CN115499863A

CN115499863A - Cell interruption detection method and device based on adjacent visibility hypergraph change

Info

Publication number: CN115499863A
Application number: CN202211110442.2A
Authority: CN
Inventors: 孙长印; 张燕燕; 王军选; 江帆; 万欣; 毛亚宁
Original assignee: Xian University of Posts and Telecommunications
Current assignee: Xian University of Posts and Telecommunications
Priority date: 2022-09-13
Filing date: 2022-09-13
Publication date: 2022-12-20

Abstract

The invention discloses a cell interruption detection method and a device based on adjacent visibility hypergraph change, which receive measurement information sent by active users in the coverage area of a current base station and construct a neighbor cell unit list; constructing an adjacent visibility hypergraph according to the adjacent cell unit list; calculating a change vector of the adjacent visibility hypergraph based on the adjacent visibility hypergraph at different time instants; the change vector is used as input information of a network interrupt classifier, and the network state of the cell to be detected is determined according to the classifier; the invention constructs the adjacent visibility hypergraph through the adjacent cell unit list of the active user, can combine the detection information of multiple users to improve the detection precision, and obtains signal detection data with stronger objectivity by calculating the change vector of the visibility hypergraph at different moments, thereby further improving the network signal detection precision.

Description

Cell interruption detection method and device based on adjacent visibility hypergraph change

Technical Field

The invention relates to the technical field of network detection, in particular to a cell interruption detection method and device based on adjacent visibility hypergraph change.

Background

In order to meet the requirements of different customers and application scenarios on differentiated service capability of a communication network, the scale and complexity of the 5G network are increased sharply by introducing an NFV (network virtualization) technology into a core network, introducing an SDN (software defined network) virtualization technology into a transmission network, introducing a super-large-scale antenna system into a radio access network and the like. Meanwhile, the service management mode of the 5G slice enables the network deployment to be more dynamic and complex. These all promote the network operation management degree of difficulty by a wide margin.

The cell interruption detection is a way for sensing the network state in real time and automatically positioning the network fault. The conventional cell outage detection is mainly based on signal measurement at a user cell level, and determines whether a cell is in an outage state by comparing an RSRP value and an RSRQ value received by a user with corresponding threshold values respectively. However, threshold values of various indexes in the conventional method are usually specified manually, and with the increase of network coverage, the variation range of wireless signals is large, and the threshold values are difficult to determine, which can cause the problem of low detection accuracy.

Disclosure of Invention

The invention aims to provide a cell interruption detection method and device based on the change of an adjacent visibility hypergraph, which are used for establishing the adjacent visibility hypergraph, taking the variable quantity of the adjacent visibility hypergraph as detection data, taking objective data as detection quantity, avoiding setting a threshold value and improving the detection precision.

The invention adopts the following technical scheme: a cell interruption detection method based on adjacent visibility hypergraph change comprises the following steps:

receiving measurement information sent by active users in the coverage area of the current base station and constructing a neighbor cell unit list;

constructing an adjacent visibility hypergraph according to the adjacent cell unit list;

calculating a change vector of the adjacent visibility hypergraph based on the adjacent visibility hypergraph at different time instants;

and taking the change vector as input information of a network interrupt classifier, and determining the network state of the cell to be detected according to the classifier.

Further, the receiving measurement information sent by active users within the coverage area of the current base station and constructing the neighbor cell unit list includes:

receiving a physical layer cell identifier which is sent by an active user and meets a configuration request;

inquiring whether the physical layer cell identification exists in a neighbor cell list of the current base station;

when the physical layer cell identification is not in the adjacent cell list, sending a global cell identification measurement configuration request to an active user;

receiving a global cell identifier which is sent by an active user and meets a configuration request;

the global cell identity is added to the neighbor cell unit list.

Further, the receiving of the physical layer cell identifier sent by the active user and satisfying the configuration request comprises:

when the active user detects that the beam signal strength of the serving cell is lower than a first threshold value, detecting second signal strength of all other received beam signals;

and when the second signal strength is larger than a second threshold value, taking the corresponding beam signal as a signal meeting the configuration request.

