Field of the Invention

This invention relates to an antenna and, more especially, this invention relates to a surface array antenna.

Description of Prior Art

It is well known that a two-dimensional array or surface of selectable feeds can be used in isolation or in conjunction with a microwave-optic device such for example as a lens or reflector, in order to generate a beam pointing in a desired azimuth and elevation direction. The use of parallel plate antenna structures to guide an electromagnetic wave is also well known. Furthermore, as disclosed in USA Patent No. 7,271 ,683, directed selected beam antennas are able to be produced by the use of controllable reflective structures between the parallel plates of a parallel plate antenna structure, in conjunction with electronically controlled switched reflective devices positioned between the parallel plates. The electrically controlled switched reflective devices may be, for example, a reflector composed of plasma PiN diodes or micro-actuators. The use of parallel plate structures to distribute RF power to a number of output ports is also well known. In most cases these ports are arranged around the edge of the waveguide. If they are arranged in the body of the waveguide, their presence disrupts the propagation of the RF in the parallel plate waveguide, hence limiting both their location and number that can be placed in the waveguide.

Brief Description of the Invention

It is an aim of the present invention to simplify and extend the application of the above mentioned prior art antenna designs.

Accordingly, in one non-limiting embodiment of the present invention there is provided a surface array antenna comprising:

(a) a launch structure comprising a single feed or a pair of feeds;

(b) a parallel plate structure comprising a single parallel plate cavity or a pair of parallel plate cavities;

(c) an array of reconfigurable reflectors and an array of selectable field probes or a pair of arrays of reconfigurable reflectors and a pair of arrays of selectable field probes located in the parallel plate structure, where the array of reconfigurable reflectors or the pair of arrays of reconfigurable reflectors, when configured, is able to focus on a sub-group of selectable feeds or a pair of sub-groups of selectable feeds, and where the selectable field probes are such that they can be reconfigured as transparent to the wave propagating in the parallel plate waveguide, or coupled to the wave in the parallel plate waveguide; (d) a transition structure that contains a pattern of printed tracks, radio frequency cables or waveguides or two interlaced patterns of the tracks or waveguide that magnifies the pattern of selectable field probes or pair of patterns of selectable field probes; and

(e) an array of single or dual polar launch elements that are positioned at or close to the focal surface of a lens or reflector, or lie on a curved surface; and the surface array antenna being such that:

(f) the launch structure launches an electromagnetic wave or a pair of electromagnetic waves into the parallel plate cavity or the pair of parallel plate cavities in the parallel plate structure;

(g) the parallel plate structure contains the array of reconfigurable reflectors and the array of selectable field probes or the pair of arrays of reconfigurable reflectors and the pair of arrays of selectable field probes;

(h) the parallel plate structure has a first layer which contains the reconfigurable reflector which, when configured, is able to focus on the sub-group of selectable feeds or the pair of sub-groups of selectable feeds, which are positioned in either the first layer of the parallel plate structure or a second layer of the parallel plate structure, linked by a U-turn from the first layer;

(i) the selected feeds transit to the transition structure;

(j) the transition structure transits to the array of single or dual polar launch elements;

(k) the array of single or dual polar launch elements produces a selected beam directed according to a sub-group of selected field probes; and

(I) the transition of the selected feeds to the transition structure is such that the selectable field probes are able to be selected individually or collectively in order to launch via the transition structure through either single polarisation or dual polarisation antenna arrangements directly into free space or into a lens or reflector means and thereby optically perform electronic scanning in two orthogonal directions.

The surface array antenna of the present invention simplifies and extends the application of the prior art antenna designs by allowing the use of parallel plate structure to select feeds from the fixed feed and to steer the resulting beam in two orthogonally independent directions. The fixed feed may be a two-dimensional array or surface of fixed feeds. The two orthogonally independent directions may be, for example azimuth and elevation. The focal surface array antenna may allow for a single stage selection among many feeds and thereby to select from very large two- dimensional arrays of beams. This is unlike more conventional switch matricies that require many layers of switching when the number of selectable beams is large, for example typically more than four. The single stage selection minimises losses, complexity and associated implementation costs, which are all important for facilitating the commercialisation of the focal surface array antenna. In many applications, a very low insertion loss also obviates the need for separate amplification directly behind each feed.

The surface array antenna may be one in which the reconfigurable reflector is a plasma device, and in which the reflecting elements are PiN diodes which when excited form plasma between the parallel plates, with the plasma reflecting the electromagnetic wave.

Alternatively, the surface array antenna may be one in which the reconfigurable reflector is a radio frequency MEMS device, in which the reflecting elements are radio frequency MEMS reflecting elements, and in which when actuated the radio frequency MEMS reflecting elements form localised radio frequency shorts between the parallel plates, with the radio frequency shorts reflecting the electromagnetic wave.

The surface array antenna may be one in which the selectable field probes are excited plasma devices which are located within the parallel plate structure, and which are able to selectively transition a proportion of the travelling electromagnetic wave into a surface track or waveguide which is located outside the parallel plate structure. Alternatively, the surface array antenna may be one in which the selectable field probes are actuated radio frequency MEMS devices which are located within the parallel plate structure, and which are able to selectively transition a proportion of the travelling electromagnetic wave into a surface track or waveguide which is located outside the parallel plate structure.

The surface array antenna may be one in which the reconfigurable reflectors, the feeds and the selectable field probes are positioned individually or collectively on conducting pillars within the parallel plate structure, and in which the conducting pillars have parallel or sloping sides to form circular conducting frusta which electrically connect via their parallel faces to parallel plates and individual terminals of the reconfigurable reflectors, feeds or selectable field probes respectively.

The surface array antenna may be one in which the feed is in the form of a corporately connected sub-array composed of reconfigurable electromagnetic feeds and optional fixed feeds which collectively facilitates the generation of an electromagnetic wave between the parallel plates, which is controllable in intensity and phase and which can be tuned in frequency according to the spacing, spatial pattern, intensity and phase weighing of the sub-array's elements to produce a directed wave between the parallel plates of either omnidirectional or selectively directed form.

