CN101848482A - Method and device for acquiring interference matrix - Google Patents
Method and device for acquiring interference matrix Download PDFInfo
- Publication number
- CN101848482A CN101848482A CN 201010133225 CN201010133225A CN101848482A CN 101848482 A CN101848482 A CN 101848482A CN 201010133225 CN201010133225 CN 201010133225 CN 201010133225 A CN201010133225 A CN 201010133225A CN 101848482 A CN101848482 A CN 101848482A
- Authority
- CN
- China
- Prior art keywords
- data
- cluster
- measurement report
- telephone traffic
- frequency
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Landscapes
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses a method and a device for acquiring an interference matrix. The method mainly comprises the clustering of sweep-frequency data according to frequency-domain characteristics, the telephone traffic distribution estimation based on measurement report classification, and the data fusion and interference matrix computation based on class characteristics and telephone traffic distribution. The device mainly comprises a module for clustering according to the frequency-domain characteristics, a telephone traffic distribution estimating module, a data fusion module and an interference matrix computation module. The invention enables the generated interference matrix to accurately and completely reflect the actual situation of minizone interference and the real feeling of a mobile user via the complementation of the sweep-frequency data and the statistical data of the measurement report. The method of the invention has no special requirement on the input data, the required statistical data of the measurement report can be easily extracted on various OMCR devices of manufactures; and the treatment complexity is low, a large number of measurement reports do not need to be analyzed, a three-dimensional map and other auxiliary tools are not needed, and the engineering practicability is strong.
Description
Technical field
The present invention relates to the mobile communication technology field, relate in particular to a kind of about obtaining the method and apparatus of interference matrix in the network planning process.
Background technology
For mobile communication system, frequency resource is very precious.Satisfy the demand of user for making full use of the finite frequency resource to network capacity and quality, the GSM network using channeling technology.Channeling has improved the power system capacity of network greatly, but it can cause presence of intercell interference, reduces speech quality.Thereby reasonably network planning is the core of networking.
In the daily operation process of GSM network, frequency needs along with the variation of coverage internal traffic, communication environments is adjusted, to keep the performance of network in the distribution of minizone.It is interference matrix (Interference Matrix that ripe at present network changes frequency method, IM) technology, promptly the interference table between the sub-district is shown an interference matrix by analysis to measured data, element in the matrix minimizes total interference of whole network then as the tolerance of disturbing size between two sub-districts by the heavily distribution of frequency.Network changes the accuracy that expection effect frequently depends on interference matrix to a great extent.
According to the difference of used measured data, the existing method of obtaining interference matrix mainly contains two kinds: a kind of drive test frequency sweep data that are based on are obtained interference matrix, and another kind is based on mobile phone measurement report and obtains interference matrix.Frequency sweep data frequency domain information completely, and have latitude and longitude information, can reflect the disturbed condition on the collecting location exactly; Can only carry out in some areas but gather, the interference matrix that generates based on the frequency sweep data can not reflect the disturbed condition that all have traffic scene place, does not also contain telephone traffic information.Mobile phone measurement report includes telephone traffic information, but does not contain positional information, owing to only comprise the received power grade of 6 the strongest adjacent sub-districts of current main plot and signal, so its frequency domain information is imperfect, the interference matrix of generation has certain error again.
In addition, application number is to disclose a kind of method of obtaining interference matrix in 200810225983.3 the Chinese patent application: use the ray trace model, obtain coverage information in conjunction with base station data, Three-dimensional Numeric Map and frequency sweep data; The measurement report initial data analyzed obtain traffic distributed data with rough geographical location information; Traffic distributed data and coverage information are mated, and soon the traffic value in the unit grid is evenly distributed in the unit grid that is equal to the coverage information precision in the traffic map; Calculate interference matrix, realize the network planning of lattice level.Owing to adopt the ray trace theoretical model, can not accurately reflect the actual conditions of presence of intercell interference, be difficult to reflect user's sense of reality, and need a lot of auxiliary resources such as base station data, Three-dimensional Numeric Map.Owing to reasons such as technology barriers and costs, the measurement report initial data usually can't be gathered again.And the analysis of magnanimity initial data also causes very high amount of calculation.
Summary of the invention
Order of the present invention ground is to overcome the deficiencies in the prior art, makes full use of the complementarity of frequency sweep data and measurement report, and a kind of method and apparatus that obtains interference matrix is provided.
The objective of the invention is to be achieved through the following technical solutions:
A kind of method of obtaining interference matrix comprises the steps:
(1) the frequency sweep data are by the frequency domain character cluster: the input data are sub-district frequency sweep data; The space or the cell ID space of frequency point number space, frequency point number and BCC combination got in the cluster space; The cluster object is got the Boolean that performance number, carrier/interface ratio value, carrier/interface ratio quantized value or the characterization signal of signal on the cluster dimension have or not; The mean value of specific cluster object appears in the category feature that cluster obtains on each cluster dimension for such signal; Cluster result is carried out hypothesis testing.
(2) estimate based on the telephone traffic distribution of measurement report classification: the input data are the category feature that obtains in cell measurement reporting statistics and the step (1); Adopt multiple linear regression model that the composition analysis of measurement report statistics is estimated the telephone traffic of each class signal correspondence, explained variable is the measurement report statistics, and explanatory variable is the category feature that obtains in the step (1); In conjunction with the latitude and longitude information of frequency sweep data with find the solution the coefficient vector that linear regression model (LRM) obtains and obtain cell telephone traffic amount geographical distribution valuation; Overall fit goodness and regression coefficient to regression model are carried out hypothesis testing.
