GB2444996A - Inter-Relay Interference Avoidance in a Wireless Communications Network - Google Patents

Abstract

A method and network for providing inter-relay interference (IRI) avoidance in a cellular communications network comprising at least one base station defining a cell and a plurality of relay stations associated with the base station. The interference avoidance method establishes at least one threshold in at least one performance criterion (e.g. signal strength, propensity to cause interference, etc.) for communication between mobile stations and relay stations in said cell. The performance threshold(s) is/are used to define transmission groups and each mobile station is allocated to a transmission group on the basis of its performance against said threshold(s). Links for simultaneous communication are then selected on the basis of membership of said transmission groups. This selection preferably involves determining, for a desired system performance (e.g. high throughput, fair transmission, etc.), transmission groups that can compete or operate at the same time, such that mobile stations belonging to said groups can communicate simultaneously. For example, if high throughput is desired for two adjacent relay station zones, a transmission group based on high power and low interference may be allowed to operate in one relay station zone at the same time as a transmission group based on high power and low interference operates in an adjacent relay station zone.

Description

M&C Folio: GBP95391 1 2444996 The present invention relates to a

wireless communications system. It is particularly, but not exclusively, concerned with a wireless communications system involving a base stations, relay stations and mobile stations. This may include, but is not limited to, mobile cellular network systems.

In traditional mobile cellular network systems, a cell is defined by the extent of coverage of a base station, primarily in terms of the limits of the range of mobile stations from the base station. A cell is thus defined as a region extending substantially omnidirectionally around the base station. No distinction is normally made between parts of the cell and thus the position of mobile stations in the cell. Indeed, the base station is operable to radiate a signal in all directions within the cell in order to provide radio coverage for possible reception by mobile stations. As such, the base station does not, in conventional arrangements, use information on the position of mobile stations within the cell -the base station is only concerned with the presence, or otherwise, of a mobile station in the cell.

In transmission from a base station, to a mobile station, much transmission power is wasted because the base station in such circumstances has no knowledge of the position of the mobile unit for which the transmission is intended. The base station may be expending transmission power to provide radio coverage to parts of the cell may not contain a mobile station. Likewise, in reception, the base station antenna receives signals coming from all directions, whether or not those signals actually emanate from mobile stations. In directions not associated with mobile stations, either through direct transmission to the base station or through reflection from other features, such signals will be dominated by noise and interference signals, which can lead to deleterious performance.

Space division multiple access (SDMA) is a class of technologies developed to overcome these issues. Smart antenna technology at the base station enables detection of at least the angular position of mobile stations; this position information can be used to direct transmission and reception power as effectively as possible. It will be understood that, with this information, a base station equipeed with smart antenna technology can focus transmission power on directions associated with mobile stations, and can inherently divert attention from receive directions which would otherwise cause deterioration of receved signal quality.

Despite the use of sector/adaptive arrays at the base station, multi-user transmission with SDMA results in multi-user interference, due to application environment.

Conventionally, multi-user interference cancellation (IC) techniques have been widely employed. SDMA concepts incorporating effective multi-user IC and multi-user detection techniques have been addressed in a variety of publications. More specifically, detection techniques typically employed in CDMA systems are well known, such as: Minimum Mean-Square Error (MMSE) combining, Successive Interference Cancellation (SIC), Parallel Interference Cancellation, Maximum Likelihood Detection (MLD) and Maximum Likelihood Sequence Estimation (MLSE).

"Reduced Complexity Space Division Multiplexing Receivers" (G. Awater, VTC2000, Tokyo, Japan) describes MLD and MIMSE in detail; likewise, other techniques are described in "Low Complexity Space-Frequency MLSE for Multi-User COFDM," (M.

Speth, A. Senst, and H. Meyr, Globecom99, Rio de Janeiro, Brazil) and "Adaptive Antenna Arrays for OFDM Systems with Cochannel Interference," (Y. Li and N. R. Sollenberger, IEEE Transactions on Communications, vol. 47, pp. 217-229, Feb 1999).

However, applying interference cancellation for each mobile station may not guarantee performance convergence and hence the system may not be stable, particularly in the case where no more spatial freedom is available for extra interference. Consequently, the quality of service, for the multi-user system as a whole, is not guaranteed.

In contrast, interference avoidance (IA) algorithms can be more suitable in terms of system stability. They also outperform interference averaging techniques by a factor of 2-3 in spectrum efficiency, as explained in "Advanced cellular internet service (AdS)," (Cimini, Chuang, and Sollenberger, IEEE Comms. Magazine, pp. 150-159, Oct.1998) and "A novel broadband wireless OFDMA scheme for downlink in cellular communications" (J. Li, et al, WCNC 2003).

