WO2014012363A1 - Service hotspot detection method, determination method and locating method and device - Google Patents

Abstract

The present invention is applied to the technical field of wireless networks, and provided are a service hotspot detection method, determination method and locating method and device. The detection method comprises: periodically counting the average data rate of the worst cell user in a cell, the number of blocked users and the number of active users; and if the counted average data rate of the worst cell user is reduced and the number of blocked users and the number of active users are increased, a service hotspot possibly existing in the cell. When detecting the state of a service hotspot, the present invention takes into account the changes of a plurality of network state parameters, including the average data rate of the worst user, the number of blocked users, the number of active users and the like. It is judged that a service hotspot possibly exists in the current cell only when all the data changes meet the requirements. When detecting the service hotspot, the present invention takes into account multi-aspect network state information, and the detection result thereof is more reliable as compared with the existing detection technology.

Description

 Description of a service hotspot detection method, determination method, positioning method and device

 The present invention belongs to the field of wireless network technologies, and in particular, to a method, a determining method, a positioning method and a device for detecting a service hotspot. Background technique

 As the demand for user data rates continues to increase, the trend toward miniaturization of base stations is becoming more and more obvious. Therefore, the dynamics of mobile communication networks are increasing, and the number of network elements to be maintained by operators is increasing rapidly. The bigger it is. In addition, the high mobility of user applications has led to the emergence of hot spots in network services. The self-organizing network SON (self-organization network) technology is proposed to achieve the goal of saving operating costs (OPEX) by achieving as much automation as possible in the planning, deployment, and operation and maintenance phases of the mobile communication network.

 In the operation and maintenance of the existing cellular network, the detection of the service hotspot is the voice reporting of the operator through the network management system (EMS or NMS) or the prediction based on the network load. When the load changes beyond the preset range of the operator, that is, the network performance cannot reach the predetermined value, the operation and maintenance and network management personnel can learn the problems that may occur in the network operation through experience analysis methods, such as whether it is a service hotspot. However, this kind of load-based change cannot accurately determine the service hotspot, because this load change may be caused by strong interference. If the error is judged, it may cause subsequent misoperation. Therefore, the currently detected service hotspot detection method is not suitable for a network in which service hotspots frequently occur due to high mobility of users and increasingly smaller cells. technical problem

In view of the above problems, the present invention provides a method for detecting a service hotspot, which aims to solve the technical problem that the existing service hotspot detection method only relies on load changes to determine service hotspots, resulting in inaccurate service hotspot detection. Technical solution

 The present invention is implemented as a method for detecting a service hotspot, and the method includes the following steps:

 Periodically statistics the average data rate, the number of blocked users, and the number of active users of the worst users in the cell;

 If the average data rate of the worst user of the statistical cell becomes smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell.

 Another object of the present invention is to provide a method for determining a service hotspot, the method comprising the method for detecting a service hotspot, and the method further comprising the following steps:

 Counting the reference signal receiving quality RSRQ of each user in the cell, and clustering the users in the cell;

 Determine whether the current cluster is a service hotspot cluster according to the clustering result.

 Another object of the present invention is to provide a method for locating a service hotspot, and the method includes the method for determining the service hotspot, and further includes the following steps:

 The area where the service hotspot cluster is located is located according to the location information of the user terminal in the service hotspot cluster. Another object of the present invention is to provide a device for detecting a service hotspot, the device comprising: a data statistics unit, configured to periodically count the average data rate, the number of blocked users, and the number of active users of the worst user in the cell;

 The hotspot judging unit is configured to: if the average data rate of the worst user of the cell is smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell.

 Another object of the present invention is to provide a device for determining a service hotspot, the device comprising the device for detecting a service hotspot, and further comprising:

 a user clustering unit, configured to collect a reference signal receiving quality RSRQ of each user in the cell, and perform clustering on users in the cell;

The service hotspot determining unit is configured to determine, according to the clustering result, whether the current cluster is a service hotspot cluster. Another object of the present invention is to provide a positioning device for a service hotspot, the device including the The determining device for the business hotspot further includes:

 The area locating unit is configured to locate the area where the service hotspot cluster is located according to the location information of the user terminal in the service hotspot cluster. Beneficial effect

 In the embodiment of the present invention, a method and an apparatus for detecting, determining, and locating a service hotspot in combination with a network state are provided. When detecting a service hotspot state, the embodiment of the present invention considers changes of multiple network state parameters, including The average data rate of the worst user, the number of blocked users, and the number of active users, etc., can determine that there may be service hotspots in the current cell when all the data changes meet the requirements, and then cluster the users, and determine the services in the cell according to the clustering result. Hotspot cluster, finally locate the location of the business hotspot cluster. The embodiment of the invention can effectively detect a cell in which a service hotspot cluster exists, determine a service hotspot cluster in the cell, and can effectively locate a hotspot area in the network, so that the network triggers corresponding

SON action, to achieve the purpose of saving operating costs. DRAWINGS

1 is a flowchart of a method for detecting a service hotspot according to a first embodiment of the present invention; FIG. 2 is a flowchart of a method for detecting a service hotspot according to a second embodiment of the present invention; FIG. 3 is a third embodiment of the present invention; FIG. 4 is a flowchart of a method for determining a service hotspot according to a fourth embodiment of the present invention; FIG. 5 is a flowchart of a service hotspot according to a fifth embodiment of the present invention; Figure 6 is a flow chart of a device for locating a service hotspot according to a sixth embodiment of the present invention; Figure 7 is a block diagram showing the structure of a device for detecting a service hotspot according to a seventh embodiment of the present invention; FIG. 9 is a block diagram showing the structure of a service hotspot determining apparatus according to a ninth embodiment of the present invention; FIG. 10 is a block diagram showing the structure of a service hotspot determining apparatus according to a ninth embodiment of the present invention; A block diagram of a device for determining a service hotspot; FIG. 11 is a block diagram showing the structure of a device for determining a service hotspot according to an eleventh embodiment of the present invention; 12 is a block diagram showing the structure of a positioning device for a service hotspot according to a twelfth embodiment of the present invention.