(3) data fusion and the interference matrix based on category feature and telephone traffic distribution calculates: the data fusion based on category feature and telephone traffic distribution comprises: measurement report statistics frequency domain information replenishes and the frequency sweep data are pressed telephone traffic distribution weighting reconstruct, measurement report statistics frequency domain information replenishes and is first data fusion, and the frequency sweep data are second data fusion by telephone traffic distribution weighting reconstruct.Described measurement report statistics frequency domain information replenishes: according to the frequency domain information that lacks in the report of the category feature additional survey on linear regression model (LRM) and the exclusive cluster dimension of the frequency sweep data statistics.Described frequency sweep data are reconstructed into by telephone traffic distribution weighting: the composition of frequency sweep data is adjusted according to the telephone traffic distribution, made that the frequency sweep data after the reconstruct have the geographical distribution identical with telephone traffic.Interference matrix calculates and to comprise: replenish that measurement report statistics after complete is calculated interference matrix or with the frequency sweep data computation interference matrix of pressing after the telephone traffic distribution weighting reconstruct with frequency domain information, wherein, replenish measurement report statistics after complete with frequency domain information and calculate interference matrix and be first interference matrix and calculate, use to be second interference matrix by the frequency sweep data computation interference matrix after the telephone traffic distribution weighting reconstruct and to calculate.
A kind of device that obtains interference matrix, described device comprises:
One presses frequency domain character cluster module: connect the frequency sweep data, with the data based frequency domain character cluster of frequency sweep, obtain category feature;
One telephone traffic distribution estimation module: connect and press frequency domain character cluster module, do linear regression with measurement report statistics and category feature and estimate that the cell telephone traffic amount distributes;
One measurement report statistics complementary module, i.e. first data fusion module: connect the measurement report statistics, press frequency domain character cluster module and telephone traffic distribution estimation module, be used for the frequency domain information that lacks according to linear regression model (LRM) additional survey report statistics, obtain frequency domain information and replenish measurement report statistics after complete; With
One first interference matrix computing module: connect measurement report statistics complementary module, be used for generating interference matrix IM-MR ' according to the measurement report statistics that frequency domain information replenishes after complete.
Further, describedly comprise by frequency domain character cluster module:
One data format preliminary treatment submodule: connecting the frequency sweep data, is carrier/interface ratio quantized data and the cluster data with specific cluster space and object with the frequency sweep data conversion;
One cluster submodule: connect data format preliminary treatment submodule, it is the K class that cluster data is gathered, and carries out the hypothesis testing of cluster result; With
One category feature generates submodule: connect data format preliminary treatment submodule and cluster submodule, be used for generating category feature according to cluster result and carrier/interface ratio quantized data.
Further, described telephone traffic distribution estimation module comprises:
One Data Matching submodule: connect measurement report statistics and category feature and generate submodule, be used to select category feature and measurement report statistics with identical cluster dimension; With
One regression coefficient estimator module: connect the data matched sub-block, be used for carrying out the linear regression model (LRM) coefficient and find the solution, and the overall fit goodness and the regression coefficient of regression model are carried out hypothesis testing according to the measurement report statistics of coupling and category feature.
A kind of device that obtains interference matrix, described device comprises:
One presses frequency domain character cluster module: connect the frequency sweep data, with the data based frequency domain character cluster of frequency sweep, obtain category feature;
One telephone traffic distribution estimation module: connect and press frequency domain character cluster module, do linear regression with measurement report statistics and category feature and estimate that the cell telephone traffic amount distributes;
Sweep the audio data reconstructed module, i.e. second data fusion module: connect the frequency sweep data, press frequency domain character cluster module and telephone traffic distribution estimation module, be used for the frequency sweep data being weighted reconstruct, obtain by the frequency sweep data after the telephone traffic distribution weighting reconstruct according to cell telephone traffic amount distribution valuation; With
One second interference matrix computing module: connect frequency sweep data reconstruction module, be used for according to generating interference matrix IM-DT ' by the frequency sweep data after the telephone traffic distribution weighting reconstruct.
Beneficial effect of the present invention is: frequency sweep data and measurement report statistics are complementary to be utilized, the interference matrix reflection minizone actual interference relation of generation and user's sense of reality; Engineering practicability is strong, only need frequency sweep data and measurement report statistics, required measurement report statistics all can easily be extracted on the OMCR of each manufacturer equipment, need not the original measurement report of magnanimity is analyzed, and need not aids such as three-dimensional map.
Description of drawings
Fig. 1 is the schematic diagram of technical solution of the present invention;
Fig. 2 obtains the flow chart of interference matrix for the present invention;
Fig. 3 is for obtaining the schematic diagram of interference matrix based on the frequency sweep data;
Fig. 4 is the structural representation of a kind of implement device of the present invention;
Fig. 5 is the structural representation of the another kind of implement device of the present invention;
Fig. 6 is a routine cluster result figure.
Embodiment
Technical solution of the present invention is based on the following fact: the frequency sweep data are relevant with measurement report, and they all are the measurements to the mobile communications network wireless signal strength, are same overall different samples; Frequency sweep data and measurement report are and are complementary that the former can be provided by the latter telephone traffic information of disappearance, and the incomplete frequency domain information of the latter can be replenished by the former.