However, conventional IA also only considers maximising the SINR (signal to interference plus noise ratio) by assigning different frequencies and/or sub-channels to different users. Even using IA, system instability is inevitable.

In recent years, network interference avoidance has attracted considerable interest especially in networks employing radio resource sharing. For a relaying system, it is possible to support multi-user transmission simply by applying the concept of SDMA (Spatial Division Multiple Access), due to the large separation of users in space. For high efficiency, it is necessary to consider radio resource sharing between Relay Stations (RSs) and a Base Station (BS). In an ideal system design, no extra radio resource is imposed by insertion of an additional RS into the network. However, as with any multi-user transmission, multi-user interference still has the capability to seriously limit system performance.

Figure 1 illustrates an example of a portion of a cellular network to illustrate problems encountered in the prior art. Figure 1 illustrates a base station BS with its respective cell 10 highlighted among other cells.

Figure 2 illustrates the cell 10 in further detail. The cell 10 comprises at its centre a base station having an area of effective transmission and reception coverage which defines the extent of cell. The cell is illustrated as hexagonal, primarily to illustrate the tessellation of cells. In practice, cells could be circular or any other shape defined by the directionality of the antennas at the base station BS.

Relay Stations (RS) are located around the base station BS, within a base station cell, to enhance the coverage in the outlying regions of the cell. Relay Stations RS comprise antennas which are generally directional, and aligned away from their respective BS, thereby transmitting to and receiving from regions outlying the RSs with respect to the BS.

The overall cellular network is controlled by means of radio network controllers (RNC) located in each base station. The RNCs control radio transmitters and receivers in node base stations, and perform other radio access and link maintenance functions, such as soft handoff, in a 3G wireless network. An RNC is similar in function to a base station controller (BSC) as used in some technologies. In such systems, a BSC is part of the wireless systems infrastructure and controls radio signals in one or more cells, thus reducing loading on switches. It performs radio signal management functions for base stations, managing functions such as frequency assignment and handoff.

In general, each cell in the cellular system can be separated to an acceptable level from other cells by appropriate cell planning. Cell planning can include selection of optimal locations for Base Stations, and code or frequency assignment, depending on the technology used. Accordingly, in conventional systems, each cell operation can be relatively independent. However, intercell interference (ICI) arises and, to maintain cell independence, interference cancellation must be used at both base stations and mobile stations.

The arrangement illustrated in Figure 2 essentially comprises a relaying system within a cell. Each relay station can be treated as a small cell system. Conventional interference cancellation techniques can therefore be applied. However, for the limited radio resource associated with a relay station, interference is substantially more significant.

This imposes a heavy demand on conventional interference cancellation.

This is illustrated in further detail in Figure 1. In Figure 1, one mobile station (MS) is illustrated as an example. It will be appreciated that further mobile stations will normally be in use in a cellular network. The mobile station is shown in Figure 1 receiving intercell interference (IC!) from two other cells, namely 1 and Ij,2. In Figure 2, these two intercell interferences are also supplemented by inter-relaying interferences (IRI). These LRIs arise from Relay Stations RS_1 and RS_2 which are further away from the mobile station concerned than the Relay Station RS which is in use. This example implies that all relay stations use the same frequency for transmission to their mobile stations. The example shows that, without relaying, there are only two interferences to take account of(Ic,l arid 1ci 2), whereas, with relaying, two additional interferences, namely the IRIs, also need to be taken into account.

Clearly, the mobile station has substantially more interference cancellation to execute compared with conventional cellular architecture. This could be computationally difficult.

In general terms, aspects of the invention involve separation of consideration of ICI from WI. The approach taken in aspects of the invention is to group transmission with interference avoidance, to reduce the impact of interference from other RSs within the cell. However, the techniques proposed can be also extended to relaying outside a cell.

According to a first aspect of the invention, there is provided a method of controlling a cellular communications network comprising at least one base station defining a cell and a plurality of relay stations being associated with the base station or with one of the base stations as the case may be, the method comprising establishing at least one threshold in at least one performance criterion for communication between mobile stations and relay stations in said cell, said performance thresholds defining transmission groups, allocating each mobile station to a transmission group on the basis of performance against said thresholds, and selecting links for simultaneous communication on the basis of membership of said transmission groups.

A second aspect of the invention provides a cellular communications network comprising a plurality of base stations defining a cell and a plurality of relay stations being associated with each base station, the network being operable to define at least one threshold in at least one performance criterion for communication between mobile stations and relay stations in said cell, said performance thresholds defining transmission groups, each mobile station being allocated to a transmission group on the basis of performance against said thresholds, and means for selecting links for simultaneous communication on the basis of membership of said transmission groups.

