GB2332600A - Method of providing a matrix indicating interference relationships between cells of a cellular communication system - Google Patents

Abstract

A method of providing a matrix indicating interference relationships between cells of a cellular communication system comprises the steps of receiving measurement reports from mobile stations, the reports containing signal strength metric data of a serving cell and signal strength metric and identity data of a plurality of other cells, comparing the signal strength metrics of each other cell with those of the serving cell to provide a set of comparison values and determining the matrix from the comparison values and identity data. The step of comparing signal strengths involves subtracting the signal strength value N i of the other cell from the signal strength value S i of the serving cell. A number of S i -N i values are calculated for a particular other cell and a histogram is formed from the count occurrence of each value. The levels of occurrence are normalised and totalised within predetermined ranges of S-N value on the histogram to provide an interference coefficient for the particular other cell. The coefficients for all other cells form a matrix of values which is used for frequency planning.

Description

2332600 MEMOD OF PROVIDING A MATRIX INDICATING IN'ITA1FERENCE

RELATIONSE11PS BETWEEN CELLS OF A CELLULAR COM31LTMCATION SYSTEM The invention relates to a method of providing a matrix indicating interference relationships between cells of a cellular communication system.

BACKGROUND OF THE INVENTION

In a cellular communications system, the area over which service is provided is divided into a number of smaller areas called cells. Each cell is served from a base station which has a corresponding antenna or antennas for transmission to and reception from a user station, normally a mobile station.

A specific radio frequency or frequencies is assigned to each cell. Because only a limited spectrum is available and only a finite number of subdivisions within the limited spectrum are possible, certain restrictions arise. One restriction is that it becomes necessary to re-use a specific frequency in different cells of a system. A further restriction that arises is that adjacent frequency values have to be used that are closer in value to each other than might otherwise be desirable. Such frequency restrictions give rise to so-called co-channel interference and adjacent channel interference. Co-channel interference refers to the interference arising between two different cells that are using the same frequency as each other. Adjacent channel interference refers to the interference arising between two different cells when adjacent frequency sub-divisions of the available spectrum are being used in the respective cells. When planning frequency allocation of a system, one concern is how to reduce the likely levels of co-channel interference and adjacent channel interference.

A prior art method involves undertaking so-called propagation predictions. For a given cell, the total area is divided into yet smaller areas. For each said smaller area, an estimation is made of what relative signal levels would be received by a mobile station from base stations of different cells compared to the base station of the present cell. This procedure is carried out for a large number of cells. Based on the resulting estimates, frequencies are allocated throughout the system, or existing frequency 2 allocations are altered. This process involves the provision of so-called C/I matrices. Such first matrices contain CII values which are calculated from the above described propagation predictions. Value C refers to the carrier signal strength, i.e. that signal which it is estimated would be received from the base station of the cell under consideration. Value I refers to the interfering signal strength, i.e. that signal which it is estimated would be received from the base station of the other cell being considered in the respective propagation prediction.

In the provision of C/I matrices, propagation predictions are augmented by drive testing. Drive testing involves an operator literally driving around the coverage area of a cellular vehicle.

Drive routes. that is routes that are intended to be representative of the communications traffic contained in the cell under investigation, are chosen. The drive-testing method involves driving the chosen routes making multiple telephone calls in both the mobile-to-land and the landtomobile directions. The information from each call is recorded and postprocessed.

A disadvantage with both the basic propagation prediction method and the additional drive testing contribution is that all values used therein are merely unrepresentative samples or average values taken from specific locations. These locations are artificially chosen and hence do not represent a full cross-section of the real usage in the cell. Another disadvantage is that unless such procedures are frequently repeated, no account is taken of any changes to the system. Another disadvantage with the prior method in which C/I matrices are provided is that the values within such matrices constitute absolute values of the calculation CII.

Such absolute values therefore do not discriminate between favourable or less favourable frequency re-use candidates. In particular the values do not include any reflection of levels of subscriber usage, although such levels in reality play a role in the question of whether any frequency planning results in a successful frequency allocation within a system.