The principle of technical solution of the present invention is as follows: in the cell mobile communication systems, the wireless signal that the optional position receives had both comprised that this subdistrict frequency point signal also comprised the frequency signal of a plurality of adjacent sub-districts on every side, and the signal strength signal intensity of each frequency changes along with the geographical position.Thereby there are strong correlation in frequency domain information and geographical position in the wireless signal, can be used as " fingerprint " feature in geographical position.The data based frequency domain information of frequency sweep is carried out cluster analysis, can obtain some typical wireless signal types, signal type is corresponding one by one with the region.Measurement report is classified by above-mentioned frequency domain character, can estimate the distribution of telephone traffic with geographic location.Based on one of frequency sweep data and measurement report, the information that it lacked is replenished with another person, is called data fusion.A kind of data fusion is based on measurement report, and the frequency domain information that it lacked replenishes with the frequency domain information in the frequency sweep data of the same area.Replenish the interference matrix IM-MR ' that the measurement report after complete generates with frequency domain information and compare with the interference matrix IM-MR that generates with measurement report, the former reflects more complete interference relationships on the multifrequency point, thereby more approaches the reality of presence of intercell interference.Another kind of data fusion is that the telephone traffic information that it lacked provides through estimating with the measurement report of the same area based on the frequency sweep data.Compare with the interference matrix IM-DT that generates with the frequency sweep data with the interference matrix IM-DT ' that generates by the frequency sweep data after the telephone traffic distribution weighting reconstruct, the former reflects the geographical distribution of telephone traffic, promptly give different attention rates with the duration of call to zones of different, change yupin effect and can benefit mobile subscriber/business as much as possible thereby make by user's dense degree.The measurement report that replenishes after complete with frequency domain information generates interference matrix IM-MR ', or with by the generation of the frequency sweep data after telephone traffic distribution weighting reconstruct interference matrix IM-DT ', all can adopt prior art.
Should be pointed out that owing to reasons such as technology barriers and costs original measurement report is difficult to obtain, can only obtain the measurement report statistics of process statistical disposition on the OMCR equipment in the engineering usually.So, the distribution of above-mentioned telephone traffic with geographic location can not be with original measurement report by signal characteristic through simple classification, report that by class the simple statistics of bar number obtains, and should mate with measurement report statistics and signal type, coefficient finds the solution and obtains.And type matching, coefficient are found the solution, and need realize by mathematical statistics methods such as multiple linear regressions.Above-mentioned first kind of data fusion is actually with the frequency domain information in the frequency sweep data of the same area and comes additional survey to report the frequency domain information that lacks in the statistics.Above-mentioned second kind of data fusion is actually the telephone traffic that obtains through recurrence with the measurement report statistics and distributes the frequency sweep data are weighted reconstruct.
The present invention obtains the method for interference matrix, comprises the steps:
A. the frequency sweep data are pressed the frequency domain character cluster;
B. estimate based on the telephone traffic distribution of measurement report classification;
C. data fusion and the interference matrix based on category feature and telephone traffic distribution calculates.
In the described steps A, space, the CELLID space, sub-district of frequency space, frequency and BCC combination can be got in the cluster space, the cluster object can be got the Boolean that performance number, carrier/interface ratio value, carrier/interface ratio quantized value, the characterization signal of swept-frequency signal on the cluster dimension have or not, the category feature that cluster obtains is specific cluster object appears in such signal on each cluster dimension a mean value, and it is correct to guarantee on its statistical significance that cluster result must carry out hypothesis testing.
Among the described step B, adopt the constituent of linear regression model (LRM) analysis to measure report statistics, estimate the telephone traffic of each class signal correspondence, explained variable is the measurement report statistics, and explanatory variable is the category feature that obtains in the steps A.Obtain the distribution of cell telephone traffic amount in conjunction with the latitude and longitude information of frequency sweep data and the coefficient vector of linear regression model (LRM) again.The overall fit goodness of regression model and regression coefficient are all carried out hypothesis testing to guarantee the correctness of regression result on statistical significance.
Among the described step C, the data fusion that distributes according to category feature and telephone traffic comprises: measurement report statistics frequency domain information replenishes, and the frequency sweep data are by telephone traffic distribution weighting reconstruct.The additional category feature and the linear regression model (LRM) that obtain with the frequency sweep data of referring to of measurement report statistics frequency domain information comes additional survey to report the frequency domain information that lacks in the statistics.The frequency sweep data are the constituent ratios that distributes and adjust the frequency sweep data in telephone traffic in telephone traffic distribution weighting reconstruct.
In described step C, interference matrix calculates and to comprise: calculate interference matrix with the measurement report statistics that frequency domain information replenishes after complete, or with the frequency sweep data computation interference matrix of pressing after the telephone traffic distribution weighting reconstruct.The calculating of interference matrix can be adopted any existing method.For example, available ICD well known in the art (Inter Cell Independence) method is calculated co-channel interference matrix and adjacent interference matrix frequently.
The present invention obtains the device of interference matrix, comprising:
Press frequency domain character cluster module: connect the frequency sweep data,, obtain category feature with the data based frequency domain character cluster of frequency sweep;
Telephone traffic distribution estimation module: connect and press frequency domain character cluster module, do linear regression with measurement report statistics and category feature and estimate that the cell telephone traffic amount distributes.
Measurement report statistics complementary module (first data fusion module): connect the measurement report statistics, press frequency domain character cluster module and telephone traffic distribution estimation module, be used for the frequency domain information that lacks according to linear regression model (LRM) additional survey report statistics, obtain frequency domain information and replenish measurement report statistics after complete;
The first interference matrix computing module: connect measurement report statistics complementary module, the measurement report statistics of being replenished after complete by frequency domain information generates interference matrix IM-MR '.
Describedly further comprise by frequency domain character cluster module:
Data format preliminary treatment submodule: connecting the frequency sweep data, is carrier/interface ratio quantized data and the cluster data with specific cluster space and cluster object with the frequency sweep data conversion;
The cluster submodule: connect data format preliminary treatment submodule, it is the K class that cluster data is gathered, and carries out the significance test of cluster result;
Category feature generates submodule: connect data format preliminary treatment submodule and cluster submodule, be used for generating category feature according to cluster result and carrier/interface ratio quantized data.
Described telephone traffic distribution estimation module further comprises:
Data Matching submodule: connect measurement report statistics and category feature and generate submodule, be used to select category feature and measurement report statistics with identical cluster dimension;
Regression coefficient is found the solution submodule: connect the data matched sub-block, be used for carrying out the linear regression model (LRM) coefficient according to the measurement report statistics of coupling and category feature and find the solution, and the overall fit goodness and the regression coefficient of regression model are carried out significance test.