The surface array antenna may be one in which the selectable field probe is in the form of a sub-array composed of reconfigurable and optional fixed field probes which facilitates the generation of a set of driving signals with intensity/phase characteristics and which are tuneable in frequency according to the spacing, spatial pattern, intensity and phase weighting of the sub-array's elements that feed individual or groups of antenna elements with collective free-space radiation patterns that are steerabie and depend on the direction and form of the electromagnetic wave between the parallel plate, any dielectric between the parallel plate, and any external tuning applied to the driving signals.

The surface array antenna may be one in which the transition board comprises a set of tracks or guides that connect to the printed launch elements.

The surface array antenna may be one in which the printed launch elements are single polarisation printed launch elements.

Alternatively, the surface array antenna may be one in which the printed launch elements are dual polarisation printed launch elements, in which there are two of the fixed feeds and two of the lens or reflector means, and in which the two fixed feeds and the two lens or reflector means are used independently to feed and control two ports of the dual polarisation launch elements.

The surface array antenna may be one in which the lens and/or reflector means is used to point and improve the collimation of an electromagnetic wave emanating from a selection of one or more of the printed launch elements of the surface array, and thereby to produce a beam directed in the two orthogonal directions according to the positions and orientations of the selected printed launch elements within the surface array of printed launch elements. The two orthogonal directions may be azimuth and elevation. In this case, the focal surface array antenna may be one in which the lens or reflector means is an off-set reflector. Alternatively, the lens or reflector means may be a Luneberg lens. Alternatively, the lens or reflector means may be a multi-faceted configuration.

Brief Description of the Drawings

Embodiments of the invention will now be described solely by way of example and with reference to the accompanying drawings in which:

Figure 1a shows a single layer focal plane array (ray trace);

Figure 1 b shows a dual layer focal plane array (ray trace);

Figure 2 shows a reflective electromagnetic switch;

Figure 3 shows an active electromagnetic feed;

Figure 4 shows a reconfigurable electromagnetic field probe;

Figure 4A(a) shows a reflective reconfigurable element positioned on a cylindrical pillar within a parallel plate structure;

Figure 4A(b) shows a reconfigurable electromagnetic feed positioned on a cylindrical pillar within a parallel plate structure;

Figure 4A(c) shows a reconfigurable field probe positioned on a cylindrical pillar, with illustrated sloping sides, within a parallel plate structure;

Figure 4B(a) shows in cross-section a corporately connected sub-array composed of reconfigurable electromagnetic feed elements within a parallel plate structure; Figure 4B(b) shows in cross-section a sub-array composed of reconfigurable field probe elements with a parallel plate structure, feeding a set of free-space antenna elements;

Figure 4C(a) shows in plan form a corporately connected sub-array composed of four active reconfigurable electromagnetic feed elements within a parallel plate structure radiating an approximately omnidirectional electromagnetic wave;

Figure 4C(b) shows in plan form a corporately connected sub-array composed of four active reconfigurable field probe elements within a parallel plate structure feeding a set of four antenna elements;

Figure 4C(c) shows in plan form a corporately connected sub-array composed of three active configurable electromagnetic feed elements within a parallel plate structure radiating a directed electromagnetic wave;

Figure 4(d) shows in plan form a corporately connected sub-array composed of four reconfigurable field probe elements within a parallel plate structure with a single active reconfigurable field probe feeding a large circular patch antenna element;

Figure 5 shows the combined selectable electromagnetic feed and electromagnetic reflector (common layer);

Figure 6 shows a single parallel plate feed to a field probe reflected off an electromagnetic reflector;

Figure 7 shows a dual parallel plate feed with a U-turn transition to a field probe;

Figure 8 shows a centre fed two layer parallel plate structure; Figure 9 shows hyperbolic reconfigurable reflectors of different eccentricity with a U-turn boundary (note different focal point);

Figure 10 shows an elliptical reconfigurable reflector with a U-turn boundary (note different focal point);

Figure 11 shows a selected reconfigurable reflector, with the left hand part of Figure 1 showing double layers of electromagnetic switches, and the right hand part of Figure 1 showing radial control tracks;

Figure 12 shows an electromagnetic field between parallel plates for a selected reflector, with the left hand part of Figure 12 being in the off condition and the right hand part of the Figure 12 being in the on condition;

Figure 13 shows a radial or Cartesian array;

Figure 14 shows one out of two radio frequency feed or electromagnetic field probe array;

Figure 15 illustrates port selection for all-ports (omni-case) and two ports;

Figure 16 shows elements selection for all-elements (omni-case) and two elements;

Figure 17 shows radial and Cartesian twenty five element selection; Figure 18 shows offset reflective configuration (centre beam position illustrated);

Figure 19 is a side view of a single layer focal plane antenna;

Figure 20 is a plan view of a focal plain antenna and highlights three radial transmission; Figure 21 shows layer 1 of a parallel plate centre feed, reconfigurable reflector;

Figure 22 shows the layer 1 parallel plate electromagnetic field probes, and transition to layer 2;

Figure 23 shows the layer 2 with transmission lines and with impedance transmission to printed patches (layer 3);

Figure 24 shows the layer 3 focal plane of element patches;

Figure 25 is a side view of a dual layer focal plain antenna;

Figure 26 is a plan view of a focal plane antenna, and shows radial transmission lines;

Figure 27 shows layer 1 of a parallel plate centre feed reconfigurable reflector, and shows a U-turn transition to layer 2;

Figure 28 shows layer 2 parallel plate electromagnetic field probes, and a transition to layer 3;

Figure 28A shows a radial track board : field expander;

Figure 28B shows a radial track board : field concentrator;

Figure 28C shows a radial tack board and illustrates amplification and no amplification;

Figure 29 shows a spherical segment array using a spherical lens to provide an wide field of view;

Figure 30 show the surface fed low dielectric Luneberg lens to provide a wide field of view;

Figure 31 shows Luneberg lenses providing 360° planar coverage with + 60° in the orthogonal plane (requires three front ends); Figure 32 shows four Luneberg lenses providing full 4π steradians spherical coverage (requires four front ends);

Figure 33 is a side view of a single layer dual polar focal plane antenna;

Figure 34 is a plan view of a dual polar focal plane antenna, and highlights three radial transmission lines;

Figure 35 illustrates a planar array of feeds;

Figure 36 shows two examples of multi-faceted arrays of selectable feeds to extend the field of view of a spherical lens;

Figure 37 shows a flexible radio frequency printed circuit board folded at vertices to produce facetted arrays of selectable feeds;

Figure 38 shows a small cell backhaul;

Figure 39 shows a mid-range 60GHz access point;

Figure 40 illustrates a circular distribution network; and

Figure 41 shows how the circular distribution network of Figure 40 may be used to generate the beam selection and beam steering phase and amplitude signals for a non-planar array.