 The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It is understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

 In order to explain the technical solutions of the present invention, the following description will be made by way of specific embodiments.

 Embodiment 1:

 FIG. 1 shows a method for detecting a service hotspot according to a first embodiment of the present invention, and only parts related to the embodiment of the present invention are shown for convenience of description.

 Step S101: Periodically count the average data rate, the number of blocked users, and the number of active users of the worst user in the cell.

 In this step, the network status information of users in the cell needs to be periodically collected, including the average data rate of the worst users, the number of blocked users, and the number of active users. Service hotspot detection can be performed in each statistical period, and the service hotspot change information can be flexibly grasped, so that the corresponding SON operation can be automatically performed.

 Step S102: If the average data rate of the worst user of the cell is smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell.

 When there are service hotspots and no service hotspots in the cell, the average data rate, the number of blocked users, and the number of active users of the worst users in the cell are different. When the number of active users in the cell increases, the number of user terminals in the cell indicates the number of user terminals in the cell. If the number of blocked users in the cell increases, it indicates that the mobile network is blocked. When the average number of users in the cell decreases, the communication quality of the user terminal begins to deteriorate. The average number of data of the worst user becomes smaller, and the number of blocked users and the number of active users become larger, and it can be judged that there may be a service hotspot in the current cell.

The embodiment of the present invention can effectively detect that there may be a service hotspot in the cell by acquiring the change of the network state information. Embodiment 2:

 FIG. 2 shows a method for detecting a service hotspot according to a second embodiment of the present invention. For the convenience of description, only parts related to the embodiment of the present invention are shown.

Step S201: The statistical period is divided into W sub-cycles, and the average data rate, the number of blocked users, and the number of active users of the worst user are respectively sampled, and W sample values are respectively obtained: (r _m (nW + l), .. .,rm(n)), (bm(nW + !),..., bm(n)), — W + 1),...,^«( 2)).

Step S202: Dividing a statistical interval of the average user data rate, the number of blocked users, and the number of active users into a statistical interval and normalizing to generate a corresponding histogram, respectively: hM ^r = ( ),... , _m2N (n)), hH d.., h _m ^b ₂ , h _m ^x (n) = (h _m ^x _l (n),...,h _m ^x _2N (n)) , where 2 is the division The number of intervals.

 In the embodiment of the present invention, the statistical period is further divided into W sub-cycles, and the average data rate, the number of blocked users, and the number of active users of the worst user are respectively sampled, and W sample values can be obtained.

Here, the W sample values of the average data rate of the worst users of the cell m are:

 The sample value is further divided into 2^ intervals according to the size, wherein each of the minimum sample value and the maximum sample value is an interval, and the minimum sample value and the maximum sample value are equally divided into -2 intervals, and then the calculation is performed. Enter the number of samples in each interval, normalize the number of samples in each interval with the total number of samples, and obtain a histogram of the average user's average data rate, as follows:

= ... ₂ ""));

 Similarly, the W sample values of the number of blocked users and the corresponding histograms are:

(b _m (nW + lX..., bm(n)) , (η) = ( Λη),..., _{2Ν ϋ} ) _; The w sample values of the active user number and the corresponding histogram are:

(x _m (nW + l\^x _m (n)) , = O )). ―

Step S203, detecting

ε'·) value. Step S204, detecting h (bm(n), Sb) values. Step S205: Detect h(x _m (j , value). Steps S203-S205 are respectively used to detect a change in the average user average data rate, a change in the number of blocked users, and a change in the number of active users.

For detecting the worst user average data rate change:

 When Hl=l, it means that the average user's average data rate is reduced, that is, the determinant function

Var'

h{r _m {n), ε. rm{n)- Ε where r _m ( ) is the arithmetic mean of ε r _m nW + \),...,r _m n)),

E '=!>■* ., the value is expected, =| '— ² * ., is the variance, based on Chebyshev bound, the detected false alarm probability is > 0)|/7 ,(«)| < ε, , In the present invention, it is taken. . ₀₅ .

For detecting changes in the number of blocked users:

H2=l, indicating that the number of blocked users increases. Similarly, the upper limit of the detected false alarm probability is £6, which is 0.05, and the determinant function b _m (n» = where ii) is

6

Replacement page (Article 26 The arithmetic mean of (b _m (n - + 1),..., bm n)), and the E ^b 4 Ku value is expected and is the variance. For detecting changes in the number of active users:

 H3=l, indicating that the number of active users increases. Similarly, the upper limit of the detected false alarm probability is

0.05. Where h(x _m (n), s _x ) =

, where Xm(n) is

算术0- Γ + 1),...,: )) The arithmetic mean, ^ ^χ =∑ ., · The value is expected, 3⁄4 is the variance.

Step S206, when 0^* 2(/^("),8,) + 0:2* 7(6/«("), 6) + (3⁄43* 2( /«("), . >0, Then, it is determined that there may be a service hotspot in the cell, where cn, ² is a weighting factor and satisfies + a2 + a3=l.