Above-mentioned measurement report statistics complementary module also can replace with frequency sweep data reconstruction module, and the first interference matrix computing module replaces with the second interference matrix computing module simultaneously.
Frequency sweep data reconstruction module (second data fusion module): connect the frequency sweep data, press frequency domain character cluster module and telephone traffic distribution estimation module, be used for the frequency sweep data being weighted reconstruct, obtain by the frequency sweep data after the telephone traffic distribution weighting reconstruct according to cell telephone traffic amount distribution valuation.
The second interference matrix computing module: connect frequency sweep data reconstruction module, be used for according to generating interference matrix IM-DT ' by the frequency sweep data after the telephone traffic distribution weighting reconstruct.
As shown in Figure 1, the present invention is based on frequency sweep data and measurement report statistics, obtain interference matrix by merging two kinds of data.Below in conjunction with specific embodiment and accompanying drawing the present invention is described in further details.
Embodiment one
As shown in Figure 2, the present invention's method of obtaining interference matrix comprises the steps:
Step 21: the frequency sweep data conversion is the carrier/interface ratio quantized data
Each bar frequency sweep data provides signal strength signal intensity (dBm), location area identifier (LAC), cell identifier (CELL ID), the BCC (BSIC) on all frequencies.Table 1 provides an example.
Table 1 frequency sweep data format example
| The frequency sweep data | Frequency point number | Signal power value (dBm) | ??LAC | ??CELL??ID | BCC (BSIC) | Frequency point number | Signal power value (dBm) | ...... |
| Article one | ??1 | ??□83.56 | ??22580 | ??40411 | ??70 | 2 | ??□49.12 | ...... |
| ??...... | ??...... | ??...... | ??...... | ??...... | ??...... | ...... | ??...... | ...... |
In order to realize the fusion of frequency sweep data and measurement report statistics, needing is the carrier/interface ratio quantized data with the frequency sweep data conversion at first, constitutes carrier/interface ratio quantized data matrix T, is used to calculate the category feature of each class frequency sweep data.Conversion process comprises the steps:
1), obtains the frequency sweep data of carrier/interface ratio value form according to the signal power value signal calculated carrier/interface ratio value in the frequency sweep data.Table 2 provides an example.
The frequency sweep data instance of table 2 carrier/interface ratio value form (supposing signal strength signal intensity 40dBm)
| The frequency sweep data | Frequency point number | Carrier/interface ratio value (dB) | ??LAC | ??CELL??ID | BCC (BSIC) | Frequency point number | Carrier/interface ratio value (dB) | ...... |
| Article one | ??1 | ??43.56 | ??22580 | ??40411 | ??70 | 2 | ??9.12 | ...... |
| ??...... | ??...... | ??...... | ??...... | ??...... | ??...... | ...... | ??...... | ...... |
2) combination (frequency colour coding to) of getting frequency point number and BCC is as the cluster dimension, and all frequency colour codings are to constituting the cluster space.Also can get frequency point number as the cluster dimension, all frequency point number constitute the cluster space.Can also get cell ID (LAC+CELL ID) as the cluster dimension, all cell IDs constitute the cluster space.
3) each the carrier/interface ratio value in the frequency sweep data of each bar carrier/interface ratio value form is mapped to dimension corresponding in the cluster space, a frequency sweep data map is a vector.Table 3 provides an example.
The vector that table 3 frequency sweep data map obtains
4) in each vector, find out 6 neighboring area signals of carrier/interface ratio value minimum and its carrier/interface ratio value is quantized, the carrier/interface ratio value of other signals is changed to zero.Carrier/interface ratio value quantized interval is interval consistent with the carrier/interface ratio primary system meter of measurement report statistics, altogether J interval; Quantized value j=1,2...J, as shown in table 4.Interval unification is for follow-up data fusion.
The corresponding relation of table 4 interval and quantized value
5) generate carrier/interface ratio quantized data matrix T, the element value of the capable m row of its n is carrier/interface ratio quantized values of n dimension in above-mentioned m the vector,
n＝1，2，...，N
dt
m＝1，2，...，M
In the formula, N
DtBe the right sum of frequency sweep data intermediate-frequeney point colour coding; M is the bar number of frequency sweep data; J=1,2...J.Carrier/interface ratio quantized data matrix T is for example as table 5.
Table 5 carrier/interface ratio quantized data matrix T (J=10)
Step 22: the carrier/interface ratio quantized data is converted into cluster data
The cluster data matrix D can be changed to 1 by the numerical value of non-zero in the carrier/interface ratio quantized data matrix T and obtain, also can equal carrier/interface ratio quantized data matrix T, also can get in each bar frequency sweep data each frequency colour coding to last carrier/interface ratio value, also can directly get in each bar frequency sweep data each frequency colour coding last signal power value.That is to say that the cluster object can be the Boolean that characterization signal has or not, or the carrier/interface ratio quantized value, or the carrier/interface ratio value, or the signal power value.
When the cluster data matrix D by carrier/interface ratio quantized data matrix T in the numerical value of non-zero be changed to 1 when obtaining, the capable m column element of its n is
n＝1，2，...，N
dt；????????????。
m＝1，2，...，M
It is in order to obtain stable cluster result that cluster data adopts the Boolean form.
The cluster data matrix D has been described the signal frequency-domain structure feature, for example as table 6.
Table 6 cluster data matrix D one example
Step 23: cluster
The effect of cluster is that the coverage with the sub-district is divided into K class zone, and each class zone (one or more geographic area) has unique adjacent area interference relationships.
