WO2021043146A1 - 检测方法、装置及系统 - Google Patents

检测方法、装置及系统 Download PDF

Info

Publication number
WO2021043146A1
WO2021043146A1 PCT/CN2020/112902 CN2020112902W WO2021043146A1 WO 2021043146 A1 WO2021043146 A1 WO 2021043146A1 CN 2020112902 W CN2020112902 W CN 2020112902W WO 2021043146 A1 WO2021043146 A1 WO 2021043146A1
Authority
WO
WIPO (PCT)
Prior art keywords
terminal
connection information
threshold
feature
target
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.)
Ceased
Application number
PCT/CN2020/112902
Other languages
English (en)
French (fr)
Inventor
李长路
包德伟
魏启坤
孙福清
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to JP2022513994A priority Critical patent/JP7293502B2/ja
Priority to KR1020227008650A priority patent/KR20220048014A/ko
Priority to EP20861088.1A priority patent/EP4021060B1/en
Publication of WO2021043146A1 publication Critical patent/WO2021043146A1/zh
Priority to US17/683,995 priority patent/US12279138B2/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W24/00Supervisory, monitoring or testing arrangements
    • H04W24/08Testing, supervising or monitoring using real traffic
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B17/00Monitoring; Testing
    • H04B17/20Monitoring; Testing of receivers
    • H04B17/27Monitoring; Testing of receivers for locating or positioning the transmitter
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B17/00Monitoring; Testing
    • H04B17/30Monitoring; Testing of propagation channels
    • H04B17/309Measuring or estimating channel quality parameters
    • H04B17/318Received signal strength
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W48/00Access restriction; Network selection; Access point selection
    • H04W48/16Discovering, processing access restriction or access information
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04BTRANSMISSION
    • H04B17/00Monitoring; Testing
    • H04B17/30Monitoring; Testing of propagation channels
    • H04B17/391Modelling the propagation channel
    • H04B17/3912Simulation models, e.g. distribution of spectral power density or received signal strength indicator [RSSI] for a given geographic region
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W24/00Supervisory, monitoring or testing arrangements
    • H04W24/02Arrangements for optimising operational condition
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W84/00Network topologies
    • H04W84/02Hierarchically pre-organised networks, e.g. paging networks, cellular networks, WLAN [Wireless Local Area Network] or WLL [Wireless Local Loop]
    • H04W84/10Small scale networks; Flat hierarchical networks
    • H04W84/12WLAN [Wireless Local Area Networks]
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W88/00Devices specially adapted for wireless communication networks, e.g. terminals, base stations or access point devices
    • H04W88/12Access point controller devices