Description of Preferred Embodiments

In the following description, the reconfigurable elements are bidirectional and fully reciprocal, so although the field probe may be viewed as sampling the local electric field on transmit, it may also be viewed as establishing the local electric field on receive. Similarly, although the electromagnetic feed may be viewed as establishing the local electric field on transmit, it may also be viewed as sampling the local electric field on receive. The following description and terms used to describe the surface array antenna and associated reconfigurable elements are based on the transmit case.

Referring to Figure 1 , there is shown a single or dual layer focal plane array. In the ray trace rendering, a reconfigurable reflector device 2 is placed in the focal plane of a microwave lens 1. The microwave energy, at the focus of the reconfigurable reflector, is electronically selected and channelled into the lens or reflector system either directly or by a radio frequency transition to a radio frequency printed circuit board (PCB). The printed circuit board is shown by way of example with a patch or dipole element array. Depending on the array position of a selected launch element in the focal plane of the lens, the microwave beam 3 is steered. The following description covers a number of preferred embodiments, using a parallel plate waveguide approach.

A parallel plate waveguide is the most efficient form of waveguide since it has only two, rather than 4, conducting walls for the electromagnetic wave to dissipate its energy. A parallel plate waveguide supports many electromagnetic modes. However the most common mode is the fundamental mode where the distance between the plates is less than half a wavelength, and the electric field Έ' in perpendicular to the plates and magnetic field Ή' is transverse to the plates. This is a transverse electromagnetic (TEM) wave and is also the mode supported by a planar wavefront propagating in free space. The parallel plate waveguide can be either air or dielectric filled. In general, the parallel plate waveguide presents a very low impedance, that is both reactive and dissipative to the wave, i.e. there will be resistive losses in both the plates and in the dielectric. The impedance of the guide depends on the distance between the plates and the dielectric constant of the material in the guide. In most circumstances, a radio frequency transition, or impedance transformer, is required to move between the parallel plate waveguide and radio frequency cables, transmission lines or free space.

Figure 2 shows a parallel plate waveguide 4, sandwiching a dielectric 5, for example silicon, in which there is positioned a reflective electromagnetic switch 8/10. The switch 8/10 has the property of being either highly transparent when 'off 8 (i.e. mostly the energy 6 is transmitted through the switch 7, and only a very small percentage of the illuminating energy is reflected 9 and absorbed) or extremely reflective 10 (i.e. a very high percentage of the illuminated energy is reflected 12, and little energy is transmitted 13 and absorbed) when On'. Typically the parallel plate is an insulating dielectric material, for example a polymer or silicon, with metallised surfaces. The metallised surfaces would be highly conductive, of low surface roughness and less than a half wavelength in dielectric apart, so supporting a transverse electromagnetic wave. The electromagnetic reflective switch would be actuated either electronically or electromechanically. For the electronically actuated case, the switch may be a PiN diode, where current is injected (i.e. forward biased) into the intrinsic T volume from the P and N doped regions to form a highly reflective solid state plasma composed of a high density of free carriers that is, electrons and holes (>10 ¹⁷ cnr ³ ). For the electromechanically actuated case, a dielectric column might be changed using a MEMS device, (e.g. a micro-actuator or a micro-pump, for example an insulating dielectric (e.g. an oil) replaced by a conducting material (e.g. mercury). In both cases, the conducting conduit when activated should cause a radio frequency short between the parallel plates.

Figure 3 shows an active electromagnetic feed 16/17, fed from an radio frequency source 14, via a coaxial cable, that only behaves as a feed when activated 17, launching radio frequency power 15 into the dielectric 21 , with little power 19 reflected back at the radio frequency source 14. Otherwise, when inactive 16, the inactive electromagnetic feed does not emit significant power 20 into the parallel plate waveguide and presents a reflective termination 18.

Figure 4 shows a selectable field probe 23/24, configured in active and inactive states and which is essentially the same set-up as for the active feed shown in Figure 3, except that the active feed now works in reverse, as a field probe with the parallel plate waveguide, collecting power 22, which in turn drives a radiating element 25, such as a patch. The strength of the field 26/27 depends on the whether the field probe is 'off 23 or 'on' 24. Significant power 28 passes by the field probe if it is 'off 23. The power 29 is reduced if the field probe is 'on' 24.

The reconfigurable reflector, electromagnetic feed and the reconfigurable field probe may not necessarily extend through the entire thickness of the parallel plate waveguide. As shown in Figure 4A, for the reflective (a), feed (b) and field probe (c) configurations, the switch 8 may be positioned on a conducting pillar 8A, so reducing the size of the closure gap of the RF MEMS switch or the thickness of the intrinsic region of the PiN diode. By adjusting the relative and absolute dimensions of the conducting pillar and the parallel plate spacing, the performance of the feed (signal injection case) or field probe (signal extraction case) may be tuned to minimise such design parameters as activation and de-activation times, insertion loss and bandwidth, at both the individual reconfigurable element and complete array levels, for a variety of active and inactive element cases. It is recognised that the introduction of a conducting pillar will generally limit the bandwidth of the surface array antenna when the switch is inactive, but should reduce activation powers and switching times of the reconfigurable elements when PiN diodes perform the switching function. It is noted that sloping the walls of the conducting pillar 8A to form a circular frustum 8B may help control the frequency sensitive effects in all three cases shown in Figure 4A.