With the frequency colour coding to be cluster dimension, Euclidean distance as the distance metric criterion, cluster data is carried out cluster, obtain class members table, as shown in table 7, class sequence number k=1,2 ... K.Also can select Minkowski distance, coefficient correlation, mahalanobis distance etc. to carry out cluster as the distance metric criterion.
Table 7 class members shows (cluster membership) example
| Frequency sweep data sequence number m | ??1 | ??2 | ??3 | ??4 | ??5 | ??6 | ??7 | ... |
| The sequence number k of affiliated class | ??1 | ??2 | ??2 | ??3 | ??3 | ??4 | ??4 | ... |
In order to ensure the correctness and the reliability of cluster result, cluster result must carry out significance test.
Step 24: category feature calculates
With class members's table and carrier/interface ratio quantized data matrix T, compute classes feature.Category feature p
N, k(j) be meant k class frequency sweep data n frequency colour coding on the carrier/interface ratio value that records fall into j the probability that statistics is interval,
In the formula, T
N, m, kBe k class m bar frequency sweep data n frequency colour coding on the carrier/interface ratio quantized value, k=1,2 ... K; α
kIt is the bar number of k class frequency sweep data; Decision function
So, 0≤p
N, k(j)≤1.
Step 25: telephone traffic distributes and estimates
To the measurement report statistics, analyze its constituent with linear regression model (LRM), find the solution regression coefficient, obtain cell telephone traffic amount distribution valuation.Linear regression analysis be frequency colour coding that measurement report statistics and frequency sweep data have jointly on carry out.
Described linear regression model (LRM) is
R
n，j＝β
0+β
1·p
n，1(j)+…+β
K·p
n，K(j)+u
n(j)
n＝1，2，…，N
C
j＝1，2，...，J
In the formula, explained variable R
N, jBe that n frequency colour coding drops on j the measurement report number in the statistics interval, explanatory variable p to uploading dried ratio
N, k(j) be the category feature of k class, u
n(j) be random error, N
cThe right sum of frequency colour coding that has jointly for frequency sweep data and measurement report statistics.Regression coefficient β
kPhysical significance be the valuation of the telephone traffic (measurement report bar number) that produces on the k class zone, k=1,2 ... K.The constant term factor beta
0Be the valuation that does not belong to the telephone traffic of any class, owe complete, frequency sweep data and measurement report statistics part mismatch causes by the frequency sweep data.
According to R
N, jAnd p
N, k(j) (j=1...J), find the solution and draw regression coefficient vector β=[β
0β
1β
K], i.e. cell telephone traffic amount distribution valuation.
Significance test is all carried out in the overall fit goodness of regression model and regression coefficient valuation.
Step 26: first data fusion (measurement report statistics frequency domain information replenishes)
Based on the measurement report statistics,, obtain frequency domain information and replenish measurement report statistics after complete (hereinafter to be referred as the measurement report statistics of replenishing after complete) with the frequency domain information that lacks in the frequency sweep data additional survey report statistics.The frequency domain information of disappearance refers to that in the measurement report statistics unexistent frequency colour coding uploads dried ratio to n ' and drop on j the measurement report number of adding up in the interval.
At first in the measurement report statistics unexistent frequency colour coding to n ' on, estimate that the carrier/interface ratio quantized value drops on j the measurement report number of adding up in the interval
n′＝1，...，N
dt-N
C
j＝1，2，...，J
Again from (N
Dt-N
C) filter out strong interferers among the individual n ', promptly satisfy
The frequency colour coding right.Wherein, J
ThFor judging whether neighboring area signal produces the carrier/interface ratio threshold value of interference, its numerical value is set by mobile operator usually, is used for the calculating of interference matrix.
Strong interferers frequency colour coding is dropped on the measurement report number of j statistics in the interval and changed into by original null value uploading dried ratio
Step 27: first interference matrix calculates
To the measurement report statistics after additional complete, can generate interference matrix IM-MR ' with existing method calculating.For example, available ICD well known in the art (Inter Cell Independence) method, the capable b column element of a of calculating interference matrix
In the formula, subscript a represents the main plot, and b represents interfered cell, and n is that the pairing frequency colour coding of interfered cell b is to sequence number.Work as J
ThBe taken as co-channel interference thresholding J
Th CoThe time, obtain the co-channel interference matrix.Work as J
ThBe taken as adjacent interference threshold J frequently
Th AdThe time, obtain adjacent interference matrix frequently.The concrete computational details is known for those skilled in the art, repeats no more here.
Replace step 26 with step 28, replace step 27 with step 29 simultaneously, obtain obtaining the another kind of method of interference matrix.
Step 28: second data fusion (the frequency sweep data are by telephone traffic distribution weighting reconstruct)
Based on the frequency sweep data, utilize the telephone traffic distribution valuation that obtains in the step 25 that the frequency sweep data are weighted reconstruct, obtain by the frequency sweep data after the telephone traffic distribution weighting reconstruct (hereinafter to be referred as the frequency sweep data after the reconstruct).Described reconstructing method is, for k class frequency sweep data (k=1,2 ... K), with original α
kBar frequency sweep data extending is β
kBar.For example can adopt following method: find the solution linear equation β
k=A
kα
k+ β
k, obtain coefficient A
kAnd β
kEach bar of k class frequency sweep data is copied as A
kBar is again from original α
kPicked at random B in article k class frequency sweep data
kBar promptly obtains the frequency sweep data β altogether after the reconstruct of k class
kBar.
Step 29: second interference matrix calculates
To the frequency sweep data after the reconstruct, can calculate generation interference matrix with existing method.For example, available carrier/interface ratio value distribution characteristics method well known in the art as shown in Figure 3, is got the carrier/interface ratio value CIR at the following α quantile place of carrier/interface ratio value distribution
αElement as the capable b row of a constitutes interference matrix IM-DT '.The concrete computational details is known for those skilled in the art, repeats no more here.