3 BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 is an illustration of a typical cellular communication system that may be used with the present invention. Figure 2 shows a representation of a neighbour list. Figure 3 is a process flow chart of an embodiment of the present invention. Figure 4 is an illustration of a histogram generated in the method of an embodiment of the present invention.

DESCRIPTION OF A PREFERRED EMBODIMENT OF THE INVENTION

The preferred embodiment relates to a cellular communication system which is a GSM system, although it will be appreciated that the invention is not limited to such a system and could equally be used in other time division multiple access (TDMA) systems, in code division multiple access (CDMA) systems, or other cellular communication systems including combined TDMA/CDMA systems.

FIG. 1 illustrates a typical cellular communication system 10 having a coverage area formed by a number of cells 12-22. A conventional schematic representation showing a hexagonal cell pattern has been used to depict the cell areas, but in practice the shape and size of each cell will be digerent. In each cell, often at the centre, there is located a base station 25 24-34 which controls communications traffic 36, 38, 40 in its cell in accordance with procedures known to one skilled in the art. Each base station (BS) 24 - 34 may receive 38 and transmit 40 signals from/to mobile stations (MS) 42 - 46 that move throughout the communicatibn system 10. In the preferred embodiment each of base stations 24, 26, 28 and 30 is 30 coupled to mobile services switching centre (MSC) 50 through one base station controller (BSC) 47 and each of base stations 32 and 34 is coupled to MSC 50 through another BSC 48. An operations and maintenance centre (OMC) 49 is also coupled to MSC 50. Subject to the size of the communications system, OMC 49 may be responsible for the whole system 4 or alternatively there may be a number of OMCs provided on a regional basis. MSC 50 is coupled to a public switched telephone network (PSTN) 51.

MS44 is assigned to BS 28 as its serving cell by means of the following procedure. When MS 44 is switched on, it scans, by processes well known in the art, in an attempt to receive signals from base stations. In this scanning stage the MS scans using all the frequencies of the system. MS 44 receives signals from BS 28 and, in this example BSs 24, 26, 30 32 and 34.

The MS determines which received signal is of the greatest strength, and then communicates to the corresponding BS, in the present case BS 28, that it wishes to be connected to the system. BS 28 undertakes a dialogue with MS44 to establish a communication path, and also at this stage interrogates the MS 44 to ascertain details from it including its identity.

The identity of MS 44 and other details, such as cryptographic information, are transmitted by the BS to the BSC 47, and on to the MSC 50.

The MSC 50 stores these details so that when the mobile station is paged it can route the call from the PSTN 51 via the MSC 50 to BSC 47, and on to BS 28, where communication between MS 44 and the system occurs. At this stage of the assignment procedure MS 44 becomes under the control of BS 28. Control information is transmitted from BS 28 to MS44. This control information includes inter-alia a control message indicating which subset of frequencies should be scanned from now on. The frequencies of the sub set correspond to the frequencies employed by the BSs on the current neighbour list of BS 28. The current neighbour list is explained more fully below. In addition, in the present embodiment, the control information includes an instruction that MS 44 should also scan some of the other frequencies used in the system. In the cominlini cation system, certain frequencies, i.e. frequency channels, may be dedicated to carrying control messages and certain other frequencies, i.e. frequency channels, may be dedicated to carrying communication traffic. One choice of which other frequencies are to be scanned is that of choosing to instruct the mobile station to scan all the frequencies that are used for control messages in the system, as opposed to the frequencies that are used for communication traffic. The outcome of the assignment procedure is that MS44 is assigned to BS 28 which is therefore termed its serving cell. Thereafter MS44 will communicate with BS 28 in response to a page or on a periodic basis to confirm to BS 28 that it is still operating normally.

MS 44 will remain assigned to BS 28 as its serving cell until it is either turned off, or handed over to another cell in the vicinity. Handover to another cell involves the current neighbour list, which is now explained in detail.

