GB2372173A - Increasing capacity of radio coverage by overlaying sectored cells with offset sectored cells - Google Patents

Abstract

A wireless cellular communications network provides increased capacity by overlaying existing cells with new cells, in which the antenna beam patterns of the sectors 41, 43, 44, 46 of the overlay cells are offset from the antenna beam patterns 11-16 of the original existing underlay cell sectors. This results in reduced antenna cusps, being regions of reduced antenna gain which occur between cell sectors in a conventional cellular network. Reducing antenna cusps ensures that radio link budgets can be maintained at a uniform level across the cell, and bit error rates can be kept as low as possible. The effect of the invention is therefore to increase cellular capacity to a further extent than is possible with conventional underlay/overlay network architecture.

Description

INCREASING CAPACITY AND IMPROVING COVERAGE OF A CELLULAR RADIO NETWORK Technical Field The present invention relates to a wireless cellular communications network and in particular to methods for simultaneously increasing the capacity of such a network as well as improving the area coverage provided by the network.

Background of the Invention Cellular networks such as those used by mobile telephone systems such as the Pan European Global System for Mobile Communications (GSM) and the US Digital Cellular System Personal Communication System (PCS) are well known and their use by members of the public is on the increase. The development of a typical cellular network can be divided into two phases. In the first phase, the primary objective of the network is to provide coverage to a target customer base, and hence the deployment of the network equipment is made across as large a target area as possible to maximise potential customer base. Once this initial coverage has been provided, the network deployment moves to a second phase in which capacity of the network is increased amid growing customer demands.

Prior Art Various methods are known for increasing the capacity of a cellular radio network, and these are reviewed in detail in US 6,091, 955 issued to Nokia Telecommunication OY, on 18 July 2000. More particularly, US 6,091, 955 identifies four known prior art methods of increasing capacity on a cellular network, and these are discussed briefly below.

The simplest way to supplement capacity is by increasing the number of channels available to the network. In general, however, the allocated cellular spectrum per network operator is usually very limited, and hence this method is usually unavailable to give relief from capacity problems.

The second method is to decrease the size of the cells and to add more base stations, thereby effectively splitting the already deployed larger cells into smaller cells. This second method is accomplished by reducing the transmit power of each base station and/or subscriber terminals to minimise interference with adjacent cells, and then filling in the coverage gaps with additional base stations. There are, however, several engineering problems which result from the reduction in size of cells, and these are particularly enhanced in urban areas with high capacity requirements-More particularly, with cell sizes of radius of less than 1 km there is usually a need to lower antenna height below roof top level to ensure coverage at street level, but this results in the problem that prediction of ranges for these types of installations is less well understood and furthermore interference management and cell overspill cannot be equally controlled. Additionally, major investments in base stations and antennas is also required.

A third option for increasing capacity in highly localised regions with high capacity requirements is to deploy a"micro cellular"network with cells having small coverage areas and antennas below roof top level. The term"micro cellular" is not clearly defined in the art, as discussed in US 6,091, 955, but can be characterised by the above features. A micro cell can be overlaid onto the existing first phase cellular network, although this is not necessary, and micro cells can themselves simply be recognised as small cells in their own right. Micro cells provide the advantage that they tend to experience more line of sight signal propagation due to their extremely small size, but localised blockages in the antenna line of sight can produce serious shadowing effects which significantly

alter the coverage provided by such cells. As with cell splitting, it is generally thought that to implement a micro cellular network to increase capacity would present network operators with a significant investment in the required base stations sites and transmission connections.

A fourth method of increasing capacity is to use what is termed an "underlay-overlay"network wherein a radio network has two or more separate cell layers overlaid on top of each other. Frequently such multi layer networks are arranged into a macro cell layer providing overall coverage and a micro cell layer which provides capacity. Subscriber terminal handovers can be made between the layers on the basis of field strength or power budget measurements. The invention of US 6,091, 955 relates to an underlay-overlay network wherein handovers of subscriber terminals can be formed at the call set up phase and later on during the call using a handover procedure which is invoked based upon measurements of down link co-channel interference in the cell separately for each ongoing call.

In addition to the above four methods for increasing cell capacity based on cells with omni-directional antennas, it is also known that cells can be split up into sectors, with highly directional antennas being used to cover a sector of the cell using a specific frequency and/or polarisation. An example of such sectorisation of a cell will be described next with reference to Figures 1 and 2.