Further, before constructing the adjacency visibility hypergraph according to the neighbor cell unit list, the method comprises the following steps:

and receiving a neighbor cell unit list sent by a neighbor cell in the neighbor cell unit list.

Further, constructing the adjacency visibility hypergraph according to the neighbor cell unit list comprises:

establishing an adjacent visibility hypergraph by taking a beam ID in the adjacent cell unit list as a node of the adjacent visibility hypergraph, taking all visible beams in a cell as a hypergraph of the adjacent visibility hypergraph and taking a visible beam of each active user as a conventional edge;

expressing the adjacency visibility hypergraph by a matrix; the rows in the matrix are adjacent to nodes of the visibility hypergraph, and the columns in the matrix are adjacent to the hyperedges and regular edges of the visibility hypergraph.

Further, calculating a change vector for the adjacency visibility hypergraph based on the adjacency visibility hypergraphs at different time instants comprises:

calculating the change digit number of all columns in the matrix;

the change bit numbers of all columns are combined to obtain a change vector.

Further, the training method of the network interrupt classifier comprises the following steps:

collecting change vectors of a cell with a normal network state, and labeling each change vector with a label as a first training data set;

collecting change vectors of a network-interrupted cell, and labeling each change vector with a label as a second training data set;

the network interrupt classifier is trained based on the first training data set and the second training data set.

Further, the network outage classifier is a knowledge-based classifier or a multivariate statistical data-based classifier.

The other technical scheme of the invention is as follows: a cell interruption detection device based on adjacent visibility hypergraph change comprises a memory, a processor and a computer program which is stored in the memory and can run on the processor, wherein the processor executes the computer program to realize the cell interruption detection method based on adjacent visibility hypergraph change.

The invention has the beneficial effects that: the invention constructs the adjacent visibility hypergraph through the adjacent cell unit list of the active user, can combine the detection information of multiple users to improve the detection precision, and obtains signal detection data with stronger objectivity by calculating the change vector of the visibility hypergraph at different moments, thereby further improving the network signal detection precision.

Drawings

Fig. 1 is a schematic diagram illustrating an ANR interaction process in an LTE system according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of an adjacent visibility hypergraph in an embodiment of the invention;

fig. 3 is a schematic diagram illustrating an adjacency visibility hypergraph after a Cell 2 downtime according to an embodiment of the present invention;

fig. 4 is a flowchart of training a network interrupt classifier and detecting a downtime in an embodiment of the present invention.

Detailed Description

The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

The invention discloses a cell interruption detection method based on adjacent visibility hypergraph change, which comprises the following steps: receiving measurement information sent by active users in the coverage area of the current base station and constructing a neighbor cell unit list; constructing an adjacent visibility hypergraph according to the adjacent cell unit list; calculating a change vector of the adjacent visibility hypergraph based on the adjacent visibility hypergraph at different time instants; and taking the change vector as the input information of a network interrupt classifier, and determining the network state of the cell to be detected according to the classifier.

Aiming at the problem of service interruption caused by faults or parameter configuration problems of cell network facilities, the invention constructs the adjacent visibility hypergraph through the adjacent cell unit list of the active user, can combine the detection information of multiple users to improve the detection precision, obtains signal detection data with stronger objectivity by calculating the change vectors of the visibility hypergraph at different moments, and further improves the network signal detection precision.

In one embodiment, as shown in fig. 1, receiving measurement information from active users in the coverage area of the current base station and constructing the neighbor cell list includes: receiving a physical layer cell identifier which is sent by an active user and meets a configuration request; inquiring whether the physical layer cell identification exists in a neighbor cell list of the current base station; when the physical layer cell identification is not in the adjacent cell list, sending a global cell identification measurement configuration request to an active user; receiving a global cell identifier which is sent by an active user and meets a configuration request; and adding the global cell identification into the adjacent cell unit list.

Specifically, before receiving the physical layer cell identifier that satisfies the configuration request and is sent by the active user, the method includes:

in the embodiment of the invention, a source base station (such as an LTE system eNodeB) issues UE (user side) measurement configuration information to instruct the UE to measure peripheral cells according to configuration requirements. The measurement configuration requires that when the active user detects that the beam signal strength of the serving cell is lower than a first threshold value, the second signal strength of all the beam signals of other cells is detected; and when the second signal strength is larger than a second threshold value, taking the corresponding beam signal as a signal meeting the configuration request.