 In this step, if the average data rate of the worst 5% of users in the cell becomes large, and the number of call blocking and the number of active users become large, there may be a service hotspot in the cell.

h(r _m (n), £r) + a2* h(bm(n),£b) + a2* h(xm(rt),s _x ) > 0

Ise

 When H=l, it indicates that the cell status is abnormal (there may be a service hotspot); otherwise, the status is normal (there is no service hotspot). Where cn, "2," 3 are weighting factors, satisfying οη + α2 + α3 = 1, which can be set to 0.3, 0.3, and 0.4, respectively.

Replacement page (Article 26 ₇ ^ , . j. , , _ ^ ι r- ^ _Θ is a ⁷ ff real body to buy Shi Wan, this purchase example, the community to use the corpse 4 and the worst 5% of the user, obviously It is also within the scope of the present invention to select other suitable ratios.

 The embodiment of the present invention comprehensively determines whether there is a possible service hot spot in the current cell by detecting data changes of the worst user average data rate, the number of blocked users, and the number of active users, and the judgment accuracy is higher than that of the prior art. Embodiment 3:

 FIG. 3 shows a method for determining a service hotspot according to a third embodiment of the present invention, and only parts related to the embodiment of the present invention are shown for convenience of description.

 Step S301: Periodically count the average data rate, the number of blocked users, and the number of active users of the worst user in the cell.

 Step S302: If the average data rate of the worst user of the cell is smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell.

 The above two steps are used to determine whether there is a service hotspot in the cell in the current statistical period. When it is determined that there may be a service hotspot, it is necessary to further determine whether there is a service hot spot. Proceed as follows:

 Step S303: The reference signal of each user in the cell receives the quality RSRQ, and the users in the cell are clustered.

 Step S304: Determine, according to the clustering result, whether the current cluster is a service hotspot cluster.

 In the embodiment of the present invention, the user is clustered based on the user RSRQ, and the user is clustered.

The RSRQ probability distribution is used to identify the similarity of users, thereby determining whether there is a service hotspot in the cell. Embodiment 4:

 FIG. 4 shows a method for determining a service hotspot according to a fourth embodiment of the present invention. For the convenience of description, only parts related to the embodiment of the present invention are shown.

 Step S401: Periodically count the average data rate of the worst user in the cell, and the number of blocked users.

8

Replacement page (Article 26 ^

 And ^^ with corpse;

 Step S402: If the average data rate of the worst user of the cell is smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell.

 Steps S401-S402 are the same as steps S301-S302, and are not described herein again.

 Step S403: Counting the RSRQ of each user in the current statistical period, dividing the statistical period into W sub-periods, sampling the RSRQ to obtain W RSRQ sample values, and dividing the W RSRQ sample values into statistical intervals and returning to the RSRQ. Histogram.

 This step is used to obtain the RSRQ histogram of the user. The acquisition method is the same as that of the implementation 2. First, the current statistical period is divided into W sub-cycles to sample the RSRQ, and W sample values are obtained. After normalization, an RSPR histogram is generated, for example, Users k and j can get a histogram of two users = (3⁄4, ... and ... ^ ).

 Step S404: If it is determined that there is a service hotspot in the current cell, all users are clustered according to the RSRQ histogram of the user, and the optimal clustering is found, so that the KL distance between the user and the cluster core of the corresponding cluster is the smallest.

In this step, all users are clustered based on the user RSRQ histogram. The hypothesis sum (?) represents the RSRQ probability distribution of users k and j, respectively, and its KL distance ( Kullback-Leibler (KL) divergence ) is:

The histograms of the RSRQs of users k and j are and = (3⁄4ΐ'... H then the KL distance can be approximated as:

 It is assumed that the user can be divided into C clusters, 2 = ( ,···, representing the cluster core of C clusters, and the probability that the user k is assigned to the cluster j (ie, the cluster member function) is ^' e [0, l] , then the goal of clustering is to find the optimal

Replacement page (Article 26) ^ _, , _(J _ _ ^ , ^ . . ^ , , ^ ,, „"

 The cluster heart and the sub-entertainment member function make the KL distance between the corpse and the cluster core of the cluster still dare to 'J,:

 Min J (U,Q) s.t.J(U,

i )=£ io _g

Where 0 = ( ,··', _Λ ) is the cluster core of cluster j, that is, its RSRQ histogram.

 Let oAU, 2) = 0 and ukjJ (U,Q) = ^ get the optimal clustering:

 Step S405: Obtain a cluster number and a cluster core of each cluster.

 In this step, the correct number of clusters C and the cluster core of each cluster can be obtained by initializing the clustering, as follows:

 The number of S405K initialization clusters is a large value N;

S4052: Calculate the mean and variance of each user RSRQ histogram. Taking user k as an example, the RSRQ histogram is Wk ⁼ (3⁄4, ... ^h j ), and the mean and variance are:

Mean: E _k ^q ^^i*h _k ^q _j

Variance:

10 Replacement page (Article 26 _τ . .. , '丄, , , . , ^ , , _t

 S4053, 标准 in standard K-means, for mean and 10,000 data

(^ k = ..., Km), Xk = [El Var _k ^q for clustering, iteratively updating the cluster heart [v , y = l,..., N} _? until convergence:

 54054. If the distance between the cluster cores of any two clusters is less than a certain threshold, the N value is decremented by 1, and the process returns to step S4053 until the cluster center distance of any two clusters is sufficiently large, and the clustering is completed. The value is the number of clusters C;

 54055. For each cluster, find the closest point to the cluster heart, and its corresponding history is the cluster core of the cluster in the clustering algorithm.

 Step S406: Perform Hoeffding test on the user members in the cluster. When the number of remaining members is greater than a predetermined threshold, the current cluster may be determined to be a service hotspot cluster.