The method of the first aspect can be directed by a base station, by a mobile station, or by a relay station, or in a distributed fashion.

Examples of the operation of the invention will now be described with the assistance of the accompanying drawings, in which: Figure 1 illustrates a cellular mobile communications network in accordance with a

prior art example;

Figure 2 illustrates a cell of the network illustrated in figure 1; Figure 3 illustrates a cell of a network in accordance with a specific embodiment of the invention; Figure 4 illustrates the cell of figure 3 at another stage of operation; Figure 5 illustrates uplink interference avoidance in the cell of figures 3 and 4; Figure 6 illustrates downlink interference avoidance in the cell of figures 3 and 4; Figure 7 illustrates operation of a specific embodiment involving a self certification mobile station; Figure 8 illustrates a conventional arrangement illustrating a scenario to be improved; Figure 9 illustrates operation of a specific embodiment of the invention in the arrangement illustrated in figure 8; Figure 10 illustrates a typical fixed relaying system in illustration of the field of the invention; Figure ii illustrates an example of an architecture of a single cell of the cellular network illustrated in figure 10; Figures 12 to 17 illustrate graphs of performance of the system in accordance with three scenarios; and Figure 18 illustrates graphs of probability distribution functions for conventional systems and a system in accordance with the specific embodiment, for comparison purposes.

For simplicity, an example is described herein with merely two relay stations (RS) in the cell, as illustrated in figures 3 (uplink) and 4 (downhink). For efficient use of radio resource, the two RSs use the same transmission frequency. Therefore, interference will be encountered at these two RSs during up-link (UP) communication and at their respective MSs during down-link (DL) communication. This is illustrated in figures 3 and 4, where each relay station is equipped with a directional antenna for its specific coverage zone, namely an RS-zone, in accordance with a conventional arrangement.

It will be appreciated that this is similar to a conventional relaying system. In order to have IRT-free transmission, it will be necessary to limit operation such that the two RSs are not operated simultaneously. However, this will halve the resource efficiency. For higher efficiency, it is highly desirable to operate the two RSs simultaneously. In this case, if each RS operates in accordance with its own performance criteria, without reference to performance of other RSs, the operation of the overall system may not

always be stable.

For a stable and efficient relaying system, interference avoidance is acknowledged to be a better approach than interference cancellation. However, interference cancellation can be operated by a single device, whereas interference avoidance requires system level operation.

The presently described example of the invention involves grouping transmission to categorise all MSs covered by an RS into several small groups according to their own signal strength at the BS (i.e. high or low power) and their propensity to cause interference to other RS coverage areas (i.e. high or low interference).

Group I: High-power-High-interference (G 1 -HpHi); Group 2: High-power-Low-interference (G2-HpLi); Group 3: Low-power-Lowinterference (G3-LpLi); Group 4: Low-power-High-interference (G4-LpHi).

In order to establish these groups in a particular RS zone, each RS-zone is able to define two thresholds, for signal and interference respectively. These two thresholds can be pre-set according to the characterisation of propagation in the RS-zone. More thresholds make more precise categorisation but may result in a more complicated implementation.

By grouping transmission, the network can provide high efficiency and a high performance service. In particular, it can offer benefit to the following applications as examples. In the notation below, the symbol -* denotes groups which are permitted to operate in a competing manner, recognising that competition with other groups would generate a high risk of interference. This is therefore an application of an interference avoidance technique, rather than an attempt to reduce or negate interference.

This accounts for conflicts between competing groups which, in the context of the system requirement, will cause performance deterioration, but does not involve arbitrary application of interference avoidance in circumstances which, in reality, will not generally cause such problems.

In a first example, it is desired to guarantee QoS at Cell Edge, where the Cell Edge is defined with reference to low received MS power at the BS. Therefore, it is necessary to avoid pairing groups together which will cause one group to impart high interference on another, otherwise low interference, group: RSI-zone: G3-LpLi -RS2-zone: G3-LpLi RSI-zone: G2-HpLi RS2-zone: G3-LpLi RS 1-zone: G3-LpLi -RS2-zone: G2-HpLi For the first example, in which a guaranteed Cell Edge QoS is assigned, the competing MSs from both RS-zones are competing with low interference since an MS positioned at a cell edge always has low received power. The principle of this application is to avoid unfair competition, such as low power to high interference. Therefore, Group I of any particular RS-zone, which experiences high power but high interference, is not permitted to compete with another group as this would caused inter-group interference.