With reference to Figure 1, as an example of a frequency spectrum which can be used by cellular networks, Figure 1 shows the channelisation of the United States UNII band which provides ten 20MHz channels across the middle and upper UNII band. In order to increase the reuse of the available spectrum, it is known that signals can be vertically or horizontally polarised to provide for a degree of signal separation. Therefore, as shown in Figure 1 a particular channel within the band can be split into a vertical channel and a horizontal channel. With respect to. Figure 1 itself, the reader should be advised that the fill patterns used to

illustrate the vertical and horizontal channels within Figure 1 are used in the other Figures to denote sectors of a cell in which that channel may be used, and hence the fill patterns within Figures 2 to 6 describing the prior art and the present invention can be related back to Figure 1 to determine which channel and which polarisation of a channel is illustrated as being used in a particular cell sector.

Figure 2 illustrates an example cell of the prior art which has been split into six sectors to provide a 3600 range of coverage. Each sector is illuminated by an independent base station which is referred to hereafter as an Access Point (AP), which is the physical transmitting and receiving equipment located at the cell base station. There are therefore six separate access points for a single cell; each access point operates on a separate one of the 20 vertical or horizontal channels illustrated in Figure 1. Therefore, as shown in Figure 2, a single cell comprises six sectors, 1,2, 3,4, 5,6 provided by six separate access points (not shown), each arranged to cover a different 60'are from the other cells and each on a separate vertical or horizontal channel.

As acknowledged in US 6,091, 955, however, although it is usually convenient to draw a cellular network as a combination of circles of hexagons, these representations are in reality quite different from the actual world, where in fact the coverage area of a particular access point in any one cell is defined by the beam pattern produced by the access point antenna, and the direction in which the antenna is pointing. An example of an antenna beam pattern known in the prior art and which is highly suitable for use as the antenna of an access point base station is shown in Figures 7 and 8, which respectively illustrate polar plots of the horizontal and vertical gains provided by the antenna, and from which it will be seen that the antenna provides an antenna gain of 17dBi at boresight, which drops to 14.7dBi at 300 either side of boresight in the horizontal plane. The antenna is provided with a single main lobe and as will be apparent to the intended reader

provides very good front to back isolation and controlled side lobes to minimise provi 1 1 1 mast head interference. The antenna pattern is therefore particularly adapted to covering a sector of 60'arc.

Figure 3 illustrates a polar plot of the antenna patterns of 6 antennas with the horizontal beam pattern of Figure 7 when deployed to cover 600 sectors as shown in the example cell of Figure 2. Such an antenna beam pattern within a cell is typical of a multi-sectoral cell of the prior art, and although it is capable of providing network coverage across the whole geographic area of the cell, the coverage provided is not uniform across the whole area of the cell due to the separate antenna beam patterns providing areas known as"cusps"which occur between two antenna beam patterns. The location and nature of cusps are illustrated in Figure 3, where it can be seen that cusps 30,32 and 34 (amongst others) are provided at those points within a cell coverage area on a sector boundary. In particular, from Figure 3 it will be seen that cusps are areas of reduced antenna gain, of a magnitude of up to 2.3 decibels with the example antenna discussed earlier, and which occur at every intersection of a sector boundary.

The effect of a cusp is to reduce the gain of the access point antenna within the cusp by the magnitude of the cusp, in this case, 2.3 decibels. A cusp impacts on the link budget for a radio link between an access point and a subscriber terminal in the cusp by reducing the link budget by the magnitude of the cusp being 2.3 decibels. This reduction in the link budget for the radio link between a subscriber terminal located in a cusp to and from an access point will be manifested at first as an increase in the bit error rate for the link, as the transition from no BER to high BER can be 1 or 2 dB if forward error correction (FEC) is used, and just a few dB if there is no FEC. As will be known to the man skilled in the art, however, cellular networks, and particularly fixed wireless networks, are

usually designed to include a link margin, which allows for various changes in the link budget after equipment installation, due to antenna misalignment, water vapour, daily changes in propagation characteristics, and other interfering sources.