 After the clustering is completed, the similarity of users in the same cluster is not necessarily large enough, so it cannot be determined that the cluster belongs to a service hotspot cluster. In this embodiment, the Hoeffding test is also needed to determine the service hotspot cluster: Assume ^ Α^ β') represents a decision criterion based on cluster 0', which contains a series of probabilities

Λ , if * Ε ^Λ then user k is deleted from cluster β'. When the number of remaining members is greater than a predetermined threshold, it may be determined that the current cluster is a service hotspot cluster.

The optimal decision criteria are based on the Hoeffding test:

Is an adjustable factor, by controlling this factor, you can get the probability of the decision error:

11

Replacement page (Article 26 Where W is the number of periods of the histogram data, when ^ ⁷ → ∞, D{Pk Q) → 0 _o The embodiment of the present invention provides a specific user clustering method, and the cluster is further determined by Hoeffding test for each cluster. Whether it is a business hotspot.

Embodiment 5:

 FIG. 5 shows a method for determining a service hotspot according to a fifth embodiment of the present invention. For the convenience of description, only parts related to the embodiment of the present invention are shown.

 Step S501: Periodically count the average data rate, the number of blocked users, and the number of active users of the worst user in the cell.

 Step S502: If the average data rate of the worst user of the cell is smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell.

 Step S503: Counting the RSRQ of each user in the current statistical period, dividing the statistical period into W sub-periods, sampling the RSRQ to obtain W RSRQ sample values, and dividing the W RSRQ sample values into statistical intervals and normalizing. Generate a user RSRQ histogram.

 Step S504: If it is determined that there is a service hotspot in the current cell, all users are clustered according to the user RSRQ histogram, and the optimal clustering is found, so that the KL distance between the user and the cluster core of the corresponding cluster is the smallest.

 Step S505: Obtain a cluster number and a cluster core of each cluster.

 Steps S501-S505 are the same as steps S401-S405, and are not described herein again.

 Step S506: When the PRB utilization rate of the total physical resource block of the cell is less than 100%, and the PRB utilization rate of the members in the cluster is greater than a certain threshold Thrl, or when the total PRB utilization rate of the cell is equal to 100%, and the satisfaction of the members in the cluster Less than a certain threshold Thr2, it can be determined that the current cluster is a business hotspot cluster.

 When the total PRB utilization of a cell is less than 100%, the cluster load can be measured by PRB utilization:

LCI = U _PRB r = ^ > Thrl

 c PRB

12 Replacement page (Article 26 . _rt ^ , , . _α , . „ ^ , ^ * , -r , ., If the family-loaded LCI (ie, the PRB utilization rate in the cluster) is greater than a 疋U-limit Thrl, the cluster is a hotspot cluster.

When the total PRB utilization of the cell is equal to 100%, the cluster load is defined as: LC2 = k * S + S < Thr2 s 实际 actual rate per user _ Non - GBR user average rate where ^GM ― ∑ per 6^ ^ user Guaranteed rate and - ^GBR = Non-GBR user ^ ffiW average is GBR ( Guaranteed Bit Rate) service and Non-GBR (non-guaranteed bit rate) service user satisfaction; k is greater than or equal to 1 The weight coefficient can be set to 2. Aggregated Maximum Bit Rate (AMBR) is the aggregated maximum bit rate of the Non-GBR service.

 If the cluster load LC2 (that is, the satisfaction of the members in the cluster) is less than a certain threshold Thr2, the cluster is a hot cluster.

 This embodiment provides another method for determining a service hotspot. The difference from the fourth embodiment is that the embodiment further determines whether the current cluster is a service hotspot cluster by using the total PRB utilization rate and the cluster load value.

Example 6:

 FIG. 6 shows a method for locating a service hotspot according to a sixth embodiment of the present invention. For the convenience of description, only parts related to the embodiment of the present invention are shown.

 Step S601: Periodically count the average data rate of the worst user in the cell, the number of blocked users, and the number of active users. Step S602: If the average data rate of the worst user of the cell is smaller, and the number of blocked users and activities are blocked. If the number of users becomes large, there may be service hotspots in the cell;

 Step S603: The reference signal of each user in the cell is received by the RSRQ, and the users in the cell are clustered.

 Step S604: Determine, according to the clustering result, whether the current cluster is a service hotspot cluster.

 Steps S601-S604 are used to determine whether the current cluster is a service hotspot cluster, which is the same as that in the third embodiment, and details are not described herein again.

13 Replacement page (Article 26 S6, according to the position of the corpse terminal in the business hotspot cluster, the area where the entertainment is located.

 In the embodiment of the present invention, location information of the user terminal may be obtained by using various methods. If the user terminal includes the GPS module and is enabled, the location information of the service hotspot cluster may be located by acquiring the location information uploaded by the GPS module in the user terminal in the service hotspot cluster; if the user terminal does not include the GPS module or the GPS module currently If the vacancy is not available, the location of the user terminal can be estimated by obtaining the radio frequency fingerprint information reported by the user terminal in the service hotspot cluster, and the area where the service hotspot cluster is located is located. After the location of the service hotspot in the network is located, it is easy to select the correct optimization action.

 To estimate the location of the user terminal by obtaining the radio frequency fingerprint information reported by the user terminal in the service hotspot cluster, the steps for locating the service hotspot cluster are as follows:

 S605 obtains measurement report data reported by the user terminal.

 During the statistical period, the user reports multiple measurement reports, especially the RSRP (Reference Signal Receiving Power) measurement level.

 S6052: Select a grid point with the smallest matching degree according to the matching degree between the measurement report data and the corresponding data of the feature database.