Figure 4B depicts configurations for an electromagnetic feed (a) or a selectable field probe (b) where the reconfigurable element's coupling and matching are controllable through the reconfigurable element being composed of a sub-array of elements, where the elements, for example, are individually controllable RF MEMS or PiN diodes.

By way of example only, Figure 4B(a) shows a cross-section of a 2D sub-array of four electromagnetic feeds, which is further depicted in plan form in Figure 4C(a), where it can be seen that a 4 way radial split is used to divide the power equally between four reconfigurable elements. The spacing, relative intensity and phasing between the elements control the radial pattern between the parallel plates produced at a particular frequency. The resulting pattern may range from approximately omni-directional, single lobe or multi- lobe and be controllable in terms of its point direction. That is, by varying the number and selected pattern of active reconfigurable elements within the sub- array, the sub-array may be controlled in terms of its frequency/magnitude/phase characteristics. Figure 4C(c) illustrates in plan- view the case where 3 of the 4 elements are active. Reconfigurable feed elements may also be controlled through their termination to the parallel plate waveguide (e.g. capacitive or inductive, implemented as tuned stubs or varactor diode configurations) to provide similar levels of control in terms of their frequency/magnitude/phase responses. As a further example, a reconfigurable element may be implemented as a central PiN diode with small ring of diodes around it. The combination of diodes and their DC biasing would give the reconfigurable element different responses and also, potentially, some spatial directivity. A set of fixed electromagnetic feeds may also be introduced within the sub-array of reconfigurable electromagnetic feeds to a facilitate highly efficient fixed feeds with some controllable directivity.

Again, by way of example only, Figure 4B(b) shows a cross-section of a 2D array of four field probes, which is depicted in plan form in Figure 4C(b), where it can be seen that the four field probes feed radiative elements. The phasing between the field probes will depend on the interaction with any travelling wave between the parallel plate and how the tapped energy is coupled to the antenna element, possibly in the form of a patch or sub-array of patches. Due to the dielectric constant between the parallel plates possibly being greater than free space, the phase difference between the feed points to the antenna elements (e.g. patches) may be greater than that inferred by their free space distances. This may cause the beam pattern to be preferentially steered. The spacing between the field reconfigurable probes is likely to be constrained by the dimensions of the antenna element or sub- array of antenna elements. By selecting a single element of the sub-array, (see Figure 4C(d)), the transmitted polarisation and the phasing of the patch may be controlled. Left and right hand circular and all linear polarisations are possible according to the position feed points to the antenna element (e.g. a patch). Further control of phasing is possible by dividing the patch into concentric annular regions that can be spatially selected through the choice, positioning and weighting of the selected plasma feeds. The reconfigurable field probe elements may also be controlled through their termination to the parallel plate waveguide (e.g. capacitive or inductive, possibly implemented as tuned stubs or varactor diode configurations) to provide similar levels of control in terms of their frequency/magnitude/phase responses. A set of fixed electromagnetic field probes may also be introduced within the array of reconfigurable electromagnetic field probe elements to a facilitate a combined field probe of higher efficiency but still possessing some spatial selectivity. Figure 5 shows a combined active feed 30 and active reflector 31 , both in their active 'on' state, as described above, for Figures 3 and 4. If desired, a single active feed 30 might be enclosed within many active reflector elements 31 which collectively form a two dimension reflector, as discussed in later paragraphs.

Figure 6 shows a single parallel plate feed 30 to a field probe 32 that has been reflected off a reconfigurable electromagnetic reflector 32.

Figure 7 shows the same sort of configuration as shown in Figure 4, that is an active feed 31 and a field probe element 32, except that the configuration shown in Figure 7 has been folded and a U-turn 34 has been introduced.

Figure 8 illustrates a two layer, circular parallel plate waveguide 34 which is fed from a centre feed 33. In this structure, a U-turn 35 is introduced around the circumference of the disc. Assuming the dielectric media of the guide is uniform, then the energy introduced by the feed 33 comes to a focus on the other side of the disc.

As illustrated by Figure 9, the focal point on the other side of the disc may be moved by introducing a hyperbolic reflector of variable eccentricity 36 and 37.

Figure 10 illustrates how the focal point may also be moved by introducing an elliptical reflector and varying its eccentricity 36A and 37A.

A hyperbolic reflector causes the focus to be within a half disk radius of the centre. An elliptical reflector places the focus within half a radius of the circumference. Figure 11 shows typical elliptical reflector shapes 39. In Figure 11 an active reflector 40 is highlighted in the left-hand depiction, and radial control tracks 42 to the same reflector shape 41 are shown in the right-hand depiction.

Figure 12 illustrates the electromagnetic field intensities for a circular disk with centrally fed reflectors 38. The inactive state is shown as 34A and the active state is shown as 34B. The second focus of the ellipse 40 is close to the circumference.

The selectable electromagnetic field probes and radio frequency feeds are normally configured as either a radial array 36 or a Cartesian array 38, as shown in Figure 13. The array elements are either shown a circular elements 35 or square patches 37.

The elements within the array may be electronically addressed individually or selected as groups. A simple Cartesian addressing scheme is shown in Figure 14. Here a x addressing multiplexer 40 and a y addressing multiplexer 42 are controlled by an address selector 40A to select required elements 39. This required functionality is normally achieved by a single computer programmable logic device (CPLD) or it can be integrated into the array.

As shown in Figures 11 and 12, a reconfigurable elliptical reflector 40B can be configured to illuminate part of the circumference of a cylindrical parallel plate device. The contiguous illuminated regions transit into microwave transmission lines 41 , as shown in Figure 15 for a disk device 34 which fed from a central feed 38. These lines 41 can be configured to selectively illuminate groups of one, two launch elements 42, or more than two launch elements as 42 and 43 combined of a 2D array, as shown in Figures 16 and 17. The advantages of this approach over other more conventional electronic switch techniques are:

Flexibility in terms of providing single pole to many throw

(SPMT) switching, while maintaining a reasonable match of high percentage bandwidth.

Inherent ability to act also as a tapered power divider while switching between a very large number of states or poles while maintaining a low insertion loss.