Likewise, Group 4 of any particular RS-zone is also subject to high interference so no inter-group competition is permitted for this group either.

In a second example, high throughput is desired for both RS-zones: RS1-zone: G2-HpLi *-* RS2-zone: G2-HpLi In this example, only Group 2 MSs are permitted to compete as any further competition will have a deleterious effect on throughput. The fact that Group 2 MSs are able to operate with high power and low interference means that they are likely to be the least affected by increased competition for the medium, and so grouping of Group 2 from one RS-zone with Group 2 of another RS-zone will, in accordance with this example, be likely to improve system performance.

In a third example, high throughput is desired for a specific user: RSI-zone: G2-HpLi *-* RS2-zone: G2-HpLi RSI-zone: G2-HpLi -+ RS2-zone: G3-LpLi RS1-zone: G3-LpLi RS2-zone: G2-HpLi RS 1-zone: G1- HpHi 4-* RS2-zone: G2-HpLi RS1-zone: G2-HpLi RS2-zone: Gl-HpHi This third example takes advantage of similarities in characteristics between MSs of different RS-zones, allowing them to compete where possible, being mindful of avoiding unfair competition' such as involving low power, high interference devices -such devices would be too sensitive to be able to compete successfully with a device of another group and need to be provided with sufficient resources to be able to communicate without additional competition.

In a fourth example, a fair transmission regime is required: RS1-zone: G1-HpHi RS2-zone: G1-J-IpHi RSI-zone: G3-LpLi RS2-zone: G3-LpLi RS 1-zone: G4-LpHi RS2-zone: G2-HpLi In the fourth example, the fairness is exhibited in pairing groups with similar properties -allowing devices to compete where there would be no inherent likelihood of devices of one RS-zone always successfully communicating at the expense of devices of another RS-zone.

In order to make available this approach to grouping transmission with interference avoidance (GTIA), a new transmission architecture and system operation are described.

As already discussed previously, it is important to distinguish between ICI and IRI. In an RS-zone, the received signal of either the RS and a MS can be expressed as YX+!JR,+IIC,+fl, (1) where x is the transmitted signal, IRI is the interference from other RS zones, !,., is the interference from other cells, and n represents noise. The noise level is normally fixed, and is mainly due to the receiver architecture.

As already mentioned, it is difficult to control inter-cell interference in any network. In accordance with the specific embodiment of the invention, efficient avoidance of inter-relaying interference IRI is used in order to achieve a required quality of service. Consequently, it is essential to detect and control the IRJ* A process to achieve this can be operated by either a RS or a MS. Both modes are illustrated below as their precise implementations can differ in terms of architecture and operation.

In any event, it is firstly assumed that power detection and interference detection are possible to any BS, RSs and MSs.

in a first mode of operation, i,, detection and control is performed at the RS. At each RS, if all user in the cell are transparent to the RS, which is possible for specific network setup, it is possible for the RS to detect the IRI and to group users in its RS-zone for interference avoidance operation. There are two difference cases of this mode of operation: In the up-link case, before transmission, the mobile station should firstly communicate with BS and/or RSs for routing purposes, as in any traditional networking. The up-link case is shown in figure 5, in which the MS and its routing have been selected, after the initial communication, as MS2-RS2-)BS. After this routing has been resolved, it is important for the RS2 to determine if a link with MS2 can be made. Consequently, RS2 needs to collect all information and status of MS2, including its received power at the RS2 (r) and its interference to other RSs (s3 -+ 1R13 S1 -+ /ffl).

The received power (F) detection is straightforward, as it can be detected directly by the RS2. However, the interference IRII and!R13' as shown in the figure, cannot be detected by the RS2 but can be detected by RS I and RS3 respectively. Therefore, links with the relevant relay stations are required, such as RS1 -* BS & BS - RS2 or the direct link RS1 - RS3, for delivery of the information describing /RJ) and i,.

Having been furnished with F,, and JRJI & 11R13' it is now possible for the RS2 to categorise the MS2 into a group.

In the down link case, as shown in figure 6, the RS2 needs to detect the received power i, and the interference, & 1,. Then, the MS2 needs to feed the information back to the RS2. Clearly, it is possible to detect and control lip, at the RSs. The advantage of this is its easy control mechanism -all user grouping and transmission are through the RSs andlor the BS. However, links between BS and RSs are required for the up-link and feedback is required for the down-link.

In a second mode of operation, ip! detection and control is performed at an MS. Such an MS is therefore called a self-certificate mobile station. This is illustrated in figure 7.