In the system of the present embodiment it is necessary to allocate, for each BS, those specific other BSs that will serve as candidate handover BSs. The list of identities of allocated handover BSs, and hence allocated handover cells, is called the neighbour list. In the system according to the present embodiment the neighbour list is stored at the base station. However, in other systems it could alternatively be stored at the BSC, the OMC, or indeed even remotely. The candidate handover BSs in the present example of BS 28 are chosen from those located in close geographical proximity to the cell area of BS 28, and in the absence of any other factors will be allocated on a simple basis of all those cells geographically surrounding it. It is noted however, that were other factors influential, such as geographical and construction features, e.g. obstructions, subscriber densities, subscriber movement patterns, and uneven sized or shaped cells, then it would be preferable to allocate cells on a less systematic basis. The resulting allocation arrangement of candidate handover relationships between the BSs (and hence cells) of the cellular communi cations system is termed the topology of the system. In the present embodiment each of the five cells shown schematically around cell 16 in FIG. 1, that is cells 12, 14y 18 20 and 22 are present on the current neighbour list of BS 28. Moreover, in the present embodiment, no further cells are in the current neighbour list. It is pointed out however that in general, in a given cellular communication system there will typically be a. urn number of cells that can be contained on the current neighbour list of a cell.

6 In the GSM system of the present embodiment each BS broadcasts on its control frequency continuously in the form of a broadcast control channel (BCCH). In the present GSM case the control channel frequency is in the form of the. absolute radio frequency channel nuinber (ARFCN). Encoded in the BCCH is a further BS identifier termed the BS identifier code (BSIC). The overall BCCH-ARMN and BSIC combination thus forms identity data relating to the identity of the base station, i.e. identity data relating to the identity of the cell.

As previously mentioned, MS44 scans the frequencies determined by the control message from BS 28, i.e. the frequencies contained in the socalled scan list of BS 28. By virtue of receiving the control frequency signals, MS44 determines a signal strength metric, that is a measure of the strength of the signal in the case of each received BCCH. It also decodes the BSIC.

Thus MS44 which is assigned to BS 28 as its serving cell acquires identity data and signal metric data in the form of three pieces of information from each BS, i.e. other cell, it has successfully scanned, namely the BSIC, the signal strength metric, and the respective control channel frequency. In the present GSM case the control channel frequency is in the form of an ARMN index. These three pieces of information are attached together in the form of a module of data. Respective modules of data determined from a plurality of different BSs, i.e. a plurality of cells other than the serving cell, are passed back to BS 28. In the present GSM embodiment, MS44 ranks the modules of data on the basis of the signal strength metric, and transmits only the six highest ranked modules of data to BS 28. These are transmitted to BS 28 in the form of a measurement report. Such a measurement report is fidly defined in ETSI GSM Specifications 4.08.

The signal metrics in the measurement report contain two values. One value pertains to transmissions that are continuous, and employs what is termed the full rate channel. The other pertains to transmissions that are intermittent but follow the input signal and periods of silence, and employs what is termed the sub-rate channel. The intermittent signalling is generally termed discontinuous transmission (DTX). A flag in the 7 measurement report indicates which value in the measurement report is valid. The modules of data are transmitted continuously.

Within BS 28 the BCCH-ARFCN and BSIC combiflation is converted to a BS identity, in this case BS 32. Such a conversion can only occur if the BCCHARMN and BSIC combination and BS identity are contained in the neighbour list. FIG. 2 schematically shows a representation of a neighbour list, from which it can be seen how a BCCH-ARFCN and BSIC combination corresponds to a specific BS identity.

The method of the preferred embodiment is shown in process flow chart 300 of FIG. 3. referring to FIG. 2, function box 305 shows the step of collecting, for a first serving cell, identity data related to identities of a plurality of cells other than said first serving cell and signal strength metric data related to respective signal strength metrics from said first serving cell and said plurality of other cells, to provide collected data. In the present embodiment the identity data and the signal metric data is originally acquired by a plurality of mobile stations by scanning the other cells and the serving cell, as explained earlier above. The data is in the form of the modules of data contained within the measurement reports. Thus, for example, in the case of data acquired by MS 44, the measurement report is transmitted from MS 44 to BS 28, and the data therein is further transmitted from BS 28 to BSC 47, where it is collected to provide collected data. In the present embodiment the corresponding data from a plurality of other mobile stations assigned to BS 28 is also transmitted to BS 28 and ftn-ther transmitted from BS 28 to BSC 47, but it is pointed out that the invention may nevertheless be usefully employed using the data from just one mobile station. It is noted that the data could according to other embodiments of this invention be collected at BS 28 itself, or alternatively at other network nodes.