In Figure 16 either all the elements or a pair of adjacent elements are active elements 46 and fed by radio frequency tracks 47. Figure 17 shows how routing between the plasma device ports and twenty five element discs 45, for radial and Cartesian element configurations. It is important to realise that if multiple active elements are to be simultaneously selected and excited, then the electrical path lengths 47 need to equalised. (N.B. This is not true for single element selection).

Figure 18 shows, using a ray trace, an offset parabolic reflector configuration 48 where the centre element of the focal plane array 47 has been excited to produce a boresight beam 49, and also shows front 50, side 51 and top 52 views. By choosing single non-central elements within the focal plane array, the beam scans away from boresight. By selecting multiple elements, the beam can be broaden and the side-lobe levels and null structures adjusted. By using an elliptical paraboloid the beam may be selectively broaden in one dimension.

A single layer focal plane antenna is shown in side view in Figure 19 and in top view in Figure 20. As shown in Figures 19 and 20, the central radio frequency/output port 54 feeds into a parallel plated device 55 with an elliptical reconfigurable reflector, and within the enclosed area of which there are active electromagnetic field probes 56. In turn, the electromagnetic field probes feed through to a radio frequency printed circuit board (PCB). The PCB 57 has tracks that feed a focal plane array 58. The focal plane array 58 may be made larger than the parallel plate device according to the ratio of the refractive indices of the dielectric (n) within the parallel plate 55 and the lens 53, for example silicon (n=~3.4) and polythene (n=~1.5). In other words, a high dielectric material such for example as silicon can be used to configure the first layer of the antenna 55, and a lower dielectric material such for example as polythene can be used for the lens 53.

The first focus of the elliptical reflector is the central feed 54. This is separately illustrated in Figure 21 which shows layer 1 of the antenna in side view on the left and in top view on the right. Two positions of elliptical reflector 59 (or different eccentricity) are shown in the top view.

The second focus is at selected field probe 56, shown separately in side view on the left and in top view on the right in Figure 22. Figure 22 shows a radial array of field probes 57, one 56 of which has been selected. Typically, for maximum compactness, the field probes 57 will be at a half wavelength spacing within the dielectric material of the parallel plate.

Continuing with the description layer by layer, Figure 23 in side and top view illustrates layer 2. In layer 2, a lower dielectric radio frequency PCB is used to magnify layer 1 in size. This is done using an array of radially directed transmission lines 60. Typically, for half wavelength element spacing, the magnification factor will be limited to the square root of the ratio high dielectric constant to the low dielectric constant.

In Figure 24, the transmission lines are shown to terminate behind the array of circular launch elements 58 to which they are attached directly, or are coupled electromagnetically 61. These launch elements are typically patches that in turn launch into the microwave lens.

Figure 25 shows a dual layer focal plane antenna in side view. The dual layer configured in much the same way as the single layer configuration, except the single layer parallel plate device is replaced by a dual layer device 63. This approach reduces the possible interaction of the reflecting elements and field probe devices. The dual layer focal plane antenna is shown in top view in Figure 26.

As shown in Figures 27 and 28, the central radio frequency input/output port feeds into a parallel plated device with an elliptical reconfigurable reflector. Typically a high dielectric material such for example as silicon can be used to configure this dual layer parallel plate device.

The operation of the focal plane antenna has already been described for a single layer parallel plate. The first focus of the elliptical reflector is the central feed, separately illustrated in Figure 27 which shows a section of the antenna which now incorporates a corner-turn 64 to an upper layer.

The second focus is at selected field probe in the upper layer, separately illustrated in Figure 28 which shows a radial array of field probes in the upper layer 65, one of which has been selected. Typically the field probes will be at a half wavelength spacing within the dielectric.

The description, layer by layer, progresses as previously described for the single layer device.

A potential extension to the U-turn approach discussed above is to transit out of a first circular disc via radio frequency transition ports that feed regular inward or outward going radial tracks. This alternative approach is depicted in Figure 28A. When the tracks are λ _ε 12 apart, then the tracks can transit into a second parallel plate disc made from a dielectric material of effective permittivity 'e'. If 'ε' is less than the dielectric constant of the first disc, the tracks travel outwards and in effect act as a spatial expander (see Figure 28A) . If 'e' is more than the dielectric constant of the first disc, the tracks travel inwards and in effect act as a spatial concentrator (see Figure 28B) . In effect, the parallel plate's field is sampled by the ports around the first disc and is weighted in phase and amplitude, and can be translated to both smaller and larger second discs that can directly feed lens or reflector systems. Air may be used as the dielectric for the second disc, and so minimise the loss to those in the metal of parallel plate, thereby potentially allowing a very cost effective construction. Amplification can be introduced into the radial tracks, so allowing intermediate amplification on either transmit or receive or both if Tx/Rx switch is introduced (see Figure 28C). The amplifiers need to matched in amplitude and phase, or compensated. The compensation may be achieved by correcting distortions of the reconfigurable reflector. Power-combining of individual amplifier outputs is achieved within the second disk on transmit. Signal averaging of individual amplifier outputs is achieved within the first disc on receive.

Figure 29 shows a high dielectric spherical lens where the focal points are buried in the lens. Although this approach produces a wide field of view, it tends to be quite high loss and narrow band, i.e. it requires a matching or blooming layer. When a wide field of view is required, an important approach is the use of a Lijneburg Lens, as shown in Figure 30. Essentially, a Lijneburg lens is low dielectric (<2) graded refractive index spherical device that produces a focus on its surface by forcing an electromagnetic wave to follow a curved trajectory 69, rather than a straight path. This approach requires the radio frequency PCB and array of elements to form a sectoral cup 67 to launch directly into the lens. One alternative is to grade the refractive index to achieve a flat focal plane. Another alternative is to select a sub-group of weighted adjacent elements using the reconfigurable reflector device across a flat focal plane that has the same phase and amplitude spatial distribution of a planar cut through a Lijneburg lens.

Due to the use low permittivity materials, the attractiveness of the Lijneburg lens is that it can have a very low insertion loss (e.g <ldB) for very high gain. The Luneburg lens is a sphere with a radially symmetric refractive index profile given by the following formula: n - c (rx ² - ry ² - rz )

where n and c are constants, R is the radius of the sphere, and rx, ry, rz are Cartesian coordinates.