By above description as can be seen, the present invention is by the deep excavation to the available data resource, adopt comparatively simple statistical method to realize the complementation of measurement report statistics and frequency sweep data, two kinds of new interference matrix that obtain can reflect the presence of intercell interference relation comprehensively, exactly.
Embodiment two
Technical scheme provided by the invention also can adopt device shown in Figure 4 to realize.Device comprises:
Press frequency domain character cluster module: connect the frequency sweep data,, obtain category feature with the data based frequency domain character cluster of frequency sweep;
Telephone traffic distribution estimation module: connect and press frequency domain character cluster module, do linear regression with measurement report statistics and category feature and estimate that the cell telephone traffic amount distributes.
Measurement report statistics complementary module (first data fusion module): connect the measurement report statistics, press frequency domain character cluster module and telephone traffic distribution estimation module, be used for the frequency domain information that lacks according to linear regression model (LRM) additional survey report statistics, obtain replenishing the measurement report statistics after complete;
The first interference matrix computing module: connect measurement report statistics complementary module, generate interference matrix IM-MR ' by the measurement report statistics after additional complete.
Describedly further comprise by frequency domain character cluster module:
Data format preliminary treatment submodule: connecting the frequency sweep data, is carrier/interface ratio quantized data and the cluster data with specific cluster space and feature with the frequency sweep data conversion;
The cluster submodule: connect data format preliminary treatment submodule, it is the K class that cluster data is gathered, and carries out the significance test of cluster result;
Category feature generates submodule: connect data format preliminary treatment submodule and cluster submodule, be used for generating category feature according to cluster result and carrier/interface ratio quantized data.
Described telephone traffic distribution estimation module further comprises:
Data Matching submodule: connect measurement report statistics and category feature and generate submodule, be used to select category feature and measurement report statistics with identical cluster dimension;
Regression coefficient is found the solution submodule: connect the data matched sub-block, be used for carrying out the linear regression model (LRM) coefficient and finding the solution, and the overall fit goodness of regression model and the regression coefficient of trying to achieve are carried out significance test according to the measurement report statistics and the category feature of coupling.
This device has very strong independence, can be added into network optimization system as module independently, does not need original system is made any modification.
Embodiment three
Technical scheme provided by the invention also can adopt device shown in Figure 3 to realize.Device comprises:
Press frequency domain character cluster module: with pressing frequency domain character cluster module in the example two described devices;
Telephone traffic distribution estimation module: with the telephone traffic distribution estimation module in the example two described devices;
Frequency sweep data reconstruction module (second data fusion module): connect the frequency sweep data, press frequency domain character cluster module and telephone traffic distribution estimation module, be used for the frequency sweep data being weighted reconstruct, obtain the frequency sweep data after the reconstruct according to cell telephone traffic amount distribution valuation.
The second interference matrix computing module: connect frequency sweep data reconstruction module, be used for generating interference matrix IM-DT ' according to the frequency sweep data after the reconstruct.
This device has very strong independence, can be added into network optimization system as module independently, does not need original system is made any modification.
The beneficial effect of technical scheme of the present invention can be verified by emulation.For example, clusters number K gets 6, gets certain sub-district frequency sweep data clusters result, sees Fig. 6.Among the figure, abscissa is a longitude, and ordinate is a latitude, the corresponding frequency sweep data of every bit.As seen from the figure, cluster result and geographical position strong correlation.
This sub-district according to the equation of linear regression that frequency sweep data and measurement report statistics obtain is:
R
n，j＝β
0+β
1·p
n，1(j)+…+β
6·p
n，6(j)+u
n(j)
n＝1，2，…，N
C
j＝1，2，...，10
Solve factor beta
0=15110, β
1=348115, β
2=627623, β
3=241825, β
4=0, β
5=565001 and β
6=329114, be the valuation that telephone traffic distributes in 6 zones.The frequency domain information that can realize the measurement report statistics with coefficient vector β replenishes (first data fusion), or realizes that the frequency sweep data are by telephone traffic distribution weighting reconstruct (second data fusion).
Replenish the strong interferers that increases newly in the measurement report statistics of complete back with frequency domain information and see Table 8.As can be seen, the technology of the present invention can more completely be described the actual interference relation of minizone.
The strong interferers that table 8 is newly-increased
| The sign of main plot (cell id) | Newly-increased strong interferers (frequency colour coding to) |
| ??30531 | ??073_58??090_62??024_57??081_61??095_49??084_49 |
Obviously, those skilled in the art can carry out various changes and distortion to the present invention and not break away from the spirit and scope of the present invention.Like this, if these modifications of the present invention and distortion belong within the scope of claim of the present invention and equivalent technologies thereof, then the present invention also is intended to comprise these changes and is out of shape interior.
Claims (5)
1. a method of obtaining interference matrix is characterized in that, comprises the steps:
(1) the frequency sweep data are by the frequency domain character cluster: the input data are sub-district frequency sweep data; The space or the cell ID space of frequency point number space, frequency point number and BCC combination got in the cluster space; The cluster object is got the Boolean that performance number, carrier/interface ratio value, carrier/interface ratio quantized value or the characterization signal of signal on the cluster dimension have or not; The mean value of specific cluster object appears in the category feature that cluster obtains on each cluster dimension for such signal; Cluster result is carried out hypothesis testing.
(2) estimate based on the telephone traffic distribution of measurement report classification: the input data are the category feature that obtains in cell measurement reporting statistics and the step (1); Adopt multiple linear regression model that the composition analysis of measurement report statistics is estimated the telephone traffic of each class signal correspondence, explained variable is the measurement report statistics, and explanatory variable is the category feature that obtains in the step (1); In conjunction with the latitude and longitude information of frequency sweep data with find the solution the coefficient vector that linear regression model (LRM) obtains and obtain cell telephone traffic amount geographical distribution valuation; Overall fit goodness and regression coefficient to regression model are carried out hypothesis testing.