 In this embodiment, the matching degree between the measurement report data and the corresponding data of the feature library is calculated, and one grid point whose matching degree (Sr ) is the smallest (ie, the best) is selected. For each measurement report data and each database data, starting from the first cell level, the level of the same cell in the database is sequentially searched, and the level difference is calculated, and the sum of the differences is the matching degree. Therefore, if there are X user terminals in the service hotspot cluster, then there will be X matching degrees, and the grid points with the smallest matching degree are selected.

An example is as follows (the number indicates the cell number, and the level values are arranged in descending order):

 Then Sr = |S5-R5| + |S2-R2| + |S1- 1 | + |S3-R3| + |S7-R7| + |S6-R6

 Some cell levels exist in the measurement report but do not exist in the database cell list, or exist in the database but do not exist in the measurement report, then you can add a penalty factor on Sr to make the search

14

Replacement page (Article 26 Sr = |S5-R5| + |S2-R2| + |S1_R1 | + |S3-R3| + |S7- R7| + |S6-R6| + |S8-R6| + |S3-R4|.

 In this implementation, the feature library needs to be established and initialized first, and updated in time. The initialization feature library and the feature database update method are as follows:

 Initial feature library: The corrected model (preferably the sub-regional area correction) used in the network gauge is calculated, and multiple level data of each geographical position in the grid is calculated in each grid to generate a probability distribution. .

 Feature Library Update: Continuously put the data with location information into the grid level database, and filter out the older data according to the data, and update the probability distribution.

 S6053. Position the area where the service hotspot cluster is located according to the location information of the one grid point with the smallest matching degree.

 In the embodiment of the present invention, the location of the user terminal is estimated by the information reported by the GPS module or the radio frequency handprint information, and the area where the service hotspot cluster is located is located. Example 7:

 FIG. 7 shows a device for detecting a service hotspot according to a seventh embodiment of the present invention, and only parts related to the embodiment of the present invention are shown for convenience of description.

 In this embodiment, the device includes:

 The data statistics unit 701 is configured to periodically count the average data rate, the number of blocked users, and the number of active users of the worst user in the cell.

 The hotspot judging unit 702 is configured to: if the average data rate of the worst user of the cell is smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell. Example 8:

 FIG. 8 shows a device for detecting a service hotspot according to an eighth embodiment of the present invention, and only parts related to the embodiment of the present invention are shown for convenience of description.

 In this embodiment, the device includes:

15

Replacement page (Article 26 _{ηΛ 1} m ^ m η - * _tt . 兀 兀 801, used for periodically counting the rate of corpse use, the number of blocked users and the number of active users in the cell.

 The hotspot judging unit 802 is configured to: if the average data rate of the worst user of the cell is smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell.

 The data statistics unit 801 includes:

The sample obtaining module 8011 is configured to divide the statistical period into W sub-periods, and respectively sample the average data rate, the number of blocked users, and the number of active users of the worst user, and obtain W sample values respectively: {r _m {nW +\ ...,r _m {n)), (b _m {nW + \ ...,b _m n)), _m {nW + \),...,x _m (n))■ Histogram generation module 8012. The sample values of the average user data rate, the number of blocked users, and the number of active users are divided into statistical intervals and normalized to generate corresponding histograms, respectively:

K d d'-'D )), = ("), '·', (")), = , where 2 ^Λ ' is the number of intervals divided.

The hotspot determining unit 802 includes: a worst user rate detecting module 8021, configured to detect <("), a value,

>Q indicates that the average data rate of the worst user in the cell becomes smaller, where - + l),...,r _w (the arithmetic mean

^) ² * ., is the variance, Sr is the upper limit of the false alarm probability; the blocking user number detection module 8022 is used to detect the value, if / 7( ("), "^)>0 indicates that the number of blocked users becomes larger, wherein Where bmi ^ is

The arithmetic mean of (b _m (nW + \),...,b _m (n)), E ^b = value period

16 Replacement page (Article 26 , '

Var ^b =^(i~E ^h ) ² *h _m ^h is the variance, _£ b is the upper limit of the false alarm probability; the active user number detection module 8023 is used to detect the value of 7(x _m O),

 Var

Indicates that the number of active users has increased, where 20("), .) = (") - where x _w (") is

The arithmetic mean of (x _m (nW + l),...,x _m (n)), E ^x =∑ i * Ki is the expected value,

Var ^x ^(i~E ^x ) ² *h _m ^x _i is the variance, which is the upper limit of the false alarm probability; the hotspot determination module 8024 is used to be cn * 2(ΓΜ("), ε·) + α 2 *

+ a3*h(Xm(n),Sx) _>Qi determines that there may be service hotspots in the cell, where Ctl, "2, CC3 are plus 4 and factor, satisfying cn + a2 + cn=l.

 Example 9:

 FIG. 9 shows a device for determining a service hotspot according to a ninth embodiment of the present invention. For the convenience of description, only parts related to the embodiment of the present invention are shown.

 The data statistics unit 901 is configured to periodically count the average data rate, the number of blocked users, and the number of active users of the worst user in the cell.

 The hotspot judging unit 902 is configured to: if the average data rate of the worst user of the cell is smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell.

 The user clustering unit 903 is configured to collect a reference signal receiving quality RSRQ of each user in the cell, and perform clustering on users in the cell.

 The service hotspot determining unit 904 is configured to determine, according to the clustering node, whether the current cluster is a service hot spot cluster. Example 10:

17

Replacement page (Article 26 ^, . , . _ _T FIG. 10 shows a service hotspot provided by the tenth embodiment of the present invention. The tool J is convenient for the description only showing the parts related to the embodiment of the present invention.