As shown in Figure 31 , to cover 360° in azimuth and ±60° (or -30° to +90°) in elevation, a three lens configuration can be used. In this way either the horizon or the zenith may covered. The left-hand diagram in Figure 31 shows three lenses 69 with selected beams at 120° azimuth spacing and 0° elevation. The right-hand diagram in Figure 31 shows three lenses 70 with selected beams at 120° azimuth spacing and a 30° upward elevation.

As shown in Figure 32, the full sphere may be covered using one more LCineburg lens 71. The left-hand diagram in Figure 32 shows four lenses with three selected beams at 120° azimuth spacing and 60° downward elevation, and with the fourth lens's beam pointing at zenith. The right-hand diagram in Figure 32 shows four lenses with selected beams at arbitrary azimuth and elevation angles. Each lens and launch array covers a field of view of at least π steradians.

A dual polar approach for a focal plane approach (or a surface approach as in the Luneburg case) is shown in cross-section in Figure 33. In Figure 33, two separate reconfigurable reflector devices are used to feed common launch patches, either individually or simultaneously to control polarisation. This is more clearly illustrated in top view in Figure 34.

Referring to both Figures 33 and 34, it can be seen that a surface array antenna for the more general dual polar case may typically be described in terms of :

(a) a launch structure (54) comprising a single feed or a pair of feeds;

(b) a parallel plate structure (56) comprising a single parallel plate cavity or a pair of parallel plate cavities;

(c) an array of reconfigurable reflectors and an array of selectable field probes or a pair of arrays of reconfigurable reflectors and a pair of arrays of selectable field probes located in the parallel plate structure (56), where the array of reconfigurable reflectors or the pair of arrays of reconfigurable reflectors, when configured, is able to focus on a sub-group of selectable feeds or a pair of sub-groups of selectable feeds;

(d) a transition structure (55) that contains a pattern of printed tracks, radio frequency cables or waveguides or two interlaced patterns of the tracks or waveguide that magnifies the pattern of selectable field probes or pair of patterns of selectable field probes (57); and (e) an array of single or dual polar launch elements (58) that are positioned at or close to the focal surface of a lens or reflector (53), or lie on a curved surface; and the surface array antenna being such that:

(f) the launch structure (54) launches an electromagnetic wave or a pair of electromagnetic waves into the parallel plate cavity or the pair of parallel plate cavities in the parallel plate structure

(56);

(g) the parallel plate structure (56) contains the array of reconfigurable reflectors and the array of selectable field probes or the pair of arrays of reconfigurable reflectors and the pair of arrays of selectable field probes;

(h) the parallel plate structure (56) has a first layer which contains the reconfigurable reflectors and which, when configured, is able to focus on the sub-group of selectable feeds or the pair of subgroups of selectable feeds, which are positioned in either the first layer of the parallel plate structure or a second layer of the parallel plate structure, linked by a U-turn from the first layer;

(i) the selected feeds transit to the transition structure (57);

(j) the transition structure (57) transits to the array of single or dual polar launch elements (58); (k) the array of single or dual polar launch elements (58) produce a selected beam directed according to a sub-group of selected field probes; and

(I) the transition of the selected feeds to the transition structure (57) is such that the selectable field probes are able to be selected individually or collectively in order to launch via the transition structure through either single polarisation or dual polarisation antenna arrangements (58) directly into free space or into a lens or reflector means (53) and thereby optically perform electronic scanning in two orthogonal directions.

By phasing signals between orthogonal mode polarisation ports (e.g. vertical and horizontal) of a patch, the element can be configured to launch any desired polarisation. For example, if a 90° phase shift is set up between the ports, then circular polarisation results.

The ray trace depicted in Figure 35 shows a spherical lens 72 (in this case and by way of example only a Luneburg lens) with an array of feeds constrained to a plane behind the lens 73. This positioning is a likely practical restriction when a printed array of electronically selectable patches is used to feed the spherical lens, since current radio frequency PCB technology cannot be easily conformed to the sphere. Essentially, this means a void exists between the plane and the lens. At the centre of the array there is no gap because the radio frequency PCB and the lens are in contact. As illustrated in the contour plot 74 in Figure 35, the pattern is perfectly symmetrical and fully focused. The gap is greatest at edge of the array feeds and, as a result, these outer elements are slightly defocused, causing a broadening of the beam, a departure from symmetry, and some reduction in gain. This effect is shown in the overlaid azimuth patterns 75 of Figu re 35, with the scanned elements being reduced in gain by ~2dB at the edge of scan 76 (i.e. ±30°).

Figure 36 illustrates two examples of faceted sets of selectable feeds, arranged so as to extend the angular field of view of a spherical lens in both azimuth and elevation. Two configurations are illustrated in perspective at 77 and 78, and in plan views at 79 and 80. The left configuration 77 shows five planar selectable feed arrays, and the right configuration 78 shows nine planar selectable feed arrays. Both configurations cover approximately a ±60° field of view. However the configuration with nine faces provides better quality higher gain beams away from the principle axes.

By faceting in a similar way the back of the lens the air gaps between the radio frequency PCBs and the lens may be avoided. For the particular case of the Luneburg lens, the outer layer of the antenna has a refractive which is close to air (i.e. ε=1). This adjustment has only a small effect on the match but adds to the overall integrity and robustness of the antenna. The resulting beams are slightly defocused, especially for the centrally launched beams. This defocusing may be reduced by adjusting the layered structure of the graded index lens to refract more at centre of the facets, and so helping to restore the focus. In regard to the construction of the multiple facet radio frequency PCBs, these may be made as a single board 81 using flexible circuits to allow folds at the vertices 82. Radio frequency transmission lines 83 , power distribution 84, and digital control signals 85 may be routed across these folds. This type of construction is illustrated in Figure 37. The radio frequency PCB may be multiplayer, and may contain active radio frequency devices (e.g. switches and amplifiers) and passive components (e.g. transmission lines, patches, power and control lines). This approach avoids the need for expensive multiple radio frequency connectors on each facet (i.e. selectable feed array face).