Generally, the UE performs the intra-frequency measurement by default, and the eNodeB may issue the intra-frequency handover measurement configuration information by default through signaling RRC Connection Reconfiguration when the UE has established the radio bearer. When the UE needs to perform the inter-frequency measurement, the eNodeB is required to issue the inter-frequency handover measurement configuration information. In general, a UE has only one receiver, and only one frequency point may receive signals at the same time, so that there is no way to receive information of two frequency points simultaneously. Therefore, the measurement GAP (inter-frequency measurement interval) is a time period from the UE leaving the current frequency point to the measurement of other frequency points, and the measurement GAP is used for inter-frequency measurement and inter-system measurement. In inter-frequency and inter-system measurements, the UE only performs measurements in the measurement GAP. And activating the GAP mode to perform pilot frequency measurement.

For example, the UE of the Cell a reports the PCI (physical layer Cell identity) of the Cell B (i.e., cell B) satisfying the measurement configuration requirement (i.e., measurement configuration request) to the source eNodeB Cell a (i.e., cell a) in a measurement report manner. The source eNodeB queries whether the PCI of Cell B exists in the intra-system NCL (i.e., neighbor Cell list) of Cell a. If yes, exiting the process; if not, the source eNodeB issues a measurement configuration request to the UE, and requests the UE to read a request for parameters of Cell B, such as ECGI (global Cell identity), TAC (Tracking Area), PLMN ID list (land mobile network identity list), and the like.

And the UE reports the read parameter information of the Cell B, such as ECGI, TAC, PLMN ID list and the like, to the source eNodeB. The source eNodeB adds the newly found neighbor information of Cell B to the intra-system NCL of the source eNodeB, and simultaneously adds the neighbor relation to the intra-system NRT of Cell a.

The UE of cell B constructs the NCL of cell B in the same manner.

In one embodiment, before constructing the adjacency visibility hypergraph from the neighbor cell unit list comprises: and receiving a neighbor cell unit list sent by a neighbor cell in the neighbor cell unit list. In order to obtain more useful information, in a certain area, base stations exchange information with each other through a standard interface such as X2, so as to provide better service.

As a specific implementation manner, constructing an adjacency visibility hypergraph according to a neighbor cell unit list includes: establishing an adjacent visibility hypergraph by taking a beam ID in the adjacent cell unit list as a node of the adjacent visibility hypergraph, taking all visible beams in a cell as a hypergraph of the adjacent visibility hypergraph and taking a visible beam of each active user as a conventional edge; expressing the adjacency visibility hypergraph by a matrix; the rows in the matrix are adjacent to nodes of the visibility hypergraph, and the columns in the matrix are adjacent to the hyperedges and regular edges of the visibility hypergraph.

Neighbor cell unit list (NCL) reporting can create a visibility relationship graph (i.e., an adjacency visibility hypergraph) that is generated from NCL reports received by a mobile terminal (UE), the adjacency hypergraph elements including hyperedges, edge weights, and vertices. The edge weight represents the number of the mobile terminals which have reported the specific neighbor relation, and the vertex represents the beam ID.

The NCL report is always generated when the UE is in an active connection state, at which time the UE continuously measures the beam signal strength and quality of the radio channel of the serving cell to which it is currently connected, and also measures the beam signals of some neighboring cells, which constitute potential handover candidates. These measurements are sent to the base station of the current serving cell to decide whether a handover has to be performed. A terminal making a connection can measure up to M base stations in its visibility range and send NCL reports at intervals of T, containing the N best measurements.

For example, in the GSM system, a connecting terminal can measure up to 16 base stations in its field of view and send NCL reports at 480ms intervals, including the 6 best measurements. NCL reports of different terminals are retrieved by a subordinate network entity. The detailed information of the measurement report is different from GSM to UMTS or LTE, but it is assumed that the neighbor cell list is reported at regular time intervals, which can be achieved in different radio access technologies.

