 8. The antenna 1 may include a high impedance surface with voltage-controlled varactor diodes and the dielectric substrate may be a metamaterial. The feed ports may take various forms. The antenna may provide a low-cost electronic tracking system.

Description

Flat-panel Leaky-wave Array Antenna with 2D Electronic Scanning

Description

A novel flat panel array antenna is described which is based on leaky-wave electromagnetic radiation from a parallel plate waveguide cavity (Fabry-Perot cavity). One of its two electromagnetic wave trapping walls, or plates, which are normally highly conducting, is replaced by a partially reflective surface. The second normally reflecting wall is replaced by a reconfigurable high impedance surface to provide frequency independent scanning.

By exciting the square cavity from a corner port, the orthogonal modes in the cavity (TEio and TED) are set up. If the partially reflective surface is then represented by a diagonal grid of periodically spaced conductive tracks, a linearly polarised pencil beam is generated with the orthogonal modes acting in concert. The beam can be scanned from broadside toward endfire in the diagonal plane of the antenna by means of impedance changes at the reconfigurable HIS. A second port located in the opposite corner of the antenna can provide scanning from broadside toward endfire in the opposite direction.

A partially reflective surface furnished with a pair of closely spaced orthogonally polarised conducting grids al so supports the propagation of simultaneous TE10 and TE01 modes within the cavity but in this case the modes act independently. Depending on which of four sidewall ports (one on each wall) is excited on its own, linearly polarised beams can be formed and scanned in the principal planes of the antenna. If both modes are excited simultaneously from a pair of ports on adjacent sidewalls, a scanned linearly polarised pencil beam in the diagonal plane is enabled, when the ports are fed with equal amplitude and in phase signals. In addition if the port signals are in quadrature phase a circularly polarised beam is procured also scanning in the diagonal plane.

The reconfigurable high impedance surface typically comprises a periodic array of microstrip patches laced with tuning varactor diodes. Voltages applied to the diodes produce capacitance changes which alter the patch resonances and hence the impedance or reflection properties of the surface. This in turn enables electronic beam scanning.

Two dimensional electronic scanning of a pencil beam is a technology usually restricted to phased array antennas. However, for low cost flat panel antennas, the phased array format is heavily over-engineered, resulting in unacceptably high cost levels. The flat panel leaky-wave array antenna outlined here provides a solution which is potentially more acceptable in this market.

Prior Art

In an original version of a flat panel leaky-wave antenna (fig. 1), a 'pencil' shaped microwave beam was made to radiate at a fixed angle to 'broadside' in a principal far-field plane of the antenna (x-z plane in fig. 1). Beam scanning is possible for the antenna shown in fig. 1, but only by sweeping the frequency of the input signal. The main features of this 'fixed beam' array, which are relevant to the proposed invention, are: a) The antenna is founded on an parallel plate electromagnetic waveguide cavity resonator (A) supporting the fundamental TEu) mode (D), or the orthogonal TE,01 mode b) This basic mode resonates between the ground plane (C) at x=0 and the porous surface at x-a, with the electric field (E) directed along y and the magnetic field (H) directed along x and z, when the guide is end fed at z-O, or at z-L (as shown in fig. 1), thus procuring leaky-wave operation.

c) For leaky-wave radiation from the porous wall, it usually takes the form of an open periodic surface (F) which may take several possible forms. Here a closely, and periodically, spaced parallel wire grid is depicted.

d) For ay-directed grid, leaky-wave behaviour requires excitation on an x-y face, either at z=0 or at z=L (7). For this grid orientation the TE10 mode is formed in the cavity with an E-field (E) in the y -direction resulting in a radiated beam at a fixed inclination from broadside in the x-z plane. The beam is linearly polarised ( y -polarised).

e) For a square cavity (L x L) an inclined beam located in the x-y plane can also be produced by a feed port at y=0, or at y=L, provided that the periodic grid wires are z-directed. In this case the TEN mode is formed in the cavity antenna with the inclined pencil beam now located in the x-y plane. The beam is now 14-polarised.