(3) data fusion and the interference matrix based on category feature and telephone traffic distribution calculates: the data fusion based on category feature and telephone traffic distribution comprises: measurement report statistics frequency domain information replenishes and the frequency sweep data are pressed telephone traffic distribution weighting reconstruct, measurement report statistics frequency domain information replenishes and is first data fusion, and the frequency sweep data are second data fusion by telephone traffic distribution weighting reconstruct.Described measurement report statistics frequency domain information replenishes: according to the frequency domain information that lacks in the report of the category feature additional survey on linear regression model (LRM) and the exclusive cluster dimension of the frequency sweep data statistics.Described frequency sweep data are reconstructed into by telephone traffic distribution weighting: the composition of frequency sweep data is adjusted according to the telephone traffic distribution, made that the frequency sweep data after the reconstruct have the geographical distribution identical with telephone traffic.Interference matrix calculates and to comprise: replenish that measurement report statistics after complete is calculated interference matrix or with the frequency sweep data computation interference matrix of pressing after the telephone traffic distribution weighting reconstruct with frequency domain information, wherein, replenish measurement report statistics after complete with frequency domain information and calculate interference matrix and be first interference matrix and calculate, use to be second interference matrix by the frequency sweep data computation interference matrix after the telephone traffic distribution weighting reconstruct and to calculate.
2. a device that obtains interference matrix is characterized in that, described device comprises:
One presses frequency domain character cluster module: connect the frequency sweep data, with the data based frequency domain character cluster of frequency sweep, obtain category feature;
One telephone traffic distribution estimation module: connect and press frequency domain character cluster module, do linear regression with measurement report statistics and category feature and estimate that the cell telephone traffic amount distributes;
One measurement report statistics complementary module, i.e. first data fusion module: connect the measurement report statistics, press frequency domain character cluster module and telephone traffic distribution estimation module, be used for the frequency domain information that lacks according to linear regression model (LRM) additional survey report statistics, obtain frequency domain information and replenish measurement report statistics after complete; With
One first interference matrix computing module: connect measurement report statistics complementary module, be used for generating interference matrix IM-MR ' according to the measurement report statistics that frequency domain information replenishes after complete.
3. device as claimed in claim 2 is characterized in that, describedly comprises by frequency domain character cluster module:
One data format preliminary treatment submodule: connecting the frequency sweep data, is carrier/interface ratio quantized data and the cluster data with specific cluster space and object with the frequency sweep data conversion;
One cluster submodule: connect data format preliminary treatment submodule, it is the K class that cluster data is gathered, and carries out the hypothesis testing of cluster result; With
One category feature generates submodule: connect data format preliminary treatment submodule and cluster submodule, be used for generating category feature according to cluster result and carrier/interface ratio quantized data.
4. device as claimed in claim 2 is characterized in that, described telephone traffic distribution estimation module comprises:
One Data Matching submodule: connect measurement report statistics and category feature and generate submodule, be used to select category feature and measurement report statistics with identical cluster dimension; With
One regression coefficient estimator module: connect the data matched sub-block, be used for carrying out the linear regression model (LRM) coefficient and find the solution, and the overall fit goodness and the regression coefficient of regression model are carried out hypothesis testing according to the measurement report statistics of coupling and category feature.
5. a device that obtains interference matrix is characterized in that, described device comprises:
One presses frequency domain character cluster module: connect the frequency sweep data, with the data based frequency domain character cluster of frequency sweep, obtain category feature;
One telephone traffic distribution estimation module: connect and press frequency domain character cluster module, do linear regression with measurement report statistics and category feature and estimate that the cell telephone traffic amount distributes;
Sweep the audio data reconstructed module, i.e. second data fusion module: connect the frequency sweep data, press frequency domain character cluster module and telephone traffic distribution estimation module, be used for the frequency sweep data being weighted reconstruct, obtain by the frequency sweep data after the telephone traffic distribution weighting reconstruct according to cell telephone traffic amount distribution valuation; With
One second interference matrix computing module: connect frequency sweep data reconstruction module, be used for according to generating interference matrix IM-DT ' by the frequency sweep data after the telephone traffic distribution weighting reconstruct.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201010133225 CN101848482B (en) | 2010-03-26 | 2010-03-26 | Method and device for acquiring interference matrix |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201010133225 CN101848482B (en) | 2010-03-26 | 2010-03-26 | Method and device for acquiring interference matrix |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101848482A true CN101848482A (en) | 2010-09-29 |
| CN101848482B CN101848482B (en) | 2012-08-15 |
Family
ID=42772908