In carrying out the present embodiment a statistical sampling procedure is employed, that is the GSM measurement reports are subjected to a statistical sampling procedure. In other words not every measurement report need be fully processed. In the present embodiment many tens of 8 thousand of measurement reports, providing a plurality of the signal metrics acquired over time, are processed so that the resulting calculations are carried out with a high degree of accuracy. This provides the advantage that the results are not biased for-example by a single long communication between MS44 and BS 28, since instead results from a plurality of communications between MS44 and BS 28 are used. Thus information across the whole of the coverage area of BS 28, i.e. over the whole cell., can be provided. By using data from a plurality of communications, the measurement report data can originate from users moving over the whole cell, due to the mobility of the mobile stations within the GSM system of the present embodiment. The same advantages can also be achieved in the case of any other mobile communication system. Sampling the measurement reports is implemented such that from the series of measurement reports being collected, one report is analysed, then a number are skipped, i.e. ignored, then the next measurement report is analysed, then a number are shipped, and so on. One advantageous way of implementing this is to select every nth measurement report where n is an integer. The sampling procedure creates a sub-set of the whole data set of statistics. This process allows more information to be processed for a given size of computer memory and processing capability. Furthermore, many measurement reports are similar, so by adopting a sampling procedure more variants of information modules within the measurement report can be listed with the given computer. Alternative statistical sampling approaches can be used in the present embodiment, and is chosen by the skilled person according to the prevalent patterns of data and desired or available amount of processing capacity. For example, the data can be collected on a sampling basis and then all the collected data further processed, and this may ihclude collecting from only some of the mobile stations in a cell.

The step of determining, at least in part from said collected data, the identities of said plurality of cells other than said first serving cell, is shown at function box 310 of FIG. 3. In the present embodiment, the way that this step is carried out is as follows. From the collected data the respective ARMN-BCCH and BSIC combinations reported within each 9 measurement report are separated and then compared against the list of such combinations held on the current neighbour list. This procedure is implemented by connecting a computer to the BSC and using routine computer methods well known in the art. An alternative means of implementing these procedures would be to use a computer or part of a computer integrated in the BSC or BS itself. Those identities and corresponding ARMN-BCCH and BSIC combinations that appear in the current neighbour list are isolated. This is also carried out by routine computer methods. Determination of the identity of the cell is most easily achieved by applying the assumption that the ARFMBCCH and BSIC combination is unique to a particular BS. If in certain circumstances this assumption proves to be imperfect, then supplementary information, for example the X and Y co-ordinates of each BS in the system, can be used. The distance between the serving BS and all other BSs is calculated, using trigonometric calculations well known to those in the art to provide a database having the ARFMBCCH and BSIC combinations of difterent BSs in correspondence with their respective distances from the serving BS. This database is sorted according to ARFW-BCCH and BSIC combination and in ascending distance from the serving BS. The database is searched to find those cells that have the candidate ARMN-BCCH and BSIC combination and have an appropriate distance from the serving BS. The distance is appropriate to range and coverage of the serving cell and its surrounding neighbours. For example in a city cells a few kilometres away would be considered, whereas in country areas the neighbours may be at a greater distance. Further information in terms of the antenna orientation and the predicted signal strength in the serving cell's coverage area can be used to distinguish when two neighbouring cells have similar distances from the serving cell and the same ARFCN-BCCH and BSIC combination in a system.

Before any calculations begin, the measurement reports are individually examined to determine which part of the signal metric is valid. This is done by examining the DTX flag and appropriately using the full or subrate value in the serving cell information contained within the measurement report. In the case of other systems compared to that of the present embodiment, it would be readily apparent to the skilled person what procedure would be required to determine the valid value from the available data, depending on the particular format of the data in that system.