The cavity resonator antenna, as depicted in fig. 1, is typically designed to locate a 'pencil' beam at broadside (x-direction). In the x-z plane the linearly polarised radiation pattern is shaped by the inductive surface created by the wire grid (F) which is leaky to the z-propagating quasi-TEu) mode. In the y-z plane the radiation pattern is shaped by the feed (G) whose field pattern in this plane as it enters the cavity (A) is minimally disturbed by the grid (F). In the y-direction the wires are invariant and in explanatory theoretical treatments are generally assumed to be unterminated. At the cut-off frequency of the parallel plate waveguide (A) the pattern maximum in the x-z plane is at broadside, so the pencil beam is located there. At frequencies above cut-off the maximum moves away from broadside toward endfire in a manner well recognised by antenna practioners, but remains at broadside in the y-z plane. It is pertinent to note, for the set up in fig. 1, that if fed through the y-0 (or the y-/) surface of the cavity the TE mode is not formed since the grid presents an open-surface to the EIVI fields.

While the fundamental form of this planar leaky-wave array antenna is limited to a fixed beam operational mode once constructed, the beam can, however, be positioned anywhere between broadside and end-fire, at any in-band frequency, by altering the cut-off frequency of the TE mode within the parallel plate waveguide cavity. The operating principle is summarised in fig. 2 which displays the TE modal behaviour in terms of the relationship between frequency (co) and the TE mode propagation phase shift (A). The elevation angle of the leaky-wave beam (0) is related to A, through the equation: - 1.1=3 -sin' Ag 10 Consequently, the use of cavity height (h) to position the beam now becomes apparent (see fig. 2) given that the nth TE resonance frequency occurs where con= ?cc/N.

Technical Description of Invention

The proposal described in this patent application, details two novel planar Fabry-Perot cavity resonator array antennas, one of which is linearly polarised while the other is circularly polarised. Both aim to provide directional beam scanning of a radiating 'pencil' beam. Scanning, at a fixed frequency, of an antenna not unlike that depicted in fig. 1, is evidently feasible if the ground plane movement which is fundamental to beam positioning can be simulated by electronic means. But for electronic scanning of a 'pencil' beam formed by the leaky-wave array on its own, the E-plane horn feed in fig. 1 must be eliminated and replaced by one or more simple, localised feed arrangements. To reach this goal it is mandatory that, in both versions of the proposed array antenna, the TElo and TE01 modes in the cavity are excited simultaneously. Independent excitation of these modes leads to the elimination of ID restrictions.

ID electronically scanned, linearly polarised, leaky-wave antenna A scanned linearly polarised 'pencil' beam can be generated from a square and layered flat-panel leaky-wave array antenna as suggested in fig. 3. This solution is achieved by taking the novel step of feeding electromagnetic power into a corner of the square cavity (1), between the partially reflective surface (2) and the impedance surface (3), which in fixed beam manifestations is usually a simple conducting surface. Thus both the z-propagating TEm and the y-propagating TEnt modes in the cavity are excited equally and simultaneously. Both must contribute to the leaky-wave process, in order to generate a pencil beam. To procure this result the partially reflective surface (2) requires the dielectric substrate supporting the periodic grid of conducting wires (4) to be orientated in a direction which is normal to the primary direction of propagation of the combined cavity modes as shown in fig. 3. That is, from the corner port (6) (or (7)) to the opposite corner. It should be noted that the leaky-wave surface need not necessarily be formed from straight conducting wires but could be evolved from a similarly polarised periodic surface of metallic dipoles deposited on a dielectric substrate, or of slots in a metallic screen. The Fabry-Perot cavity (1) is excited from a microstrip substrate (5) positioned between the cavity and the ground or impedance surface (3). It is depicted as supporting coaxial line fed patch couplers (6) and (7) located at opposite corners of the substrate. The centrally located patch feed (8) provides signal continuity through broadside in a scanning antenna. Coaxial line fed loop couplers, probe couplers, rectangular waveguide and coplanar waveguide inputs are among other port possibilities.