In this case, each MS, such as MS2, should be able to detect the received signal strength p, at the MS2 and the interference i, and l, from RS1 and RS3.

In this example, it is assumed that the MS2 has knowledge of the signal power and interference thresholds defined for the RS-zone; the MS2 is thus able to categorise itself into a group, as a self-certificate mobile station.

This self categorisation can be suitable for both down-link and up-link, although it will be appreciated that, in the down-link case, operation is most straightforward as the MS will directly detect the power and interference (RS-MS), and that in the uplink case the information gathered may relate to an uplink transmission at a different time or in relation to a different frequency. For an uplink transmission wherein the information relates to transmission at a different frequency but at the same time, self categorisation presents little difficulty. For transmission in a different time slot, the MS should perform the categorisation just before making the transmission. The MS should actively monitor other transmissions in order to detect power and interference. By means of the active detection, the MS can detect the required information based on another user's transmission. However, if the MS cannot derive sufficient information before its transmission, it is preferable for the MS to communicate first with the RSs.

With the above two cases of jp! detection and control at either RS or MS, one common technique is the i,, detection. The!RJ needs to be detected prior to any operations.

However, this detection should be operated together with a routing and scheduling procedure. This is because each MS needs to know if its transmission is going to be relayed via a RS or directly with the BS. Furthermore, each MS needs to know which RS is the best for its transmission. This naturally falls within the scope of i, detection.

Any conventional interference detection techniques can be applied if appropriate.

However, a specific detection technique will now be described which is especially appropriate for the relaying system presented in this report.

The fundamental idea is to attach a noise-level long periodical sequence to the desired transmission data packet. With this specific sequence, it is possible to detect the relevant signal power. This also distinguishes the signal power from the ICI since the long sequence should become orthogonal to ICS. Assuming the sequence is s, equation (1) can be rewritten as: Y X+Sq +i,., +71, (2) As already mentioned, s, is the noise-level sequence and constitutes a very low power signal. However, correlation at the receiver can abstract its energy, in a similar manner to the concept used in CDMA. Also, the symbols could be unevenly distributed within the sequence. Some parts of the sequence may have higher power than other part of the sequence. This depends on signal sequence. For instance, the synchronisation part of the signal is more robust to other parts of the signal since it operates correlation at the receiver. Therefore, it is possible to make the noise-level sequence, s, higher power during this part.

Also, the added sequence s, is known to the receiver, and consequently it is possible to cancel the s, (if treated as a known interference) by simple interference cancellation, such as conventional SIC or PlC.

Extending this basic idea, different sequences can be applied to BS, RSs and MSs. For simple operation, it is possible to apply different sequences to BS and each RS. For MSs, the same sequence should be applied as their respective RSs, so that RSs can detect their MSs' power strength and also measure interference from other RS-zones.

Essentially, this compares with applying respective colours' to zones, including BS-zone and RS-zones. It should be borne in mind that the insertion of s is only to enable another zone to detect interference. However, it is still possible for the sequence to be used by its own zone for any purpose, such as aiding synchronisation, frequency offset detection, etc. Another i, detection technique disclosed herein is provided to block a known signal for interference detection, which is named herein double-blockage interference detection.

This blockage involves only a small time period in the signal by using a specifically designed correlation sequence. By equation (1), signal x is firstly blocked, then the equation becomes (3) Then at the receiver, the received signal is simply all elements of the interference plus noise.

Secondly, (RI can be blocked since this interference is actually the signal from other RSs within the cell and is thus known, in contrast to the unknown i. Then, equation (3) becomes (4) By equations (3) and (4), interference IRI can easily be estimated.

In general, the process of the grouping transmission with interference avoidance can summarised as follows: * There are several relay stations (RS) in a cell of a cellular architecture and each RS has its own coverage area, which is called RS-zone; * All the RSs are employing the same frequency and perform transmission simultaneously; * Each RS has knowledge of its coverage including users' power distribution arid interference caused to other RSs by the users. This kind of knowledge can be pre-measured by the network when it established; * Each RS can then set up certain thresholds for signal strength and interference power; * The thresholds can be one for signal and one for interference. Also, the thresholds can be several for signal and several for interference. More levels of threshold makes more precise categories but the operation could be more complex; * Based on the pre-set thresholds, each RS can set up rules for grouping users, such as grouping the users with high signal power and low interference, or the users with low signal strength and low interference, etc. * The grouped users means those users falling into the same categorisation under the same threshold; * Then the RSs can negotiate with each other, or pre-set a set of rules of transmission according to its supported service and the targeting quality of service, such as guaranteed cell-edge QoS, * For those RSs who can negotiate each other, each RS might need to have communication links with other RSs or through BS; * Based on the grouping rules and the transmission rules, each RSzone can operate transmission with its grouped users. * For the case of fIR! detection and control at the RS, the RS needs to

communicate with each MS to select the appropriate users as already described.