In the prior art case, the cusp of 2.3 dB reduces the link margin (which is typically engineered to be 10 dB to 15 dB), having the effect of reducing the network uptime, being the time a subscriber terminal spends effectively connected to the network via the link. Because link margin can vary over time in a manner similar to a stochastic process, reducing the link budget can result in an increase in the amount of time that the link spends below its operational threshold. When the link's propagation characteristics result in a drop below the operational threshold, the link will rarely remain in the region where it operates with a high BER, and instead the more usual result is that the link is lost completely, meaning that the subscriber terminal is cut off completely from the network. The effect of the cusps can therefore be generally characterised as causing a higher probability of network outage to those terminals situated at the cusps of an access point's antenna pattern.

In addition to the above, further problems are that in a time division duplex system where the same antenna is used for both the up-link and the down-link, then the antenna cusp will result in the above described effects on both the up-link and down-link. Furthermore, where a TDMA architecture is used within the network, the increase in bit error rate experienced by a subscriber terminal within a cusp which is operating right at the edge of the link margin results in more packet resends to and from that terminal which will impact upon the overall data transfer rate to every subscriber terminal in the same sector as the affected terminal, by virtue of the packet resends using packet slots which could have otherwise been used for other data packets.

Summary of the Invention The present invention addresses the above-described problems of increasing capacity on a cellular network and reducing the problems caused by antenna cusps in an integrated and synergistic manner, and particularly when used in a TDMA architecture. More particularly, the present invention increases the network capacity within a cell by providing additional access points within that cell each provided with its own antenna, the beam patterns of which overlay the beam patterns of the existing access points in order to provide an overlay-underlay network to increase capacity. The beam patterns of the antennas of the additional access points are identical to those of the original access point antenna beam patterns, such that every subscriber terminal within the cell can see two or more access points, thereby providing redundancy within the cell, and allowing for dynamic load balancing between the layers. In addition to providing the additional layer, however, the present invention also addresses the problem of reducing the antenna cusp by arranging the additional access point antennas such that their beam patterns are offset from the original access point antennas with the result that the original cusp regions are at or near to bore sight of the new access point antennas. This effectively reduces the cusp loss to almost nothing, thereby improving the radio link budget within the original cusp regions and reducing bit error rate within the cusp regions. This in turn leads to a more uniform network outage probability in the entire network coverage area. When used in a TDMA architecture, the reduction in bit error rate leading to a reduction in packet resends can result in an increase of overall data traffic throughput across an entire sector, and not just on the radio links to and from those subscriber terminals which were located in the original antenna cusp regions.

In view of the above, the present invention provides a wireless cellular communications network comprising one or more cells, each cell having a first

plurality of base stations each respectively arranged to provide a different coverage area defined by the respective antenna beam patterns and boresight azimuths of each base station, the network being characterised by each cell being provided with further base stations each of which provides a coverage area defined by the respective antenna beam patterns and boresight azimuths of each further base station and each of which at least partially overlay one or more of the coverage areas provided by the first base stations, the further base stations being further arranged such that their respective antenna boresight azimuths are different from those of any of the first plurality of base stations.

The present invention provides the primary advantage of increasing capacity of the cellular network by the provision of additional sectors which overlay the original sectors, but at the same time helps to increase the capacity further by the fact that the overlay sectors are angularly offset from the original sectors such that the overlay sectors provide increased antenna gain in the original antenna cusp regions. This improves the radio link budgets between subscriber terminals located in the original antenna cusp regions and the access points which results in improved link margins for the radio links to those subscriber terminals, which further contributes to an increased capacity of a network due to a reduced probability of making packet resends.

In a preferred embodiment, each base station provides a coverage area that at least partially overlays that of another base station and each base station is arranged to transmit and receive signals on a different frequency and/or with a different polarisation to the other base station.

The advantage of transmitting on different frequencies and/or with different polarisations is that co-channel interference on radio links between two subscriber terminals in the overlay area respectively communicating with the two different access points is reduced.

Furthermore, in another preferred embodiment the respective coverage areas provided by one or more of the additional base stations at least partially overlay each other, to provide multiple overlay layers. That is, several additional base stations can be provided each overlaying the same coverage area but with slightly different boresight azimuths thereby further increasing capacity and reducing antenna cusp regions.

Preferably, each base station within a cell is mounted on the same mast or other structure. This provides the advantage that it becomes relatively trivial to increase capacity on a cellular network without having to build more base station masts.