A first wireless application for the invention is shown in Figure 38, where small cell backhaul links are shown between building rooftops. The links may also be between ground level, street furniture, installed systems and building rooftops.

A second wireless application is shown in Figure 39, where a mid- range access point, operating at a millimetre wave frequency (e.g. 60GHz), is operating in an airport lounge or similar public space. Here an electronically scanned beam providing instantaneously wideband communications is time division multiplexed between mobile clients.

Figures 40 and 41 illustrate the focal surface array antenna having a conformal surface without a lens or a reflector. More specifically, for applications, such as, mobile ground terminals for satellite communications it is useful to provide: wide angular beam forming, directed beams and fast beam selection (e.g. greater than hemispherical coverage) using a highly compact aperture. This is not possible using a flat planar antenna design which will suffer significant scan-off losses. This loss is significantly reduced and much wider fields of view achieved by interfacing a non-planar antenna array face (e.g. a segment of a sphere) directly to the beam forming and beam steering circuitry, referred to as circular distribution network. Moreover using this approach, the antenna's aperture efficiency and its associated cost, relative to a similar performing lens or reflector system, may be improved significantly for very wide fields of view, (e.g. hemispherical or greater coverage).

Figure 40 illustrates a circular distribution network composed of rings of selectable field probes and controllable elements intended for beamforming and beam selection purposes.

This circular distribution network may be used to generate the beam selection and beam steering phase and amplitude signals for a non-planar array, such as that instanced in Figure 41. The circular distribution network performs a bidirectional conformal mapping from a control plane, through selectable multiple RF transmission lines, waveguides or cables to the antenna elements making up the non-planar array face, shown in the example as a truncated icosahedron. The length of the individual RF transmission lines which form the transition structure may be used to apply further fixed beam steerage (i.e. time delays) at the array face if required and so further reduce the form factor of the antenna.

It will be seen that Figure 40 shows four examples, labelled as Example A, Example B, Example C and Example D, of a circular distribution network in a plan view set to different states. The sub-systems of the circular distribution network have previously been described as:

1. Central feed

2. Parallel plate structure

3. Array of reflectors

4. Array of field probes

In Figure 40, the state and position of the field probes have been represented by large circular symbols:

. shaded grey in its active state

. as a cross when reflective

. as a plus when transparent

A fourth state which is absorptive has also been included in the key.

The complete circular distribution/selection network may be implemented as a multi-layer RF PCB with surface mounted RF connectors at the central feed and field probe points.

Typically, each field probe would be less than one wavelength's distance from its neighbours. Associated with each field probe would be a control structure to select its state. If implemented as either an electronic or electro-mechanical switch (e.g. a PIN diode circuit or RF MEMS switch), the field probe may be implemented around a switch network to either produce short or open circuit states, the switch connecting to a metal via between the parallel plates. By routing the current generated, due to the voltage across the probe, to either an absorptive load or a reflective short circuit or a transmission line/RF cable, the required states may be achieved. The need for an absorptive mode will depend on the required performance of the antenna and may be omitted if performance (e.g. sidelobe levels) can be achieved without the need for selected absorption of residual RF energy. The switch network should generally be configured to minimise the insertion loss of the field probe when selected and active. When inactive and transparent it should be configured to minimise interaction with the electromagnetic wave across the operational band of the antenna. When reflective or absorptive, it should be when isolated from the RF transmission line or cable across the operational band of the antenna. It is noted that intermediate states between maximum reflection and absorption may be useful in optimising performance and achieved for example by controlling the bias of the PiN diode switch.

Figure 40 also shows controllable elements. Their state and position has been represented by small circular symbols: shaded white in its transparent (inactive) state shaded grey in its absorptive state

shaded black in its reflective state

These controllable elements are typically positioned at sub half- wavelength distances, and typically may lie between the field probes. They approximate to continuous reconfigurable reflectors/absorbers. In their inactive state they are transparent across the band of operation. If space allows they may also simplify the field probe switch network, obviating the need for the field probe to have reflective and absorptive modes and so improving its insertion loss when active and connected to its RF transmission line or cable.

Example A illustrates a circular distribution/selection network 101 A where the centre feed 102 is surrounded by active field probes 103 and controllable elements 104T set to their transparent state. In effect each radially positioned field probe 103 taps off a percentage of the energy associated with an outward going circular wavefront generated as a TEM wave between the parallel plates, (as discussed for the transmit case). The circular distribution/selection network is fully reciprocal and executes exactly the same process in the reverse direction for an inward travelling wave (as discussed for the receive case). Each field probe is connected via an RF cable, waveguide or transmission line to an antenna element lying on a conformal surface. If all the lines are of equal length, and the signals of roughly equal amplitude and the antenna elements approximately equally spaced at less than half a wavelength on a sphere or segment of a sphere, then the wavefront leaving this surface will be spherical and the beam either omni-directional or a broad floodlight, assuming the elements are all circularly polarised (as further discussed for View E in Figure 41).

Example B illustrates a circular distribution/selection network 101 B where the centre feed 102 is surrounded by an inner ring of active probes 03 and controllable elements 104T set to their transparent states, encircled by two layers of outer reflective rings of reflective probes 103R and controllable elements 104R set to their reflective states, and a fourth and final outer ring of absorptive field probes 103A and controllable elements 104A set to their absorptive states. In this set of states the active inner ring of probes routes the majority of the RF energy equally to the RF transmission lines or cable. The rings 2 and 3 comprising both reflective probes and reflective elements placed at approximately λ/4 and 3λ4 and conceptually reflect back the RF energy to the inner ring. Any remaining energy reaching ring 4 is absorbed by the probes 103A and elements 104A set to their absorptive states. In effect, five adjacent field probes and the inner ring have been selected, and the surrounding field probes in rings 2 and 3 configured to reflect back any energy that would have been missed if the surrounding probes and elements were transparent (or not present). When connected to an array of elements as described in Example A, a central beam is generated (as further discussed for View D in Figure 41).