Electronic scanning of the primary beam of a leaky-wave antenna array can be accomplished by introducing a ground plane modification which is generally referred to as a reconfigurable high impedance surface (HIS). The essence of it is depicted in fig. 3. Conventionally, such a surface (3) can be fabricated on a plane substrate copper clad on one face, and positioned above a conducting ground plane (9). The upper copper surface is configured into a two dimensional periodic array of square patches (10). In addition, the periodic surface is furnished with varactor diode inserts (11) positioned between each and every patch. Selected patches are connected, in a regular and predetermined fashion, both to the ground plane and to drive circuitry below the ground plane. Voltage induced capacitance changes at the diodes result in surface impedance increments or decrements at the array face. By invoking transmission line theory it can be established that these impedance changes replicate ground plane movement within the cavity, and thus contribute to main beam scanning. In the arrangement shown in fig. 3 the simulated 'ground plane movement' influences the TElo and the TE01 modes simultaneously, and to a similar extent. The beam, therefore, as described above, scans in a rotated 45° plane if excited by corner ports (6) or (7). Satisfactory simulation results at K-band have been achieved by employing modelled examples of commercially available varactor diodes, capable of voltage excursions in the 0 -30V range.

Satellite acquisition and tracking throughout the horizon-to-horizon scan, in a LEO application, can be realised by deriving correction signals from the split-feed arrangements at (6) or (7). Normally the signals received by the array enter each half-port in-phase and these signals are combined and delivered to the communications section of the tx/rx system. But additionally, difference signals, which are representative of beam-to-satellite misalignment, can be assessed by directing the antenna r.f, returns entering the half-ports to a conventional hybrid coupler and thence to a beam control system. In radar terms this is essentially monopulse operation. The control system then furnishes the motor (12) with drive voltages which are computed to minimise the difference signals and hence the misalignment.

Electronically scanned fiat-panel leaky-wave antenna with 21) scanning While a flat-panel antenna, such as that depicted in fig. 3, is limited to scanning in a fixed 45° plane it still exhibits technical features which will attract many potential applications. However, the possibility of removing this restriction would evidently improve the versatility of the antenna, particularly if motor controlled positioning can be eliminated. The route to this goal is hinted at in fig. 4. To incorporate circularly polarised radiation into a scanned leaky-wave antenna it is necessary to devise a dual-grid arrangement in which the orthogonal components of the grid are located at the same height above the quasi-ground surface, but remain electrically isolated from each other. Modelling studies indicate that the obvious stratagem of adopting straight wires orthogonally aligned on the opposite faces of a thin substrate produces severe modal mismatches. The novel solution which has been developed entails interleaving of orthogonally aligned wires (3) as illustrated pictorially in fig. 4. The proposed dual-mode dual-polarised leaky-wave array antenna which is presented in layered, or exploded, form in fig. 4, comprises many of the primary elements already described for the linearly-polarised version (fig. 3).

The y-directed and x-directed TEto and TEot propagate within the cavity (1), simultaneously and independently guided by the z-directed and y-directed conducting wires (4) threading through the substrate (2). The cavity (1) lies below the substrate (2) and above the microstrip board (5) supporting the patch feed sytem (6), (7), (8), (14) and (15). Electronic scanning is implemented by replacing the traditional metal ground plane by a reconfigurable high impedance surface (3). This surface comprises a 2D array of square metal patches (10) positioned at a predetermined height above a backing ground plane (9). The patch sizes, their periodicity and spacing, and the permittivity of the substrate are dictated by the requirement for the antenna to operate within K-band. Operation in other bands is not precluded. The array of resonant patches is laced with voltage controlled varactor diodes (1 1). Alternate patches in every row or column is connected to the ground plane (9) while the remainder are connected to a voltage source. Alterations in the voltage drive levels change the resonant characteristics of the patches, and hence the reactive impedance of the surface.

The added images (16) and (17) in fig. 4 depict practical interleaved dual-polarised grids which have been modelled on a full-wave electromagnetic field simulation package, namely HFSS. On the double sided microstrip substrate (2), the copper cladding on both sides is etched to form orthogonal dipole patterns (16). These dipoles are selectively interconnected to top and lower surface dipoles, by means of through-hole metal plated vias, to form the sinuous conducting tracks. Fabrication by 3D printing techniques has also been investigated and shown to be possible. This alternative structure is illustrated in (17). Here the dipoles and the connecting 'risers' are created by 3D printing to form the sinuous conducting tracks.