Based on the measurement reports of different terminals, a visibility hypergraph is constructed, and the observation of the change of elements in the visibility hypergraph is a key element for the cell outage problem detection provided by the invention. In particular, the contiguous visibility hypergraph is created at fixed time intervals, and any abnormal change in the contiguous visibility hypergraph may contain a hint of cell outage. As shown in FIG. 2, a specific example of such a hypergraph is depicted, two consecutive visibility hypergraphs G (t) ₁ ) And visibility hypergraph G (t) ₁ + T) is established by the time interval T. For convenience of description, hereinafter, referred to as the figure, the detection method proposed in this embodiment is on G (t) ₁ + T) any node that becomes isolated is sensitive, in which case a pattern of change of the graph is created.

In fig. 2, user A1 is served by beam V11 of Cell 1 at the current time and can receive the beam signal of beam V22 of Cell 2, and user A2 is served by beam V13 of Cell 1 at the current time and can receive the beam signal of beam V21 of Cell 2. While user B1 is currently served by beam V22 of Cell 2 and can receive the beam signal of beam V14 of Cell 1. User B2 is currently served by beam V23 of Cell 2 and can receive the beam signal of beam V12 of Cell 1.

In this example figure, the coverage visibility range of two user terminals of cell a is a circle around their respective positions, the NCL reports of its users A1 and A2 contain beams V11, V13 of the serving cell 1 as the current serving beam and beams V21, V22 of the neighboring cell 2, which are the beam IDs of the neighboring cells within its visibility range, and the NCL reports of the users B1 and B2 of the serving cell B are constructed accordingly.

In this embodiment, cell 1 and cell 2 exchange information with each other, and a visibility hypergraph is constructed according to NCL reports of four users (i.e., A1, A2, B1, and B2).

The visibility hypergraph is composed of nodes, hyperedges, and regular edges. The nodes comprise beam components under the cell, the super edge comprises a beam of a service user in the cell and a visible beam reported by users in other cells, except for the super edge, the edge between the linked nodes is a conventional edge, and two nodes are generally linked and comprise a service beam of one user and a visible beam of another cell.

In the embodiment of the invention, for convenience of description and calculation, the adjacency visibility hypergraph is expressed by a matrix; the rows in the matrix are adjacent to nodes of the visibility hypergraph, and the columns in the matrix are adjacent to the hyperedges and regular edges of the visibility hypergraph.

For example, as shown in Table 1 below, the super edge He1 (connecting 4 nodes: v11, v12, v13, v 14) includes two parts: 1) Service beam ID (v 11, v 13) formation in NCL reports for users A1 and A2; 2) The visible beam ID in the NCL report for users B1 and B2 constitutes (v 12, v 14). Similarly, the supercide He2 (connecting 3 nodes: v21, v22, v 23) also includes two parts: 1) The visible beam ID (v 21, v 22) in the NCL report for users A1 and A2; 2) The service beam ID (v 22, v 23) in the NCL report for users B1 and B2.

The super edge graph can be represented by a matrix, the super edge comprises nodes, the intersection point value of the column where the super edge is located and the row containing the nodes is 1, and otherwise, the node is 0. And connecting the two nodes of the edge, wherein the positions of the row and the column of the corresponding node are 1, and otherwise, the positions are 0.

TABLE 1

	He1	He2	E1	E2	E3	E4
							V11
	1	0	1	0	0	0
							V12	1	0	0	1	0	0
V13	1	0	0	0	1	0
							V14	1	0	0	0	0	1
V21	0	1	0	0	1	0
							V22	0	1	1	0	0	1
V23	0	1	0	1	0	0
							V24	0	0	0	0	0	0

During normal operation of the network, visibility hypergraphs may change frequently, caused by starting and ending calls, user mobility, changes in radio propagation (e.g., millimeter wave blocking), and changes in the NCL report itself.

The visibility hypergraph will change when the network goes down, e.g. the hypergraph caused by Cell 2 going down is shown in fig. 3. Compared to the original state of each cell (i.e., the non-down situation), the NCL report for user B in fig. 3 (i.e., containing B1 and B2) is no longer received, resulting in the disappearance of the visible beam for user B (i.e., user B cannot complete the measurement report), e.g., (v 12, v 14). Meanwhile, the service beam of the Cell 2 user disappears, such as (v 22, v 23), the visible beam of the user in Cell a (v 21, v 22). Each downtime condition results in the disappearance of one of the hyperedges in the visibility hypergraph.