In the preferred embodiment of the invention, the first plurality of base

stations constitutes a first set of n base stations, wherein said further base stations are arranged into a further sets of n stations, wherein a > l and n > 2, a and n being whole numbers, each set of stations covering in total an area of sector xi ana each

x, base station in each set having an antenna beamwidth of- , the network being

further arranged such that the respective antenna boresight azimuths of the base

x stations in one of said sets are angularly offset by (x) 0 to corresponding (a+l) n

respective antenna boresight azimuths of at least one other set of base stations.

The arrangement of the preferred embodiment has the primary advantage that the area of the antenna cusp regions is minimised, as is the reduction in link budget caused by the remaining cusps.

Brief Description of the Drawings Further features and advantages of the present invention will become apparent from the following description of a particularly preferred embodiment

thereof, presented by way of example only, and by reference to the accompanying drawings, wherein : Figure I illustrates an example channelisation of a particular frequency spectrum available for use by the cellular network of the present invention; Figure 2 illustrates the channel allocations which can be made in a sectorised cell of the prior art ; Figure 3 is a polar plot of antenna beam patterns produced by access point antennas when arranged in a sectorise cell such as Figure 2 of the prior art; Figure 4 is a polar plot of the gain of access point antennas when arranged according to the present invention ; Figure 5 is an illustration of how additional access points can be provided to provide overlay coverage for sectors within the present invention; Figure 6 is an illustration of an example of how cells can be tiled together to cover a larger area, and how such tiles can be overlaid in accordance with the present invention; Figure 7 is a polar plot of the gain of a single antenna of the prior art which can be used as the access point antennas within the present invention; and Figure 8 is a polar plot of the gain in the vertical plane of the same antenna as shown in Figure 7.

The reader should note that the various fill patterns given to the cell sectors in Figures 2,5, and 6 can be related directly to those given in the spectrum channelisation diagram of Figure 1. That is, the fill patterns given to each sector in Figures 2,5, and 6 relate to the channel and polarisation to be used in that sector via Figure 1.

Description of the Preferred Embodiment A preferred embodiment of the present invention will now be described with reference to Figures 4,5, and 6.

Refemng first to Figure 5, a cell 20 of a wireless communications network comprises six base station terminals, referred to hereafter as access points, geographically co-located at the centre of the cell. Each access point is arranged to cover a different sector of the cell from every other access point to provide 360 coverage of the entire cell. Thus, for example, one access point provided at the centre of the cell has an antenna provided with a beam pattern to cover the sector 11, while another separate access point is provided with an antenna whose beam pattern is arranged to cover the sector 12. Similarly, other access points are respectively provided to cover the sectors 13,14, 15 and 16. Each access point respectively covering the sectors 11 to 16 is arranged to transmit and receive on a single frequency which is of a different frequency and/or polarisation to adjacent sectors. As will be apparent with reference to Figure 1, the sectors of cell 20 of Figure 5 correspond to channels 1,6, and 9 of Figure 1 when used with both vertical and horizontal polarisation. The use of three frequency channels and two polarisations gives the six different channels which can be used in each of the sectors 11 to 16. Please note that in the present invention the cell 20 of Figure 5 is identical in every respect to the cell 20 of Figure 2 of the prior art, and constitutes part of the original cellular network over which the present invention overlays additional cells.

As mentioned previously in the discussion of the prior art, it is known in the art that the capacity of cellular networks can be increased by adding base stations to provide coverage areas which overlay the coverage areas of the original network. Within the present invention, we use this concept to increase the capacity of our network by providing an overlay cell 40 comprising six sectors 41,

42, 43, 44, 45, and 46. Each sector is provided by a single access point provided with an antenna whose beam pattern is arranged to cover the respective sector. The access points of the overlay cell 40 are geographically co-located with those access points of the original underlay cell 20, and the access points themselves are in all respects identical, and comprise the actual radio hardware and software which is the physical transmitting and receiving equipment to provide the radio link. Any known radio architecture suitable for use as a cellular base station can be used as the access points within the present invention. As will be apparent to the man skilled in the art, providing an overlay cell 40 on top of the underlay cell 20 can effectively double the traffic carrying capacity of the cell, thereby increasing the capacity of the network.

In the present invention, however, we do not merely increase the capacity by the provision of the overlay cell 40, but by further arranging the antenna beam patterns of the overlay cell 40 access points such that they are arranged on different azimuths to the antenna beam patterns of the original underlay cell access points. Thus, for example, in Figure 5 it will be seen that the coverage sectors 41, 42,43, 44,45, and 46 in the overlay cell 40 are angularly offset from the respective corresponding cells 11,12, 13,14, 15, and 16 in the underlay cell 20.