 The data statistics unit 101 is configured to periodically count the average data rate, the number of blocked users, and the number of active users of the worst user in the cell;

 The hotspot determining unit 102 is configured to: if the average data rate of the worst user of the cell is smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell.

 The user clustering unit 103 is configured to collect a reference signal receiving quality RSRQ of each user in the cell, and perform clustering on users in the cell;

 The service hotspot determining unit 104 is configured to determine, according to the clustering result, whether the current cluster is a service hotspot cluster.

 The user clustering unit 03 includes:

 The RSRQ obtaining module 1031 is configured to count the RSRQ of each user in the cell in the current statistical period, divide the statistical period into W sub-periods, sample the RSRQ to obtain W RSRQ sample values, and divide the W RSRQ sample values into statistics. Generate a user RSRQ histogram after interval and normalization;

 The optimal clustering obtaining module 1032 is configured to: if it is determined that there is a service hotspot in the current cell, cluster all users according to the user RSRQ histogram, find an optimal cluster, and make a KL distance between the user and the cluster core of the corresponding cluster. Minimum

 The number of clusters and the cluster new acquisition module 1033 are used to obtain the number of clusters and the cluster core of each cluster. The service hotspot determining unit 104 includes:

 The cluster member number determining module 1041 is configured to perform Hoeffding test on the user members in the cluster. When the number of remaining members is greater than a predetermined threshold, the current cluster may be determined to be a service hotspot cluster. Example 11:

 Fig. 11 shows a device for determining a service hotspot according to an eleventh embodiment of the present invention, and only parts related to the embodiment of the present invention are shown for convenience of explanation.

 The data statistics unit 11 1 is configured to periodically calculate the average data rate of the worst user in the cell in the cell,

Replacement page (Article 26 „ _π , _m „

 丞 丞 Γ 动 动 动 动 动

 The hotspot judging unit U2 is configured to: if the average rate of the worst user of the cell is smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell.

 The user clustering unit 1 13 is configured to collect a reference signal receiving quality RSRQ of each user in the cell, and perform clustering on users in the cell;

 The service hotspot determining unit 1 14 is configured to determine, according to the clustering node, whether the current cluster is a service hot spot cluster.

 The service hotspot determining unit 1 14 includes:

The cluster load determining module 1 141 is configured to: when the total physical resource block PRB utilization rate of the cell is less than 100%, and the _PRB utilization rate of the intra-cluster member is greater than a certain threshold Thrl, or when the total PRB utilization rate of the cell is equal to 100%, and The satisfaction of members in the cluster is less than a certain threshold hr2, and the current cluster can be determined to be a business hotspot cluster. Example 12:

 FIG. 12 is a diagram showing a positioning device for a service hotspot according to a twelfth embodiment of the present invention, and only parts related to the embodiment of the present invention are shown for convenience of explanation.

 The data statistics unit 121 is configured to periodically count the average data rate, the number of blocked users, and the number of active users of the worst user in the cell;

 The hotspot determining unit 122 is configured to: if the average data rate of the worst user of the cell is smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell.

 The user clustering unit 123 is configured to collect a reference signal receiving quality RSRQ of each user in the cell, and perform clustering on users in the cell;

 The service hotspot determining unit 124 is configured to determine, according to the clustering node, whether the current cluster is a service hot spot cluster;

 The area locating unit 125 is configured to locate the area where the service hotspot cluster is located according to the location information of the user terminal in the service hotspot cluster.

19

Replacement page (Article 26 . , , , - ,

 ^ 3^ i i J3⁄4 单 几 a few 125 packs ^":

 The GPS area positioning module is configured to acquire location information uploaded by the GPS module in the user terminal in the service hotspot cluster, and locate the area where the service hotspot cluster is located;

 Or,

 The radio frequency handprint area locating module is configured to obtain the radio frequency fingerprint information reported by the user terminal in the service hotspot cluster, and estimate the location of the user terminal, and locate the area where the service hotspot cluster is located.

 The radio frequency handprint area positioning module includes:

 a measurement report obtaining component, configured to acquire measurement report data reported by the user terminal;

 An optimal grid point selection component is configured to select a grid point with the smallest matching degree according to the matching degree between the measurement report data and the corresponding data of the feature library;

 The grid point area locating component is configured to locate the area where the service hotspot cluster is located according to the location information of the one grid point with the smallest matching degree.

 In the embodiment of the present invention, a method and an apparatus for detecting, determining, and locating a service hotspot in combination with a network state are provided. When detecting a service hotspot state, the embodiment of the present invention considers changes of multiple network states, including the most The average data rate of the poor user, the number of blocked users, and the number of active users, etc., can determine that there may be service hotspots in the current cell when all the data changes meet the requirements, and then cluster the users, and determine the service hotspots in the cell according to the clustering result. Cluster, finally locate the location of the business hotspot cluster. The embodiment of the invention can effectively detect the cell in which the service hotspot cluster exists, determine the service hotspot cluster in the cell, and effectively locate the hotspot area in the network, so that the network triggers the corresponding SON action and achieves the purpose of saving operation cost.

 It will be understood by those skilled in the art that all or part of the steps of the foregoing embodiments may be implemented by a program to instruct related hardware, and the program may be stored in a computer readable storage medium. Storage media, such as ROM/RAM, disk, CD, etc.

 The above is only the preferred embodiment of the present invention, and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection of the present invention. Within the scope.

20

 Replacement page (Article 26

Claims

 A method for detecting a service hotspot, the method comprising:

 Periodically statistics the average data rate, the number of blocked users, and the number of active users of the worst users in the cell;

 If the average data rate of the worst user of the statistical cell becomes smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell.