When the Cell 2 is down, the corresponding matrix will change due to the change of the adjacent visibility hypergraph, the changed matrix is shown in the following table 2, and the obtained change matrix is shown in the following table 3 after the matrix in the table 1 and the matrix in the table 2 are calculated.

TABLE 2

	He1	He2	E1	E2	E3	E4
							V11
	1	0	0	0	0	0
							V12	0	0	0	0	0	0
V13	1	0	0	0	0	0
							V14	0	0	0	0	0	0
V21	0	0	0	0	0	0
							V22	0	0	0	0	0	0
V23	0	0	0	0	0	0
							V24	0	0	0	0	0	0

TABLE 3

	He1	He2	E1	E2	E3	E4
							V11	0	0	-1	0	0	0
V12	-1	0	0	-1	0	0
							V13	0	0	0	0	-1	0
V14	-1	0	0	0	0	-1
							V21	0	-1	0	0	-1	0
V22	0	-1	-1	0	0	0
							V23	0	-1	0	-1	0	0
V24	0	0	0	0	0	0

Further, calculating a change vector for the adjacent visibility hypergraph from the adjacent visibility hypergraph at different times includes: calculating the change digit number of all columns in the matrix; the change bit numbers of all columns are combined to obtain a change vector.

From this, it is understood that the degree of change of the super edge 1 is 2, the degree of change of the super edge 2 is 3, the degree of change of the edge 1 is 2, the degree of change of the edge 2 is 2, the degree of change of the edge 3 is 2, and the degree of change of the edge 4 is 2. The corresponding hypergraph mode change vector is then: x = [2,3,2,2,2,2], and the network state prediction of the cell can be obtained by inputting the change vector into the classifier.

In the embodiment of the invention, the training method of the network interrupt classifier comprises the following steps: collecting change vectors of a cell with a normal network state, and labeling each change vector with a label as a first training data set; collecting change vectors of a network-interrupted cell, and labeling each change vector with a label as a second training data set; the network interrupt classifier is trained based on a first training data set and a second training data set.

As can be seen from the above, a classification model is trained and learned by using the vector X formed by the data, and the cell downtime can be judged by outputting the result. Due to the fact that the cell downtime can generate a characteristic change mode, the change mode can be distinguished from the normal fluctuation of the visible graph, and therefore the downtime detection problem can be converted into the classification problem of the change mode of the visible graph.

A classification of a group of items is one that assigns similar items to one of several different categories. The task of separating the visible graph change pattern into interrupt and non-interrupt cases can be viewed as a binary classification problem with a set of predefined classes. Classification algorithms are widely used for automatic pattern recognition, and in the embodiment of the present invention, the network interrupt classifier is a knowledge-based classifier (e.g., expert system, neural network, genetic algorithm, etc.) or a multivariate statistical data-based classifier (e.g., cluster analysis, classification, regression tree, etc.).

In summary, any of the above-described types of classification techniques may be employed in the classifier-based downtime detection algorithm. The basic steps of which refer to the training and downtime detection flow chart of fig. 4.

Collecting neighbor cell measurement reports of cell users in a normal cell, and constructing a visibility hypergraph and a visible graph change mode according to the neighbor cell measurement reports; in the abnormal downtime cell, collecting the neighbor cell measurement report of the cell user, and constructing a visibility hypergraph and a visible graph change mode according to the neighbor cell measurement report;

change vector data set for normal cell hypergraph mode { X } _m M =1,2, …, M } is labeled normal { Y } _m =1,m =1,2, …, M }, and the vector data set { X } is changed for the hypergraph mode of the abnormal downtime cell _n N =1,2, …, N } is labeled as exception { Y } _n ＝0，n＝1，2，…，N}。

Using { X _m ,Y _m And { X } _n ,Y _n Training a classifier; and finally, carrying out downtime detection by using the trained classifier: namely, collecting a neighbor cell measurement report of a cell user to be detected, constructing a visibility hypergraph and a visible image change mode vector X according to the neighbor cell measurement report, inputting the X into a classifier, and outputting a result, namely a detection result.