The antenna arrangement depicted in fig. 4 can be excited in several ways through the five ports (6,7, 14, 15, 8) as dictated by port switching options incorporated into the microwave electronics located below the HIS ground plane. With port (6) and (7) excited, a linearly y-polarised pencil beam is formed in the x-z plane, in the forward (left) quadrant for port (6) and the backward (right) quadrant for port (7). Similar behaviour in the x-y plane occurs for a z-polarised beam if ports (14) and (15) are operated in a similar manner. Orthogonality of the respective grids ensures scanning independence in these orthogonal planes. As a consequence it becomes uniquely possible, in a Fabry-Perot cavity based leaky-wave array antenna, to demonstrate controlled 2-D scanning in perpendicular planes. For circularly polarised communications the signal from the Rx/Tx is channelled through a 3dB hybrid coupler which directs equal signal components, in quadrature phase, to ports (6) and (14) (or to ports (7) and (15). Circularly polarised transmission is thus secured in both the forward direction, or the backward direction, in the diagonal plane of the antenna.

On reception, the 3dB hybrid coupler directs the 'sum' component of the of the signals at port (6) and port (14), (or port (7) and port (15)), in fig. 4, to the Tx/Rx. Simultaneously, 'difference' components in the received signal, can be directed, as described above, to an electronics controller to procure 'monpulse' tracking of a moving LEO satellite. Correction signals from the control system are subsequently directed to all four ports to modify relative power levels between (6)1(14) or (7)1(15) while maintaining total radiated power. Electromagnetic simulation of this antenna format indicates that small power adjustments are enough to correct the position of a slightly misaligned beam. In the same manner as described for the antenna option in fig. 3, port (8) provides a broadside beam, with very limited scanning capability, to maintain signal strength as the primary scanning beam passes through broadside.

Full-wave electromagnetic field simulations confirming electronic scanning For a high frequency structure simulator (HFSS) model with the following defining parameters: cavity resonator area=210mm x 210mm, resonator height=6.6mm, grid pitch=3mm, grid gap=2.1mm, grid wire width=0.9mm, resonator relative permittivity cr= 2.2, fig. 5 provides instantaneous computed E-field magnitude levels of typical modal field distributions within the cavity, with the resultant free-space radiating beam superimposed.

In image (a) the flat panel cavity antenna is excited, at the left corner, from a dual-patch source (see fig. 3 -(6)/(7)), with in-phase input signals. The cavity field pattern (19), which comprises the TEio and TE01 modes operating in unison, depicts travelling waves within the cavity, spreading with circular wavefronts as they travel away from the source. Controlled power leakage through the porous grid results in the 23dB gain pencil beam located at 48° from broadside in the diagonal plane (45° plane) of the antenna. When the input signal is switch to the port in the opposite corner of the flat-panel array, in image (b), the modal pattern (22) mirrors the pattern (19) while the 23dB gain beam is now located at -48° from broadside in the diagonal plane. In theory (see fig. 2 and the accompanying equations) these leaky-wave beams can be located anywhere between -90° and +90° in the diagonal plane by changing the cavity height from -5.6mm to -8.2mm for an antenna operating at 17GHz.

In practice the achievable range is nearer 0° to +/-700 with surface wave effects intruding above 70°. At beam angles close to broadside TE modes in the cavity are operating dose to cut-off for the low height parallel plate waveguide. Electromagnetic wave coupling into the cavity from the corner ports becomes problematic because of high input match levels for 0I<-5'). This effect is overcome by introducing a patch port at the centre of the array. Waveguide cut-off does not interfere with coupling to the TE modes for a central port as is shown in image (c). The image highlights the cut-off TE mode pattern (24) which decays in strength radially with a very long guide wavelength. The resultant radiation beam (23) is located within 5° of broadside. A beam at broadside is produced if the cavity height is reduced to -5.55mm.

In fig. 5, image (d), the split beam requirements for a satellite tracking role is depicted. This beam (26), which is positioned at 50° from broadside in the diagonal plane of the antenna, has been generated by exciting the separate halves of the split corner port with anti-phase signals. The dividing effect on the cavity mode pattern (25) is also clearly shown.