 The method for detecting a service hotspot according to claim 1, wherein the step of periodically counting the average data rate, the number of blocked users, and the number of active users of the worst user of the cell in the cell includes:

The statistical period is divided into W sub-cycles, and the average data rate, the number of blocked users, and the number of active users of the worst user are respectively sampled, and W sample values are obtained respectively: (r _m (nW + V},...,r _m (n)), (bm(n -W + 1),..., bm(n)), (x _m (nW + l),...,x _m (n));

The statistical values of the average user's average data rate, the number of blocked users, and the number of active users are divided into statistical intervals and normalized to generate corresponding histograms, respectively: h _m ^r (n) = (h _m ^r _l (n), h _2N (nj), in) = (h _m ^b ^n),...,h _m ^b _2N (w), = ( ("), ..., , ₂ w (")) , where 2W is the number of intervals divided.

The method for detecting a service hotspot according to claim 2, wherein the average data rate of the worst user of the statistical cell becomes smaller, and the number of blocked users and the number of active users become larger. Then, there may be a service hotspot step in the cell, which specifically includes: detecting a value of h(r _m (n), £ _r ), if h(r _m (n), & _r ) >o indicates an average data of the worst user of the cell The rate becomes smaller, where h(r _m (n» = 7 _m {n)-E ^r , where rmO) is

2 ^N

The arithmetic mean of (r _m (nW + l),...,r _m (n)), E ^r =∑ * j is the expected value, i=l

2 ^N

^r ^r =∑( - E ^r f * K _m . is the variance, which is the upper limit of the false alarm probability;

 i=l

Detect the value of h(b _m (n), £b). If h(b _m (n), >0 indicates that the number of blocked users becomes larger, where h(bm(n), Sb) = bm(n)-E ^t , where 1⁄20z) is ( (w - W + l),... _m (w)

2 ^N 2 ^N

Arithmetic mean, =∑^ ,, value is expected, is variance, Sb i=l i=l

The upper limit of the probability of false alarms;

Detect h(x _m (n), value, if h(x _m (n), ε _χ ) > ο indicates that the number of active users becomes larger, where Ϊ—W + 1),...,J )

E ^x f *h _m ^x _i is the variance, ε _χ

The upper limit of the probability of false alarm; when * h(r _m (n), &r) +

h(bm(n), Sb) + Of3 * h(x _m (n), ε _χ ) >o , then it is determined that there may be service hotspots in the cell, where Cn, O 2 are weighting factors, satisfying «1+^ 2 + ^3=1.

The method for detecting a service hotspot according to any one of claims 1 to 3, wherein the worst user of the cell is the worst 5% of the users in the cell.

 A method for determining a service hotspot, the method comprising the method for detecting a service hotspot according to any one of claims 1 to 4, further comprising:

 Counting the reference signal receiving quality RSRQ of each user in the cell, and clustering the users in the cell;

 Determine whether the current cluster is a service hotspot cluster according to the clustering result.

6. The method for determining a service hotspot according to claim 5, wherein the statistics are The reference signal of each user in the cell receives the quality RSRQ, and performs a clustering step for the users in the cell, including:

 In the current statistical period, the RSRQ of each user in the cell is counted, and the statistical period is divided into W sub-cycles, and the RSRQ samples are sampled to obtain W RSRQ sample values, and the W RSRQ sample values are divided into statistical intervals and normalized to generate. User RSRQ histogram;

 If it is determined that there is a service hotspot in the current cell, all users are clustered according to the user RSRQ histogram, and the optimal clustering is found, so that the KL distance between the user and the cluster core of the corresponding cluster is the smallest;

 Get the number of clusters and the cluster core of each cluster.

 The method for determining a service hotspot according to claim 5 or 6, wherein the step of determining whether the current cluster is a service hotspot cluster according to the clustering result comprises:

 Performing a Hoeffding test on the user members in the cluster, and determining that the current cluster is a service hotspot cluster when the number of remaining members is greater than a predetermined threshold;

 Or,

 When the PRB utilization rate of the total physical resource block of the cell is less than 100%, and the PRB utilization rate of the members in the cluster is greater than a certain threshold Thrl, or when the total PRB utilization rate of the cell is equal to 100%, and the satisfaction of the members in the cluster is less than a certain threshold Thr2, can determine that the current cluster is a business hotspot cluster.

 A method for locating a service hotspot, the method comprising the method for determining a service hotspot according to any one of claims 5-7, further comprising:

 The area where the service hotspot cluster is located is located according to the location information of the user terminal in the service hotspot cluster.

The method for locating a service hotspot according to claim 8, wherein the step of locating the area where the service hotspot is located according to the location information of the user terminal in the service hotspot cluster includes:

 Obtaining the location information uploaded by the GPS module in the user terminal in the service hotspot cluster, and locating the area where the service hot spot cluster is located;

 Or,

Obtaining the radio frequency fingerprint information reported by the user terminal in the service hotspot cluster to estimate the location of the user terminal, The area where the service hotspot cluster is located.

 The method for locating a service hotspot according to claim 9, wherein the obtaining the radio frequency fingerprint information reported by the user terminal in the service hotspot cluster to estimate the location of the user terminal, and locating the region where the service hotspot cluster is located, Specifically include:

 Obtaining measurement report data reported by the user terminal;

 Selecting a grid point with the smallest matching degree according to the matching degree between the measurement report data and the corresponding data of the feature database;

 The location of the service hotspot cluster is located according to the location information of the one grid point with the smallest matching degree.

 11. A device for detecting a service hotspot, wherein the device comprises:

 a data statistics unit, configured to periodically count the average data rate, the number of blocked users, and the number of active users of the worst user in the cell;

 The hotspot judging unit is configured to: if the average data rate of the worst user of the cell is smaller, and the number of blocked users and the number of active users become larger, there may be a service hotspot in the cell.