The invention also discloses a cell interruption detection device based on the change of the adjacent visibility hypergraph, which comprises a memory, a processor and a computer program which is stored in the memory and can run on the processor, wherein the processor realizes the cell interruption detection method based on the change of the adjacent visibility hypergraph when executing the computer program.

The device can be a desktop computer, a notebook computer, a palm computer, a cloud server and other computing equipment. The apparatus may include, but is not limited to, a processor, a memory. Those skilled in the art will appreciate that the apparatus may include more or fewer components, or some components in combination, or different components, and may also include, for example, input-output devices, network access devices, etc.

The Processor may be a Central Processing Unit (CPU), and the Processor may be other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

The storage may in some embodiments be an internal storage unit of the device, such as a hard disk or a memory of the device. The memory may also be an external storage device of the apparatus in other embodiments, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), etc. provided on the apparatus. Further, the memory may also include both an internal storage unit and an external storage device of the apparatus. The memory is used for storing an operating system, application programs, a BootLoader (BootLoader), data, and other programs, such as program codes of the computer programs. The memory may also be used to temporarily store data that has been output or is to be output.

It should be noted that, for the specific content of the above-mentioned apparatus, since the same concept is based on, the specific functions and the technical effects brought by the method embodiment of the present invention, reference may be made to the method embodiment section specifically, and details are not described here.

Claims

1. A cell outage detection method based on adjacent visibility hypergraph change is characterized by comprising the following steps:

calculating a change vector of the contiguous visibility hypergraph based on the contiguous visibility hypergraph at different time instants;

and taking the change vector as the input information of a network interrupt classifier, and determining the network state of the cell to be detected according to the classifier.

2. The method as claimed in claim 1, wherein the step of receiving measurement information from active users in the coverage area of the current base station and constructing the neighbor cell list comprises:

receiving a physical layer cell identification which is sent by an active user and meets a configuration request;

when the physical layer cell identifier is not in the neighbor cell list, sending a global cell identifier measurement configuration request to an active user;

and adding the global cell identification into a neighbor cell unit list.

3. The method for detecting cell outage based on adjacency visibility hypergraph change according to claim 2, wherein receiving the physical layer cell id satisfying the configuration request sent by the active user comprises:

when the active user detects that the beam signal intensity of the serving cell is lower than a first threshold value, detecting second signal intensity of all received beam signals of other cells;

4. The cell outage detection method based on the change of the adjacency visibility hypergraph according to claim 2 or 3, wherein before constructing the adjacency visibility hypergraph according to the neighbor cell unit list, the method comprises:

5. The cell outage detection method based on adjacency visibility hypergraph change according to claim 4, wherein constructing an adjacency visibility hypergraph according to the neighbor cell unit list comprises:

establishing an adjacent visibility hypergraph by taking the beam ID in the adjacent cell unit list as a node of the adjacent visibility hypergraph, taking all visible beams in a cell as a hypergraph of the adjacent visibility hypergraph and taking the visible beam of each active user as a conventional edge;

expressing the adjacency visibility hypergraph by a matrix; the rows in the matrix are nodes of the adjacent visibility hypergraph, and the columns in the matrix are the superedges and regular edges of the adjacent visibility hypergraph.

6. The method of claim 5, wherein calculating the change vector of the adjacency visibility hypergraph based on the adjacency visibility hypergraph at different times comprises:

calculating the number of change bits of all columns in the matrix;

and combining the change bit numbers of all the columns to obtain the change vector.

7. The cell outage detection method based on the change of the adjacency visibility hypergraph according to claim 5 or 6, characterized in that the training method of the network outage classifier is as follows:

8. The cell outage detection method based on adjacency visibility hypergraph change according to claim 7, characterized in that the network outage classifier is a knowledge-based classifier or a multivariate statistical data-based classifier.

9. An apparatus for cell outage detection based on adjacency visibility hypergraph change, comprising a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements a method for cell outage detection based on adjacency visibility hypergraph change according to any of claims 1-8 when executing the computer program.