That is, the boresight azimuths of the antennas of the access points in the overlay cell are arranged at different azimuths to the antenna boresight azimuths of the access points of the underlay cell. The purpose of this angular offset is to ensure that the antennas of the overlay cell access points provide an increased gain in the cusp regions of the antenna beam patterns of the underlay cell access points. This arrangement substantially reduces the effect of the antenna cusp resulting in a more uniform and consistent coverage area providing for improved link budgets and hence reduced bit error rates across the coverage area. This effect will be explained in more detail next with reference to Figure 4.

Figure 4 is a polar plot of the respective antenna gains in the horizontal plane of each access point in the underlay cell 20 when overlaid by those of the overlay cell 40. More particularly, the polar plot of Figure 4 comprises the plot of 12 antennas corresponding to 12 access points, six in each of the overlay cell 40 and underlay cell 20, the beam patterns of which are partially overlaid one on top of each other. More specifically, the beam patterns 1,2, 3,4, 5, and 6 (as more clearly depicted in Figure 3 relating to the prior art) are the respective antenna beam patterns corresponding to sectors 11,12, 13,14, 15, and 16 respectively in the underlay cell 20 of Figure 5. Overlaid on top of these beam patterns are the antenna beam patterns 21,22, 23,24, 25, and 26, which respectively correspond to the coverage sectors 41,42, 43,44, 45, and 46 of the overlay cell 40. From Figure 4 it will be seen that in the preferred embodiment being described the antenna beam patterns 21 to 26 of the overlay cell access point antennas are offset from the respective beam patterns of the access point antennas of the underlay cell by 30 , such that the respective boresight azimuths of the overlay cell access point antennas are directed towards the cusp regions of the underlay cell access point antenna beam patterns. As will be clear from Figure 4, this has the effect of reducing the antenna beam pattern cusps from, in the example shown, 2.3dB to only 0.5dB. This reduction in antenna cusps is manifested by an improvement in the link budget of radio links between access points and subscriber terminals located in the cusp regions as discussed previously, resulting in improved traffic carrying capacity on those links, and more a uniform outage probability in the entire coverage area.

Whereas Figure 4 shows the sectors of the overlay cell 40 being angularly offset from those of the underlay cell 20 by 30 , which accounts for half of the coverage area of the sector, it should be understood that the invention is not limited to such an angular offset and that other offsets can be used. The advantage

of using an offset equivalent to half the sector coverage angle is that the antenna cusps are minimised, whereas the use of any other offset with only a single overlay cell would result in the cusps being reduced, but not minimise.

It should also be understood that multiple overlay cells can be provided such that the original underlay cell 20 has two or more overlay cells 40 overlaid on top. Each overlay cell provides increased network capacity by approximately the equivalent capacity available to one cell, but at the cost of having to provide additional access points.

In addition to the above, where multiple overlay cells are provided, in the present invention the antenna beam of each access point in each cell is orientated to a different boresight azimuth in order to reduce antenna cusps. The preferred arrangement is to align the respective antennas of each access point so that the antenna boresight azimuths are distributed equally around 360 . However, in alternative embodiments it should be understood that equal distribution around 360'is not an essential element of the invention, and other distributions can be made provided that antenna cusps are reduced in areas where this is required.

In the preferred arrangement, however, generally where each cell is provided with n access points providing n sectors, and the total coverage of the

cell is x such that each sector covers-then where a number a of overlay n

cells are provided, the antenna boresight azimuths of the antennas in the base

stations of the overlay cells should be angularly offset by x-to the antenna (+1)

boresight azimuths of at least one of the other cells, which may itself to another overlay cell. By arranging the antenna azimuths of the overlay cells according to the above, antenna cusps can be minimised.