* For the case of i, detection and control at the MS (self-certificate mobile station), each MS capable of determining if it is time for it to request transmission with the RS. If not, the MS will keep quiet without requesting any transmission. Then the RS will only select some of those users who are requesting transmission at the time.

The process described above eventually avoids the interference by means of avoiding unfair competition, especially for those users with low signal power to compete high interference, which will obviously result in low quality of service. It needs to be noted here that the interference is not eliminated by the proposed scheme. However, with the proposed network operation, it is much easier to achieve high stability, high efficiency, high performance and guaranteed QoS, as compared to the conventional networking for relaying transmission.

In addition, there are several different manners of operation of this interference avoidance. Here two possible approaches are described: The first one is BSIRS controlled rules. In this operation BS and RSs set up the rules as discussed previously. In the case of self-certificate MS, each MS enters a RS-zorie and receives the rule for operation and operated accordingly. In this operation, RS can be only the amplifier-and-forward relay station. Otherwise, for RS controlled interference detection, RS should have more functions.

The second one is master-slave rule. In this operation, some RSs act as masters and some RSs act as slaves. In general, adjacent two RSs should act as one master and one slave. The master RSs can operate independently but the master RSs have to broadcast their rules of operation. In contrast, the slave RSs should operate according to its adjacent master RS(s).

Now, the network architecture of the specific embodiments and its operation are briefly summarised, with three different scenarios described previously. Firstly, the general exemplifying relay-networking is summarised for all three scenarios: 1) The network is a cellular system with multiple cells and inter-cell interference; 2) For each cell, there is one BS and several RSs to support MSs; 3) The same radio resource is applied to the BS and all RSs; 4) The BS and RSs are equipped with their own antenna which covers a specific zone; 5) The network needs handover to handle the mobility of a MS from one zone to another zone; 6) The behaviour' of a device, a term which will be used in due course, includes the received power at the device, which will include received power resultant from its own transmission, and any power received from other resources with the same frequency/time characteristics; 7) A fully loaded zone means that a zone is fully filled with MSs; 8) A fully loaded zone's signal strength distribution can be predictedlmeasured/pre-set; 9) A fully loaded zone's interference-caused distribution can be predicted, measured or pre-set; 10) Threshold(s) for signal-strength are set up based on the zone's signal strength distribution; 11) Threshold(s) for interference are set up based on the zone's interference distribution; Scenario 1: Central-control network 1) BS and RSs detect all MSs' behaviours in their zones; 2) BS and RSs categorise their users into different groups in accordance with the thresholds; 3) BS and RSs set up rules for transmission; 4) The rules should be set up to avoid unfair competition, such as low signal strength competing with high interference; 5) The rules have to consider fair transmission with fair competition, such as high signal strength competing with high interference; 6) BS and RSs allocate their MS(s) according to the rules; 7) If the number of MSs in a zone is greater than one, then the MSs should be in the same group, in order to group transmission; 8) In addition, since this is a centralised network, BS and RSs should be able to negotiate if they have links between them; 9) Negotiation operation is more suitable to guarantee Q0S; 10) Negotiation operation is also suitable to support a continuous transmission without disturbance or with less disturbance from other MS(s); Scenario-2: Self-certificate MS(s) 1) BS and RSs only set up transmission rules and the rules are the same as those in scenario 1; 2) Each MS detects its behaviours in the zone being specified; 3) Each MS categorises itself into a group upon the thresholds; 4) Each MS should be able to obtain transmission rules; 5) Upon the transmission rules, only those MSs who meet the requirement will send out a request for transmission; 6) BS and RSs will allocate their resources amongst MSs who request transmission; 7) In addition, each MS should be able to detect its behaviours from other MS(s)'s transmission; 8) This self-certificate MS(s) scheme is more suitable to packet-transmission; 9) Each MS's detection is operated on each transmission frame; Scenario-3: Master-slave 1) BS and all RSs are alternatively acting as master(s) or slave(s); 2) For master BSIRSs, itithey can operate in any way to fit its/their own self-interests; 3) Each master has to broadcast their rules in operation; 4) All slaves have to obtain their master(s)'s rules; 5) All MSs in the slaves' zones should be able to determine their behaviours in the zones; 6) The behaviour's determination can be operated as that in either scenario-I or scenario-2; 7) Only those slave-MSs meeting rules operated by master-zones can request or be made for transmission; 8) In addition, the impact of slaves should be included in master's rules.