An HESS model of a flat-panel cavity resonator antenna furnished with a dual-polarised leaky grid has also be studied. The model is based on the following defining parameters: cavity resonator area=150mm x 150mm, resonator height=6.5mm, grid pitch=2.2mm, grid gap=1.8mm, grid wire width=0.4mm, grid substrate thickness=0.33mm, resonator relative permittivity 6,-2.2. The computed results are presented in fig. 6. These, again, provide simulated E-field distributions within the cavity at an arbitrary instant in time with the predicted free-space radiating beam superimposed on the cavity patterns. Three cases are shown. In a the flat panel antenna with the above dimensions is excited from the right hand corner port. A uniformly spreading TE pattern (28) in the cavity is depicted which is formed from independent TElo and TEol cavity modes. Electromagnetic leakage through the porous dual-polarised grid, results in a pencil beam (27) with -22dB gain, located at 50° from broadside in the diagonal plane of the square array. When power is fed to the same antenna from the left hand corner (b) an almost identical TE pattern (29) spreading in the opposite direction is observed which produces a pencil beam (30) again exhibiting -22dB gain, but directed at -50° from broadside in the diagonal plane. In a beam scanning context switching ports, when the beam approaches broadside, permits a scanning range from approximately -70° to +70° in the diagonal plane. Close to and at broadside the corner ports encounter large mismatches with the planar waveguide operating near cut-off. To circumvent this phenomenon, the cavity is excited (c) at a centrally located port (8) (see fig. 4). The resultant cavity TE mode pattern at cut-off (31) is concentrated around the central port in an almost circular fashion with a single wavefront. The modal fields decay very rapidly in the radial direction away from the source. The resultant well defined pencil beam (32) exhibits an essentially circular cross-section, which equates with a directive gain of approximately 25dB.

Electronic scanning as explained with reference to fig. 3 and fig. 4 that electronic beam scanning is achieved by altering the surface impedance of a planar periodic array of patch resonators (10) interconnected by means of voltage controlled varactor diodes (11). This procedure has been simulated on HF SS using the model which generated the images shown in fig. 5. To do this the perfectly conducting ground plane employed in the model is replaced by an adjustable impedance surface (Zs=jX 0/sq). The effect on beam positioning in the diagonal elevation plane of the antenna is shown in fig. 7. The six circular diagrams presented in fig. 7 are typical two dimensional antenna radiation patterns showing directivity and elevation angle for the principal beam at a fixed frequency. For a computation at 17GHz with the resonant cavity height set to 5.9mm the radiated beam is positioned at 200 from broadside when the impedance surface is defined as perfectly conducting (Zs=j0 0/sq). Changes in Z, from j0 0/sq through j50 0/sq, j100 0/sq, j200 0/sq to j5000 0/sq, moves the radiated beam from 20° through 35°, 400, 46° to 55°. Throughout the exercise the operating frequency is held at 17GHz while the cavity height is maintained at 5.9mm. The open circuit (0.C.) case depicted in the sixth image in Fig. 8 applies to a model with the cavity height set equal to 6.6mm. The resultant beam angle measured from broadside at 74° is more than is available at a cavity height of 5.9mm (--60°). Not surprisingly cavity height strongly influences the beam scanning range. Neverless, the full-wave electromagnetic simulation of the flat panel leaky wave antenna array demonstrates the necessary link between beam scan angle and tunability of an impedance surface located at the conventional cavity ground position, thus justifying the claim that 2D electronic scanning is achievable with the proposed antenna.

Scanning options The operating mechanism, in the context of a low earth orbit communications system, is summarised in Fig. 9 with the five port, electronically scanned, leaky-wave antenna (2) at the centre of the diagram, which also depicts the possible modes of operation using broad beam direction arrows (33, 34, 35). For example ports (6) and (7) provide linear y-polarised beams (33) to the right and left. Excitation of port 6 produces a y-polarised beam (33) scanning to the right from broadside to endfire (indicated by large arrow direction), while port 7 produces a y-polarised beam scanning from the left, again from endfire to broadside in the x-z plane. With port switching at broadside implemented, a pencil beam can be scanned anywhere within the x-z plane controlled by the reconfigurable HIS (see (3) in fig. 4). Also, loss of power at switch-over can be minimised by communicating through the central port (8) which sets up a broadside beam. A similar set of operations occur for excitation of ports (14) and (15) but with scanning now in the y-z plane (34) with x-polarised beams.