The step of comparing, for each of said plurality of other cells, respective signal strength metrics with corresponding signal strength metrics from said serving cell, to provide a respective set of comparison values for each of said plurality of other cells, is shown at function box 315 of FIG. 3. In the present embodiment, the way that this step is carried out is as follows. For a given ARFCN-BCCH and BSIC combination (i.e. cell identity), from each measurement report that is sampled, the appropriate signal strength value from the measurement report information module (termed Ni, where i refers to the ith sampled measurement report) is determined. Also the corresponding value of the full or sub-rate, as appropriate, value of the signal strength, termed Si, from the serving cell in the same ith measurement report is determined. In the present embodiment the way in which the step of comparing, responsive to the previous step, a signal metric from the other cells with a signal metric from the serving cell is carried out is that a calculation of Si minus Ni is performed for each sampled measurement report. This can be automated using known computer methods. On the basis of the plurality of sampled measurement reports a list of counts for individual Si-Ni results is compiled. Resulting individual lists of Si-Ni for respective ARFCN-BCCH and BSIC combinations are statistically processed to count occurrences and a histogram is generated therefrom. Such a histogram 400 is shown in FIG. 4. This histogram thus shows the different values of S-N obtained for a single particular ARFCN-BCCH and BSIC combination (i.e.'a single other cell).

In the present embodiment, histograms of the type 400 shown in FIG. 4 are used in the step of determining, from said sets of comparison values and said identity data, a matrix indicating interference relationships between cells of a cellular communication system. This step is shown at function box 320 of FIG. 3. The sets of comparison values are the S-N values. The level of occurrence of each S-N value is represented by a shaded block in the histogram. In the present embodiment such levels of occurrence are each normalised with respect to the total number of all S-N values contained in the whole of the collected data for the serving cell. Such normalised values are then totalised within certain predetermined ranges of S-N value on the histogram to provide a respective interference coefficient for the particular other cell. The predetermined ranges are specified according to the particular system requirements. In the case of the present embodiment, the predetermined ranges are related to a frequency re-use and interference criteria employed in the communication system. The predetermined ranges also depend upon whether the interference matrix being determined is for co-channel interference relationships or adjacent channel interference relationships. In the present embodiment which is for a GSM system, the range of S-N values can extend from -63 to +63. In the present embodiment, assuming an adjacent channel frequency re-use and interference criteria of +9dB, and allowing a MB error margin which leads to a +6dB value, it follows that the normalised occurrence levels between -63 and -6 are totalised to calculate the interference co-efficient for adjacent channel interference (it is noted that a positive value of S-N means the server signal strength is greater than that from the other cell under consideration, whereas a negative value means the other cell signal strength is greater than the server signal strength, hence a positive dB level for frequency re-use criteria corresponds to a negative S-N value on the histogram). Considering now the calculation of the interference co- efficient in the case of co-channel interference, a co-channel frequency reuse and interference criteria of -9dB is assumed, which when allowing a MB error margin leads to a -12dB value. It follows that the normalised occurrence levels between -63 and +12 are totalised to calculate the interference co-efficient for co-channel interference.

In the present embodiment, a further step may be included prior to calculating the interference coefficient, namely statistically weighting said comparison values as a function of pre-determined ranges of said comparison values. An example of such ranges are shown as ranges (i), (ii) and (iii) in the histogram of FIG. 4, wherein range (i) corresponds to 12 S-N values between -63 and -6, range (ii) corresponds to S-N values between -6 and 0, and range (iii) corresponds to S-N values between 0 and +12. For example, in the case of calculating the interference coefficient for co channel interference, the statistical weighting can consist of multiplying those levels of occurrence in range (i) by a factor of 2, those levels of occurrence in range ( ii) by a factor of 1, and those levels of occurrence in range (iii) by a factor of 0.5. It is to be understood that the exact specification of any such ranges and likewise of their respective multiplication factors, will be chosen b the skilled person depending on the exact features of the system under consideration.