To provide for an increase in channel capacity there must be some method of providing for signal separation for those areas within the cell which are overlaid by two or more antenna patterns. Within the preferred embodiment which makes use of the United States UNII band this is achieved by assigning different transmission frequencies and/or transmission polarisations to those sectors which overlay each other. Therefore, turning back to Figure 5, it can be seen here that, for example, the overlay cell sector 43 has been assigned the horizontal channel of channel 4 of the UNII band, whereas the two sectors 13 and 14 of the underlay cell 20 which the sector 43 overlays have been respectively assigned the horizontal channel (H) of channel 1 of the band and the horizontal channel (H) of channel 9 of the band respectively. Thus, for example, it would be possible for a subscriber terminal or unit (SU) located in the right half of sector 43 which overlays the left half of sector 13 to access either access point 43 on channel 4 (H) or access access point 13 on channel l (H). Similarly, if two SUs are provided in the same area, then one could communicate with the access point of sector 43 on channel 4 (H) while the other communicates with the access point of sector 13 on channel 1 (H) without the two separate signals from each SU interfering with each other. As will be seen from Figure 5, channel allocations must be made for every sector of the underlay and overlay cells such that the access points of overlaid sectors do not transmit and receive on the same frequencies with the same polarisation. If enough channels are available, as is the case with the use of the UNII band and by doubling up the channels by use of polarisation, it becomes possible to ensure that such channel separation can be achieved. The highly directional antenna beam patterns also contribute towards achieving signal separation by spatial means.

Whilst the preferred embodiment makes use of different frequencies and polarisations to achieve signal separation and reduce co-channel interference, further embodiments could be envisaged which may make use of spread spectrum

techniques. Here, rather than transmitting on a different frequency channel and with different polarisations, each access point increases the band width of its signal by multiplying the signal by a high bandwidth spreading code, and then the signal is transmitted across the entire band. Signal separation is achieved by each access point transmitting with a different spreading code. At a subscriber terminal any particular access points'signal is retrieved from the spread signal by multiplying the signal with the spreading code used to spread the signal.

As another alternative, whereas the previously described preferred embodiment provides the overlay cells with the same number of sectors as the underlay cell, it is not essential that this is the case, and the overlay cell could be provided with a fewer or a greater number of sectors. The principal factor defining the coverage arc of an access point is the beam pattern of the antenna used with that access point, and hence where different antennas with different beam patterns are used then different sector patterns for the overlay cells can be achieved. For example, where three access points each covering 1200 are provided in the overlay cell then an increase in network capacity and a reduction of antenna cusps can be achieved, but without having to install a full set of six access points.

The same criteria as discussed previously with regard to the frequencies assigned to the overlay sectors such that they do not interfere with the existing underlay sectors continue to apply, however.

In addition to the advantages described previously with respect to increasing capacity and reducing antenna cusps, the use of an underlay/overlay architecture also allows the network to perform dynamic load balancing between access points in the case that one or more access points becomes overloaded.

Such dynamic load balancing is performed as follows.

When a subscriber unit (SU) (or Customer Process Equipment (CPE) ) is first powered up, the SU performs a channel scan to find available access points

before registering with the access point with the strongest signal (which provides the best radio link to an access point). The SU attempts registration by transmitting a registration message to the access point which provided the strongest signal, and the access point then passes the registration message to a Network Management System (NMS). Network Management Systems are of course known in the art, and in the present embodiment could be considered to be the equivalent of a base station controller (BSC) or mobile services switching centre (MSC) of a prior art GSM network. In the present invention, the Network Management System decides to which access point to allocate the registering SU based on centrally stored information within the Network Management System as to which access points all known subscriber units have already been registered. An available access point is allocated to an SU by the centrally located NMS with a view to balancing the loads between the available access points for that particular SU. Thus, in the example discussed earlier wherein an SU is in the region covered by sector 13 of underlay cell 20 and sector 43 of overlay cell 40, the NMS decides which of the access points providing sectors 13 or 43 an SU in that region registers with. If the NMS decides that the registering SU should register with another available access point to the one with which it is currently registered, then control signals are sent from the NMS via the access point to the SU, which then un-registers itself from the presently registered access point, and registers with the other of the access points indicated by the NMS via the control signals.

The above explanation deals with the situation when a subscriber unit is first powered up, but the same procedure can be followed to dynamically balance access point loading in each sector of a cell between layers. Here, as the Network Management System has centrally stored the information relating to with which access points all known subscriber units are currently registered with, it can

generate dynamic load balancing commands to be sent via the access points to the particular subscriber units to command them to re-register with another available access point when the present access point becomes overloaded.