Operation with circular polarisation is represented in fig. 9 by the large arrow (35) aligned at 450 to the principal planes and along the satellite track (36). The lower inward pointing portion of this arrow depicts beam locations with ports (7) and (15) equally excited and in phase-quadrature, while the outward arrow represents beam locations associated with simultaneous quadrature excitation of ports (6) and (14). Polarisation possibilities for scanning in the diagonal plane (35) are suggested by the symbols (38). As a consequence, with a reconfigurable HIS, it becomes uniquely possible, in a Fabry-Perot cavity based leaky-wave array antenna, to demonstrate controlled 2-D scanning (as illustrated by the curved arrow (37)in fig. 9), into both the forward and the backward half-plane of the leaky-wave array, if port signal magnitude switching between pairs of ports in a four port system is implemented.

Claims (16)

1. Claims 1. A square multiport leaky-wave flat panel array antenna, based on a planar Fabry-Perot resonant cavity, formed from a low loss, dielectric substrate, and comprising a porous dual-polarised periodic wall enabling independent and simultaneous excitation of forward and backward quasi-TER) and TEN modes in the cavity, while the conventional ground plane is replaced with a reconfigurable high impedance surface, enabling 2D electronic scanning of a radiating pencil beam, which may be linearly polarised or circularly polarised, within a -90° to +900 range of elevation angles.
2. 2. An antenna according to claim 1 in which the periodic wall comprises a double layer periodic surface of orthogonally directed and electrically isolated inductive conducting tracks which are straight and closely spaced.
3. 3. An antenna according to claim I in which the periodic wall comprises a double layer periodic surface of orthogonally directed and electrically isolated inductive conducting wires which are sinuous and interleaved.
4. 4. An antenna according to claim 1 in which the periodic wall comprises a double layer periodic surface of orthogonally directed and electrically isolated inductive conducting wires with prescribed track spacings for beam shape optimisation.
5. 5. An antenna according to claim I with a periodic wall which comprises an uni-polarised single layer of diagonally (relative to the sides of the square cavity) orientated conducting tracks.
6. 6. An antenna according to claim 1 with a non-planar conformal surface with enhanced scanning range or two or more flat panels with coordinated beam scanning.
7. 7. An antenna according to claim 1 and claim 4 wherein the multiports are corner ports.
8. 8. An antenna according to claim 1, claim 4 and claim 5 wherein the corner ports comprise split ports at two opposing corners, hence allowing switching between split-beam or full-beam operation of the scanning pencil beam.
9. 9. An antenna according to claim 1 with the cavity ground plane replaced by a reconfigurable high impedance surface comprising a periodic array of patch resonators populated with resonance modifying varactor diodes.
10. An antenna according to claim 1 with four sidewall positioned input ports enabling in-phase or quadrature-phase excitation of the TEio and 1E01 modes in the cavity, by applying simultaneous signal excitations at orthogonal ports.
11. 11 An antenna according to claim 1 with four sidewall positioned input ports wherein the power levels applied to othogonal ports are independently varied to achieve azimuth positioning of a linearly or circularly polarised principal beam.
12. 12 An antenna according to claim 1 with a centrally located port which supports a broadside pencil beam thus enabling scanning through broadside with minimal loss of beam power.
13. 13 An antenna according to claim 1 and claim 6 wherein split corner ports combined with a motorised positioner facilitate satellite tracking and acquisition.
14. 14 An antenna according to claim 1 and claim 7 wherein a reconfigurable high impedance surface is formed from a periodic arrangement of dual-polarised resonant elements (symmetrically shaped polygonal patches, symmetrically located dipoles, symmetrically located slots) fabricated on a copper coated dielectric substrate.
15. 15. An antenna according to claim 1, claim 4 and claim 5 wherein the dielectric substrate forming the cavity is replaced by a metamaterial.
16. 16. An antenna according to claim 1 and claim 6 wherein the ports may also be realised in waveguide, coaxial line fed probes, coaxial line fed loops, coaxial line fed resonant patches, stripline fed probes/loops, etc.