In the present embodiment all the above steps can be repeated for further serving cells to provide further interference co-efficients within the resulting matrix.

One advantage of the resulting matrix of the present embodiment is that the values therein consist of the above described interference coefficient which inherently represents a relative value directly related to the suitability of the corresponding server cell/other cell combination with respect to being assigned co-channel frequency status or adjacent channel frequency status during any frequency allocation process applied to the system and using the matrix. Also, since the interference co-efficients are calculated from, and normalised with respect to, data collected from real usage in the system, levels of subscriber usage are therefore included and hence are available for contribution to successful frequency planning. Another useful outcome of the form of the interfernce co- efficients is that since they are relative values, a further advantage can be obtained by carrying out a further normalising step in which the most 6xtreme interference co-efficient of a matrix is set to 1 and the other coefficients normalised thereto.

A further aspect of the present invention is the use of the matrix in a frequency planning process. The matrix resulting from the present invention is particularly suitable for use in automatic frequency paInning (AFP), a process well known in the art.

13 It is to be understood that the present invention is not limited to the specific examples given above. Furthermore, it is to be understood that the invention not only includes cases where the fflatrix is provided in the form of printed or other viewable data, rather it also includes cases where it is provided in other hard copy forms readable by electronic or other means, e.g. on magnetic cassette or disc, and also includes the case of when suclh a matrix is only provided in an electronic form such as when it is provided or stored in a computer memory or other electronic memory.

Likewise the present invention includes the case when the matrix as such cannot be directly analysed by a human but is nevertheless usable in the next stage of any type of frequency manipulation or planning method, such as automatic frequency planning, involving the information contained in the matrix.

Furthermore, it is to be appreciated that the different optional aspects mentioned in the above description of the preferred embodiment can be combined in numerous ways whilst implementing the invention.

14

Claims (12)

1. A method of providing a matrix indicating-interference 5 relationships between cells of a cellular commilni cation system, the method comprising the steps of.

- collecting, for a first serving cell, identity data related to identities of a plurality of cells other than said first serving cell and signal strength metric data related to respective signal strength metrics from said first serving cell and said plurality of other cells, to provide collected data; - determining, at least in part from said collected data, the identities of said plurality of cells other than said first serving cell; - comparing, for each of said plurality of other cells, respective signal strength metrics with corresponding signal strength metrics from said serving cell, to provide a respective set of comparison values for each of said plurality of other cells; and - determining, from said sets of comparison values and said identity data, said matrix.

2. A method according to Claim 1 wherein said steps are performed for a plurality of said signal metrics acquired over time.

3. A method according to Claim 1 wherein one or more of said steps is 30 performed according to a statistical sampling procedure.

4. A method according to Claim 3 wherein said system is a GSM system, said identity data and said signal strength metric data is in the form of measurement reports, and said statistical sampling procedure 35 comprises selecting a statistical sample of said measurement reports.

5. A method according to Claim 1 wherein said identity data and said signal metric data is acquired by one or more mobile stations by scanning said other cells and/or said serving cell.

6. A method according to Claim 1, wherein said steps are repeated for further serving cells.

7. A method according to Claim 1, wherein said determining step 10 comprises the steps of.

- normalising each of said respective sets of comparison values with respect to said collected data; and said other cell.

- calculating therefrom a respective interference co-efficient for each

8. A method according to Claim 7, wherein said determining step further comprises the steps of, - statistically weighting said comparison values as a function of predetermined ranges of said comparison values prior to said step of calculating said respective interference coefficients.

9. A method according to Claim 8, wherein said predetermined ranges are related to a frequency re-use and/or interference criteria employed in said system.

10. A method according to Claim 9, wherein said frequency re-use and/or interference criteria depends upon whether said interference matrix is for co-channel interference relationships or for adjacent-channel interference relationships.

11. A method of providing a matrix indicating interference relationships between cells of a cellular communication system substantially as hereinbefore described and with reference to the accompanying drawings.

16

12. A method of frequency planning in a cellular communication system comprising the use of a matrix provided by the method of any preceding claim.