The previous description relating to the first embodiment of the invention and modifications thereto has dealt solely with the invention applied within a single cell. However, in use, the present invention can be used to increase capacity across the entire network, in which case it must be applicable to every cell in the network. A second embodiment illustrating how this can be achieved will be described next with reference to Figure 6.

Figure 6 illustrates how it is possible to assign the frequencies and polarisations to each sector of a cell so that cells can be tiled together to cover a larger area without adjacent cells using the same frequency and/or polarisation.

More particularly, in Figure 6 a tile of four cells is used to provide an underlay layer 62 of a cellular network. The same frequency patterns as used in the underlay layer tile 62 can then be used in a tile of four cells to provide an overlay layer tile 64, but with the difference that the antenna beam patterns have been offset by an angle equal to half the angle of the cell sector. However, as should be apparent from Figure 6, it is not possible to directly overlay the overlay tile 64 onto the underlay tile 62 as then parts of overlay sectors would be coloured with the same frequency and polarisation as corresponding parts of corresponding underlay sectors. Therefore, in order to get round this problem the overlay layer 64 is offset by one cell to the right with respect to the underlay layer such that no two cells with the same frequency colouring pattern applied to their respective sectors are overlaid on top of each other. Therefore, as should become apparent by the dotted lines of Figure 6, cell 641 of the overlay layer tile 64 is overlaid on top of cell 621 of the underlay layer tile 62, and cell 642 of the overlay layer is overlaid on top of cell 622 of the underlay layer.

It will also be seen that cells 641 and 642 are provided with the same respective frequency covering patterns in their sectors as cells 623 and 624 of the underlay layer, whereas cells 643 and 644 of the overlay layer are respectively provided with the same frequency covering sector patterns as cells 622 and 621 of the underlay layer. In view of this, it should be apparent that it is possible to tile a plurality of four-cell tiles together in any one layer to cover a larger area. By doing so, the cells 623 and 624 of the underlay layer 62 would be overlaid by the corresponding cells two cells 643 and 644, but belonging to another four cell tile tiled next to the shown overlay layer tile 64. By tiling four-cell tiles together as required and horizontally offsetting the layers as shown, it is possible to increase the capacity across an entire metropolitan network using the present invention.

Claims (9)

CLAIMS :

1. A wireless cellular communications network comprising one or more cells, each cell having a first plurality of base stations each respectively arranged to provide a different coverage area each base station coverage area being defined by an antenna beam pattern and boresight azimuth of that base station, the network being characterised by each cell being provided with further base stations each of which provides a coverage area each further base station coverage area being defined by an antenna beam pattern and boresight azimuth of that further base station, wherein each further base station coverage area at least partially overlays one or more of the coverage areas provided by the first base stations, the further base stations being further arranged such that their respective antenna boresight azimuths are different from those of any of the first plurality of base stations.

2. A wireless cellular communications network according to claim 1, wherein each base station whose coverage area at least partially overlays that of another base station is arranged to transmit and receive signals on a different frequency and/or with a different polarisation to the other base station.

3. A wireless cellular communications network according to claims 1 or 2, wherein the respective coverage area provided by one or more of the further base stations at least partially overlays one or more coverage areas provided by other of the further base stations.

4. A wireless communications network according to any of the preceding claims, wherein each base station is mounted on the same mast or other structure.

5. A wireless communications network according to any of the preceding claims, wherein the coverage areas provided by each base station are of substantially the same size as each other.

6. A wireless communications network according to any of the preceding

claims, wherein said first pluralit claims, wherein said first plurality of base stations constitutes a first set of n base stations, and wherein said further base stations are arranged into a further sets of n stations, wherein a > : l and n > 2 a and n being whole numbers, each set of stations covering in total an area of sector x and each base station in each set having an

antenna beamwidth of- , the network being further arranged such that the n

antenna boresight azimuths of the base stations in one of said sets are angularly

je offset by (x) 0 to the antenna boresight azimuths of at least one other set of (a + I) n

base stations.

7. A wireless communications network according to any of the preceding claims and further comprising a network server arranged to control with which base station in a cell a subscriber terminal communicates when the terminal is in use.

8. A wireless communications network according to any of the preceding ing claims, wherein each base station is further arranged to communicate with a plurality of subscriber terminals geographically disposed within a base station's coverage area in a time division multiple access (TDMA) manner.

9. A wireless communications network substantially as hereinbefore described and/or with reference to the accompanying drawings.