Example C illustrates a circular distribution/selection network 101C where the centre feed 102 is surrounded by an inner ring of active and reflective probes 103A and reflective and transparent controllable elements 104R and 104T. Ring 2 is similarly composed. Ring 3 has almost the same configuration of states except only one field probe is active. The outer ring is totally composed of field probes and elements 103R and 104R set to their reflective states. In effect, five adjacent field probes in rings 1 , 2 and 3 have been selected, and the surrounding field probes configured to reflect back any energy that would have been missed if the surrounding probes and elements were transparent (or not present). The relative power of the energy intercepted by the five active probes may need to be equalised. This equalisation may be achieved by iterative adjustment of the surrounding controllable elements' states, or continuous adjustment of their biases. When connected to an array of elements as described in Example A, an off-centre beam is generated (as further discussed for View F in Figure 41).

Example D illustrates a circular distribution/selection network 101D where the centre feed 102 is surrounded by an inner ring of reflective and transparent probes 103R and 103T, and reflective and transparent controllable elements 104R and 104T. Ring 2 is composed of active and reflective probes 103 and 103R, and reflective and transparent controllable elements 104R and 104T. Ring 3 has almost the same configuration of states, except three field probe are active. The outer ring is totally composed of field probes and elements set to their reflective states 103R and 104R. In effect, five adjacent field probes in rings 2 and 3 have been selected, and the surrounding field probes configured to be either reflective or transparent in ring 1 and reflect back in ring 4. As with Example C, the relative power of the energy intercepted by the five active probes may need to be equalised. This equalisation may be achieved by iterative adjustment of the surrounding controllable elements' states, or continuous adjustment of their biases. When connected to an array of elements as described in Example A, a scanned beam is generated (as further discussed for View C in Figure 41).

It is important to recognise that the circular distribution network is highly flexible and allows field probes and controllable elements to be selected in a wide variety of configurations composed of 1 , 2, 3, 4, 5 or more active field probes. It will further be seen that Figure 41 shows a perspective view of a conformal array in the form of a truncated icosahedron, (i.e. a twenty face structure with five adjacent faces omitted) labelled as View A with circularly polarised patches shown as disks.

Five other projective views are shown as follows.

View B illustrates a top view representation where ten of the icosahedrons' faces are visible from above, and five more triangular faces from below, shown as dashed ellipses, making a total of fifteen faces. As illustrated, it will be appreciated that there are six pentagonal groupings formed from sub-groupings of five adjacent circular patches, including the pentagon positioned around zenith. Alternatively, by pointing a normal to a triangular face at zenith and only omitting a patch from the bottom face of the full icosahedron (which could be used to support the geodesic structure), nine pentagonal groupings of adjacent elements can be formed.

View C illustrates the case where five triangular faces of the icosahedron are fed via five transmission lines or five RF cables. For illustrative purposes, the inactive elements (i.e. the elements connected to inactive field probes) have not been shown. The circular distribution/selection network would be set to the states shown in Example D. When the central launch is fed with an RF signal, the phase of each active field probe may be equalised by adjusting the length of the RF transmission lines/cables. The resulting beam will then have a boresight along the radial to the centre of the pentagon formed by the five adjacent triangular surfaces, on which the antenna elements are centrally positioned. This selected geometry is one of five that produce beams in the lower hemisphere. It will be realised that the relative lengths of the RF lines associated with each ring of field probes can be set, so that the associated elevation beam's pointing direction is modified. This assumes that the antenna element pattems have sufficient angular coverage to allow for this adjustment without compromising the gain.

View D illustrates the case where the top five triangular faces of the icosahedron are fed via five transmission lines or five RF cables. For illustrative purposes, the inactive elements (i.e. the elements connected to inactive field probes) have not been shown. The circular distribution/selection network would be set to the states shown in Example A. In this particular case, the boresight of the resulting beam will be along the radial line through the centre of the activated pentagon, directed at zenith, provided all the transmission line lengths are equal. With some reduction in gain and beam symmetry, small changes to beam boresight can be achieved by deactivating one, two or three of the selected field probes within the defining pentagon.

View E illustrates the case where all fifteen triangular faces of the truncated icosahedron are fed via fifteen transmission lines or fifteen RF cables. The circular distribution/selection network would be set to the states shown in Example B. In this particular case, the resulting beam would be close to omni-directional, except at nadir.

View F illustrates the case where five triangular faces of the icosahedron are fed via five transmission lines or five RF cables. For illustrative purposes, the inactive elements (i.e. the elements connected to inactive field probes) have not been shown. The circular distribution/selection network would be set to the states shown in Example C. When the central launch is fed with an RF signal, the phase of each active field probe may be equalised by adjusting length of the RF transmission lines/cables. The resulting beam will then have a boresight along the radial to the centre of the pentagon formed by the five adjacent triangular surfaces, on which the antenna elements are centrally positioned. This selected geometry is one of five that produce beams in the upper hemisphere. It will be realised that the relative lengths of the RF lines associated with each ring of field probes can be set so that the associated elevation beam's pointing direction is modified. This assumes that the antenna element patterns have sufficient angular coverage to allow for this adjustment without compromising the gain.

The boresight gain of the antenna beam for a particular selection of antenna elements will depend on the superposition element gain complex patterns in the far field, but will be reduced by the losses incurred by the RF signal passing through the circular distribution/selection network and the RF lines feeding the antenna elements. The boresight directivity relative to a single element might be increased by 10 LogioM dB where M is number of selected elements. However this gain depends on the orientation and beam width of the selected elements and will reduce dramatically if the main beam patterns do not overlap sufficiently. In practice, the increased gain should be such as to more than overcome the losses associated with beam forming and beam selection circuitry, while also providing good angular coverage and acceptable cross-over losses.

In practice, a geodesic geometry might have many beams. It is to be appreciated that the embodiments of the invention described above with reference to the accompanying drawings have been given by way of example only and that modifications may be effected. Individual components shown in the drawings are not limited to use in their drawings and they may be used in other drawings and in all aspects of the invention. The invention also extends to the individual components mentioned and/or shown above, taken singly or in any combination.