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201010133225 Expired - Fee Related CN101848482B (en) | 2010-03-26 | 2010-03-26 | Method and device for acquiring interference matrix |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101848482B (en) |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102404773A (en) * | 2011-11-15 | 2012-04-04 | 上海百林通信网络科技有限公司 | Method for automatically optimizing power control parameters |
| CN103916863A (en) * | 2012-12-31 | 2014-07-09 | 中国移动通信集团浙江有限公司 | Frequency optimization method by adding leakage-configuration adjacent cells based on grid positioning |
| CN104918269A (en) * | 2014-03-10 | 2015-09-16 | 中国移动通信集团广东有限公司 | Method and apparatus for constructing interference matrix |
| CN104955099A (en) * | 2014-03-26 | 2015-09-30 | 中国移动通信集团浙江有限公司 | Method and apparatus for analyzing cell interference |
| WO2016090961A1 (en) * | 2014-12-08 | 2016-06-16 | 中兴通讯股份有限公司 | Method and device for network associations analysis |
| CN107786984A (en) * | 2016-08-25 | 2018-03-09 | 中国移动通信集团河北有限公司 | A kind of area interference relation determines method and apparatus |
| CN109788501A (en) * | 2017-11-10 | 2019-05-21 | 中国移动通信集团公司 | 2G+4G mobile network's voice quality joint assessment method and device |
| CN114286375A (en) * | 2021-12-16 | 2022-04-05 | 北京邮电大学 | Mobile communication network interference positioning method |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080207210A1 (en) * | 2005-12-14 | 2008-08-28 | Motorola, Inc. | Apparatus and Method for Frequency Planning for a Cellular Communication System |
| CN101304595A (en) * | 2008-06-20 | 2008-11-12 | 上海华为技术有限公司 | Method and apparatus for acquiring frequency plan |
| CN101404801A (en) * | 2008-11-10 | 2009-04-08 | 中国移动通信集团北京有限公司 | Method for acquiring interference matrix |
-
2010
- 2010-03-26 CN CN 201010133225 patent/CN101848482B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080207210A1 (en) * | 2005-12-14 | 2008-08-28 | Motorola, Inc. | Apparatus and Method for Frequency Planning for a Cellular Communication System |
| CN101304595A (en) * | 2008-06-20 | 2008-11-12 | 上海华为技术有限公司 | Method and apparatus for acquiring frequency plan |
| CN101404801A (en) * | 2008-11-10 | 2009-04-08 | 中国移动通信集团北京有限公司 | Method for acquiring interference matrix |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102404773A (en) * | 2011-11-15 | 2012-04-04 | 上海百林通信网络科技有限公司 | Method for automatically optimizing power control parameters |
| CN103916863B (en) * | 2012-12-31 | 2017-06-20 | 中国移动通信集团浙江有限公司 | The frequency optimization method of the addition miss-configured neighboring cells based on grid positioning |
| CN103916863A (en) * | 2012-12-31 | 2014-07-09 | 中国移动通信集团浙江有限公司 | Frequency optimization method by adding leakage-configuration adjacent cells based on grid positioning |
| CN104918269A (en) * | 2014-03-10 | 2015-09-16 | 中国移动通信集团广东有限公司 | Method and apparatus for constructing interference matrix |
| CN104918269B (en) * | 2014-03-10 | 2019-10-11 | 中国移动通信集团广东有限公司 | A kind of method and apparatus constructing interference matrix |
| CN104955099A (en) * | 2014-03-26 | 2015-09-30 | 中国移动通信集团浙江有限公司 | Method and apparatus for analyzing cell interference |
| CN104955099B (en) * | 2014-03-26 | 2018-08-17 | 中国移动通信集团浙江有限公司 | A kind of method and apparatus of analysis area interference |
| WO2016090961A1 (en) * | 2014-12-08 | 2016-06-16 | 中兴通讯股份有限公司 | Method and device for network associations analysis |
| CN107786984A (en) * | 2016-08-25 | 2018-03-09 | 中国移动通信集团河北有限公司 | A kind of area interference relation determines method and apparatus |
| CN107786984B (en) * | 2016-08-25 | 2021-08-13 | 中国移动通信集团河北有限公司 | Method and device for determining cell interference relationship |
| CN109788501A (en) * | 2017-11-10 | 2019-05-21 | 中国移动通信集团公司 | 2G+4G mobile network's voice quality joint assessment method and device |
| CN109788501B (en) * | 2017-11-10 | 2020-08-18 | 中国移动通信集团公司 | 2G +4G mobile network voice quality joint evaluation method and device |
| CN114286375A (en) * | 2021-12-16 | 2022-04-05 | 北京邮电大学 | Mobile communication network interference positioning method |
| CN114286375B (en) * | 2021-12-16 | 2023-08-18 | 北京邮电大学 | Mobile communication network interference positioning method |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101848482B (en) | 2012-08-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101848482B (en) | Method and device for acquiring interference matrix | |
| Phillips et al. | Practical radio environment mapping with geostatistics | |
| CN105163326B (en) | A kind of cell clustering method and system based on wireless network traffic feature | |
| EP2952047B1 (en) | Method for localizing wireless devices | |
| US20070129086A1 (en) | Method and apparatus for identifying a geographic area having undesirable wireless service | |
| CN107820255B (en) | Improved covariance absolute value cooperative spectrum sensing method | |
| CN101873623A (en) | Automatic frequency optimization method based on measurement report data | |
| EP4184981A1 (en) | Processing method and processing apparatus for saving energy of base station | |
| CN103945531B (en) | Method for WLAN indoor positioning Radio Map updating based on information entropy | |
| CN113015093B (en) | Indoor wireless positioning method based on three-dimensional depth residual error neural network | |
| Qureshi et al. | Enhanced MDT-based performance estimation for AI driven optimization in future cellular networks | |
| CN103974264A (en) | Frequency point optimal selection method | |
| CN110621025B (en) | Equipment model selection method and device | |
| CN111491255B (en) | Backfill longitude and latitude positioning and expanding method, system and storage medium | |
| Gnanasivam et al. | Efficient Centralized Cooperative Spectrum Sensing Techniques for Cognitive Networks. | |
| CN106131953A (en) | A kind of method realizing mobile subscriber location based on frequency weighting in community in the period | |
| CN114079938B (en) | Interference source positioning method, device, equipment and storage medium | |
| CN110650483B (en) | Equipment model selection method and device | |
| CN115734264A (en) | 5G network coverage evaluation method and device, computer readable medium and electronic equipment | |
| CN105682136A (en) | Indoor and outdoor separation method based on outdoor test characteristics | |
| CN110708703B (en) | Equipment model selection method and device | |
| CN113839793A (en) | Indoor distribution fault positioning method and device | |
| CN110753352B (en) | Equipment model selection method and device | |
| CN110636512B (en) | Equipment model selection method and device | |
| CN110719594A (en) | Equipment model selection method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120815 |