A relaying system in accordance with these specific examples can support multi-user transmission in combination with SDMA. In order to achieve high efficiency of radio resource, RS-zones with the same radio resource, for example the same frequency, will cause serious IRJ, which will limit the system performance. Conventional random user selection cannot provide guaranteed Q0S and may also cause system instability, or non convergence. Consequently, a conventional system normally requires complex interference cancellation techniques to recover the Q0S. However, with a relaying system, heavy emphasis is placed on interference cancellation. A typical conventional application is demonstrated in figure 8.

As shown in figure 8, there are many users in a RS-zone. Conventionally it is possible to choose two MSs such that both RSs will transmit. Unless each MS is equipped with a powerful interference cancellation engine, both MS will fail in transmission or achieve low performance.

The described embodiment of the present invention provides an amelioration of this as shown in figure 9. In figure 9, it is clearly shown that the situation which could be occurred in a conventional system can be avoided by use of the described grouping transmission architecture. Instead, a reasonable user configuration can be achieved as shown in the figure that a user with low signal power should only compete with another user producing low interference. This configuration successfully guarantees the QoS of both users.

It is noted here that this approach is more suitable for packet data transmission.

The major advantage of the grouping transmission is to group users for transmission to possibly maximise the quality of service (Q0S) of a relaying system. This approach efficiently avoids interference especially to those subjected otherwise to unfair competition, such as to the guaranteed cell-edge QoS, by designating a group with low signal power as a group with low interference power.

Besides, it is more suitable for a relay system combined with the SDMA concept to support multi-user transmission, which makes a relay system more possible for high efficient radio resource reuse.

Although the described interference detection technique is especially appropriate for grouping transmission, it could easily be extended to other transmission applications.

The core concept of the interference detection is to distinguish two or more different interference sources.

The concept of grouping transmission can also be extended to a normal cellular system if the cellular architecture and approach can be further controlled by the BSC.

Furthermore, this is even more effective for the proposed self-certificate MS. For this case, RSs can be either amplif'-forward only relaying or full functional relay station (it is well known that the full functional RS is complex, and sometimes is similar to a BS).

Finally it will be appreciated that the grouping transmission technique can be employed not only between RSs but also between RSs and BS if their coverage zones are using the same radio resource, e.g., same frequency, same time, etc. This approach is suitable for all relay systems including fixed relay and mobile relay.

For simplicity, the examples given below are mainly in fixed relaying systems.

A typical fixed relaying system is shown in figure 10. Then, a possible architecture of a single cell of the cellular network can be drawn as in figure 11, where each RS equips a sector antenna to make coverage of RS-zone.

For grouping transmission, it is necessary to understand and analyse the power distribution and interference distribution. As an example, attention is focused on the adjacent two RS-zones (see figure 11). Three typical scenarios of these distributions are shown in figures 12 to 16.

The first scenario (scenario I) is that the system fixes the received power at BS for all usersfMSs. MSs can increase power as much as possible for targeting on fixed received power at an RS. This approach is ideal for a CDMA transmission system. It is shown in figure 12 that the transmit power at RSs is very high at the far end (edge of cell).

Also the MSs at the cell-edge produces high interference. Therefore, it is more important to apply the proposed grouping and its procedure to distinguish the received power and received interference for both zones to make efficient transmission.

The second scenario (scenario-Il) is that the system fixes the transmit power for all MSs, RSs and MSs. Performance in accordance with this scenario is illustrated in figures 14 and 15. The third scenario (scenario-ill) (figures 16 and 17) is that the system fixes the received power as a reference for power control and at the same time the transmit power is limited to a level to meet the requirement of EIRP (equivalent isotropically radiated power). EIRP represents the total effective transmit power of the radio, including gains that the antenna provides and losses from the antenna cable.

If an RS is provided which is intended to support a far end user (for instance a user device located at a cell edge), the RS will produce a large interference in another adjacent zone. Since the received power for all MSs in both zones will be the same, this means that the other zone will always have a low signal to interference ratio and performance will consequently remain low.

In a practical situation, not all users will receive the high interference. Therefore the grouping transmission approach set out in Scenario 2 can readily be adopted.

For comparison, it is first assumed that random user scheduling is applied as a conventional method and only one user is considered at a time. It is noted here that for a fully loaded zone, the total received power at the RS and the total interference have only a single value. With this assumption, the property of the zone signal-to-interference distribution and several typical values are measured and are listed in the

following table.

max-SIR (dB) mm-SIR (dB) optimum-SIR (dB) RS-zone (PL-order = 4) 101.8612 -5.5757 19.4682 RS-zone (PL-order = 2) 56.2524 -2.4998 12.3955 In the above table, PL-order stands for path-loss-order, max-SIR is measured by the possible maximum signal strength to possible minimum interference level, in contrast the mm-SIR is measured by the possible minimum signal strength to possible maximum interference level.