 The device for detecting a service hotspot according to claim 11, wherein the data statistics unit comprises:

The sample obtaining module is configured to divide the statistical period into W sub-periods, and respectively sample the average data rate, the number of blocked users, and the number of active users of the worst users, and obtain W sample values respectively: (r _m (nW + l) ,...,r _m (n)), (bm(nW + l),...,bm(n)), (x _m (nW + l),...,Xm(n)); columnar a graph generating module, configured to divide a statistical interval of the average data rate, the number of blocked users, and the number of active users of the worst user into a statistical interval and normalize to generate a corresponding histogram, respectively: (n, = (h) _m ^r _l (n,"."h: _2N (n,,, 0) = ( »,.." ₂ „ 0》, = (3⁄4 ( ),..., , where 2 is the number of intervals divided.

The device for detecting a service hotspot according to claim 12, wherein the hot spot determination unit comprises: The worst user rate detection module is used to detect the «), ε;·) value. If ζ(Γ«), ε 〉0 indicates that the average data rate of the worst user in the cell becomes smaller, where h( "« & _r ) = r _m (n) - E

Where rm(n) is the arithmetic mean of (r O - W + l),...,r ( )), E ^r =∑, ^K is i=l expectation, Var'' = ^(i— E ¹ ^ ² * h _m ' i is the square, and V is the upper limit of the false alarm probability;

i=l blocking user number detection module, used to detect ζ(Κ ϊ), ) value, if h(b _m (n), Sb)>o indicates that the number of blocked users becomes larger, where k(b _m (n), , where is

The arithmetic mean of (b _m {nW + \ ..., b _m {n)), E ^b =∑^3⁄4,, the value is expected, i=l

2 ^N

v _ar ^b = iE ^b m is the variance, and _£b is the upper limit of the false alarm probability;

Active user number detection module, used to detect the value, if / £·χ)>0 indicates that the number of active users becomes larger, where h(x _m (n), ε _χ ) = xm(n) - E where XmO) is

The arithmetic mean of (x _m (nW + l),...,x _m (n)), where the E ^x value is expected,

Var ^x =^(iE ^x ) ² ^h: _i is the variance, and £x is the upper limit of the false alarm probability;

 Hotspot judgment module, used when 1 * hi min), 8r) + «2* h(bm(n), 6b) + «3* h(Xm(n), Sx)

>0, it is determined that there may be a service hotspot in the cell, where "1, on is a weighting factor, satisfying +αι + a^=\.

14. A device for detecting a service hotspot according to any of claims 11-13, characterized in that The worst user of the cell is the worst 5% user in the cell.

 A device for determining a hotspot of a service, characterized in that the device comprises the device for detecting a service hotspot according to any one of claims 11-14, further comprising:

 a user clustering unit, configured to collect a reference signal receiving quality RSRQ of each user in the cell, and perform clustering on users in the cell;

 The service hotspot determining unit is configured to determine, according to the clustering result, whether the current cluster is a service hotspot cluster.

The device for determining a service hotspot according to claim 15, wherein the user clustering unit comprises:

 The RSRQ obtaining module is configured to calculate the RSRQ of each user in the cell in the current statistical period, divide the statistical period into W sub-periods, sample the RSRQ to obtain W RSRQ sample values, and divide the W RSRQ sample values into statistical intervals. And normalized to generate a user RSRQ histogram;

 An optimal clustering acquisition module, configured to: if the current cell may have a service hotspot, ^^ according to the user RSRQ histogram, cluster all users, find an optimal cluster, and make the KL of the user and the cluster of the corresponding cluster The minimum distance;

 The number of clusters and the new cluster acquisition module are used to obtain the number of clusters and the cluster core of each cluster.

 The device for determining a service hotspot according to claim 15 or 16, wherein the service hotspot determining unit comprises:

 a cluster member number determining module, configured to perform Hoeffding test on user members in the cluster, and when the remaining member number is greater than a predetermined threshold, determining that the current cluster is a service hotspot cluster;

 Or,

 a cluster load determining module, configured to: when a PRB utilization rate of a total physical resource block of the cell is less than 100%, and a PRB utilization rate of a member of the cluster is greater than a certain threshold Thrl, or when a total PRB utilization rate of the cell is equal to 100%, and within the cluster If the member's satisfaction is less than a certain threshold Thr2, the current cluster can be determined to be a business hotspot cluster.

The device for determining a service hotspot according to any one of claims 15-17, further comprising: The area locating unit is configured to locate the area where the service hotspot cluster is located according to the location information of the user terminal in the service hotspot cluster.

 The device for locating a service hotspot according to claim 18, wherein the location locating unit comprises:

 The GPS area positioning module is configured to acquire location information uploaded by the GPS module in the user terminal in the service hotspot cluster, and locate the area where the service hotspot cluster is located;

 Or,

 The radio frequency handprint area locating module is configured to obtain the radio frequency fingerprint information reported by the user terminal in the service hotspot cluster, and estimate the location of the user terminal, and locate the area where the service hotspot cluster is located.

 The device for positioning a service hotspot according to claim 19, wherein the radio frequency handprint area positioning module comprises:

 a measurement report obtaining component, configured to acquire measurement report data reported by the user terminal;

 An optimal grid point selection component is configured to select a grid point with the smallest matching degree according to the matching degree between the measurement report data and the corresponding data of the feature library;

 The grid point area locating component is configured to locate the area where the service hotspot cluster is located according to the location information of the one grid point with the smallest matching degree.