In the system, these two max-SIR' and mm-SIR' situations will happen with the conventional operation. min-SI1R is potentially serious, as Q0S cannot then be guaranteed. The reason for this as already discussed is due to the unfair competition.

For fair competition, the weak user should compete with small interference. Then, the optimum SIR is measured by means of determining the ratio of the minimum possible signal strength to the minimum possible interference strength.

Comparing the mm-SIR with the optimum SIR, there is a substantial difference, as more than 25dB for PL-order 4 and about 15dB for PL-order 2.

The invented grouping transmission can easily solve this problem and for comparison, the PDF (probability density function) for the one-user traditionally random scheduled and the PDF for the one-user with simple grouping transmission are produced. For grouping transmission, two thresholds are set: one is the average signal power and another is the average interference level. Therefore four group users are established.

The PDF comparison is shown in figure 18, wherein it will be seen that 10dB gain can be easily achieved by the simple grouping transmission for this specific example.

Claims (16)

1. CLAIMS: 1. A method of controlling a cellular communications network

   comprising at least one base station defining a cell and a plurality of relay stations being associated with the base station or with one of the base stations as the case may be, the method comprising establishing at least one threshold in at least one performance criterion for communication between mobile stations and relay stations in said cell, said performance thresholds defining transmission groups, allocating each mobile station to a transmission group on the basis of performance against said thresholds, and selecting links for simultaneous communication on the basis of membership of said transmission groups.
2. 2. A method in accordance with claim 1 and including pairing transmission groups either from within a cell or from adjacent cells, said pairs being designated to meet system performance need, and applying interference avoidance to communication involving mobile stations allocated to said groups.
3. 3. A method in accordance with claim 1 or claim 2 wherein said performance thresholds are defined in signal strength of mobile station signal measured at its respective base station and a mobile station's propensity to cause interference to other relay station coverage areas.
4. 4. A method in accordance with any preceding claim wherein said method is performed on a distributed basis across a base station, relay stations and mobile stations of a cell.
5. 5. A method in accordance with any preceding claim wherein said interference avoidance includes identifying inter-relaying interference as opposed to inter-cell interference.
6. 6. A method in accordance with claim 5 wherein said step of identifying inter-relaying interference comprises determining a signal transmitted by a relay station in said cell and isolating the effect of this signal on the signal received at another relay station in said cell.
7. 7. A method in accordance with any preceding claim wherein said interference avoidance comprises supplementing a communication event with additional information to enable a neighbouring relay station to detect inter-relaying interference at a relay station.
8. 8. A method in accordance with any preceding claim wherein said additional information comprises a noise level sequence suitable for correlation based detection at a receiver.
9. 9. A cellular communications network comprising a plurality of base stations defining a cell and a plurality of relay stations being associated with each base station, the network being operable to defme at least one threshold in at least one performance criterion for communication between mobile stations and relay stations in said cell, said performance thresholds defining transmission groups, each mobile station being allocated to a transmission group on the basis of perfonnance against said thresholds, and means for selecting links for simultaneous communication on the basis of membership of said transmission groups.
10. 10. A network in accordance with claim 9 and including means for pairing transmission groups either from within a cell or from adjacent cells, said pairs being designated to meet system performance need, and means for applying interference avoidance to communication involving mobile stations allocated to said groups.
11. 11. A network in accordance with claim 9 or claim 10 wherein said performance thresholds are defined in signal strength of mobile station signal measured at its respective base station and a mobile station's propensity to cause interference to other relay station coverage areas.
12. 12. A network in accordance with any of claims 9 to 11 wherein said interference avoidance means is operable to identify inter-relaying interference as opposed to inter-cell interference.
13. 13. A network in accordance with any of claims 9 to 12 wherein said interference avoidance means is operable to supplement a communication event with additional information to enable a neighbouring relay station to detect inter-relaying interference at a relay station.
14. 14. A mobile station operable to direct the method of any of claims Ito 8.
15. 15. A mobile station in accordance with claim 14, comprising means for receiving information defining performance thresholds, means for allocating said mobile station to a group, arid means for communicating with other stations in a network to establish interference avoidance.
16. 16. A base station operable to direct the method of any of claims ito 8, the base station being operable to manage interference avoidance at mobile stations and relay stations in its cell.