Definitions

  • This application relates to the field of communications, and in particular to a detection method, device and system.
  • a wireless local area network usually includes at least one access point (access point, AP), and a user can access an AP in the WLAN to achieve access to the WLAN.
  • APs can provide services for users.
  • an access point controller AC can manage APs. For example, AC can optimize APs. , So that the AP can provide better services.
  • the embodiments of the present application provide a detection method, device, and system that can detect logical edge APs.
  • the technical solutions are as follows:
  • the present application provides a detection method.
  • at least one feature of a first access point AP is acquired, and the at least one feature includes requesting access to the first access point within a time period of the first time period.
  • the number of target terminals of an AP, the target terminal is a terminal with abnormal access status; according to at least one feature, it is detected whether the first AP is a logical edge AP, and the logical edge AP is the signal coverage reaching the edge of the signal coverage of the wireless local area network WLAN.
  • AP a detection method.
  • the acquired characteristics of the first AP include the number of target terminals requesting access to the first AP during the first period of time, and the target terminals are terminals with abnormal access status, it can be successful based on the characteristics of the first AP Detect whether the first AP is a logical edge AP.
  • the target terminal includes at least one of a nomadic terminal, a terminal that has not successfully accessed, or a terminal that does not belong to the whitelist of the WLAN to which the first AP belongs.
  • These types of terminals usually appear in the signal coverage of the logical edge AP, so according to the number of target terminals, it is possible to successfully detect whether the first AP is a logical edge AP.
  • the first AP when at least one feature satisfies the first condition, it is determined that the first AP is a logical edge AP.
  • the first condition includes that the number of target terminals is greater than the first number threshold, or the first AP is One of the first APs with the largest number of target terminals in the WLAN, or the difference between the number of target terminals and the average number of target terminals is greater than the difference threshold, which is equal to m times the variance value, m If greater than 1, the average number of target terminals and the variance value are obtained based on the number of target terminals of each AP in the WLAN. Therefore, whether the first AP is a logical edge AP can be successfully detected through the first condition.
  • the target terminal includes at least one of a nomadic terminal or a terminal that does not belong to the whitelist, and the connection information of at least one terminal collected by the first AP is received, and the connection information includes the terminal's connection information.
  • Identification at least one terminal is a terminal that accesses the first AP; according to the connection information of the at least one terminal, a target terminal is determined from the at least one terminal; and the number of determined target terminals is counted. Since the first AP collects the connection information of the terminal, the target terminal can be successfully determined based on the connection information, so that it can successfully detect whether the first AP is a logical edge AP.
  • the time interval for the first AP to collect connection information from the same terminal twice does not exceed the interval threshold, and the connection information also includes the collection time; for any of the at least one terminal
  • the terminal obtains the connection information sequence of any terminal, the connection information sequence includes connection information arranged by collection time, each connection information in the connection information sequence includes the identification of any terminal, and the connection information sequence in the connection information sequence
  • the interval between the collection times of two adjacent connection information does not exceed the interval threshold; it is determined whether any terminal is a nomadic terminal according to the connection information sequence. In this way, since the connection information sequence of a terminal can be obtained from the connection information collected by the AP, the nomadic terminal can be successfully determined based on the connection information sequence.
  • the start time of any terminal accessing the first AP, the stay time of any terminal in the first AP, and the disconnection of any terminal are obtained according to the connection information sequence.
  • the disconnection time of the connection with the first AP; the stay time does not exceed the first time threshold, and there is no second AP acquisition within the preset time before the start time and the preset time after the disconnect time
  • the connection information of the any terminal it is determined that the any terminal is a nomadic terminal, the second AP is an AP other than the first AP in the WLAN, and the preset duration is greater than the interval threshold. Since the stay time of any terminal, the start time of access, and the disconnection time of disconnection are obtained, based on the stay time, start time and disconnection time, it can be successfully determined whether any terminal is nomadic terminal.
  • the at least one feature further includes the total number of each connection event in the at least one connection event of the at least one terminal, and for any connection event in each connection event,
  • the first condition further includes that the total number of any type of connection event is greater than the number threshold corresponding to any type of connection event, or the first condition also includes that the total number of any type of connection event is less than that of any type of connection event The number threshold. Since the at least one feature also includes the total number of each connection event, the accuracy of detecting the logical edge AP can be improved in this way.
  • connection information of any terminal further includes at least one of the uplink signal strength of any terminal or the data transmission delay of any terminal; according to the connection information of any terminal At least one of uplink signal strength or data transmission delay included in the last piece of connection information in the sequence is used to obtain connection events of any terminal; and count the total number of connection events of any type. In this way, the connection events of the terminal can be obtained, and then the total number of connection events can be counted. Based on the total number of connection events, the accuracy of detecting logical edge APs can be improved.
  • connection information further includes at least one of the terminal's signal-to-noise ratio, packet loss rate, retransmission rate, channel utilization, and frequency band identification that the terminal accesses; at least one feature also includes information At least one of the statistical value of the noise ratio, the statistical value of the packet loss rate, the statistical value of the retransmission rate, the statistical value of the channel utilization, or the total number of terminals accessing the frequency band corresponding to the frequency band identifier.
  • the statistical value of the signal-to-noise ratio is based on at least one terminal
  • the statistical value of the packet loss rate is based on the packet loss rate of at least one terminal
  • the statistical value of the retransmission rate is based on the retransmission rate of at least one terminal
  • the statistical value of the channel utilization rate is based on at least one
  • the channel utilization rate of the terminal is obtained;
  • the first condition also includes that the statistical value of the signal-to-noise ratio is less than the statistical threshold of the signal-to-noise ratio, the statistical value of the packet loss rate is greater than the statistical threshold of the packet loss rate, the statistical value of the retransmission rate is greater than the statistical threshold of the retransmission rate, and the channel At least one of the utilization statistical value is greater than the channel utilization threshold or the total terminal number is greater than the terminal number threshold.
  • a training set is used to train a random forest model.
  • the training set includes multiple training samples and a category corresponding to each training sample.
  • the training samples whose category is a positive sample include at least one feature of the logical edge AP,
  • the training sample of the negative sample category includes at least one feature of the non-logical edge AP,
  • the random forest model after training includes at least one decision tree, and each path in the decision tree is used to detect whether any AP in the training set is It is a logical edge AP.
  • the leaf node of the path is used to save the detection result of the path.
  • the nodes in the path other than the leaf node correspond to a category and feature threshold. This node is used to determine whether the first feature exceeds the feature corresponding to the node.
  • the first feature is the feature of any AP belonging to the corresponding category of the node; determine the category corresponding to each feature of the first AP that needs to be acquired according to at least one decision tree , The characteristic threshold and the judgment condition between the characteristic and the characteristic threshold.
  • the category corresponding to each feature of the first AP that needs to be acquired can be determined, so that the feature that reflects the characteristics of the logical edge AP can be determined, and the logical edge AP can be successfully detected based on this feature. It can also reduce the number of acquired features and reduce the amount of data for calculations.
  • the access controller AC when detecting that the first AP is a logical edge AP, instruct the access controller AC to reduce the signal coverage of the first AP or instruct the AC to control the first AP to prevent downlink signal strength
  • the terminal less than the downlink signal strength threshold accesses the first AP or instructs the AC to start the delayed access function of the first AP. In this way, the performance of the logical edge AP can be optimized, and the impact caused by the logical edge AP can be reduced.
  • this application provides a detection method, in which: the connection information of at least one terminal collected by the first access point AP is received, and the connection information includes the identification of the terminal and the collection time for collecting the connection information , At least one terminal is a terminal that accesses the first AP; for any terminal of the at least one terminal, a connection information sequence of any terminal is acquired, and the connection information sequence includes connection information arranged by collection time, and the connection information Each connection information in the sequence includes the identification of any terminal, and the interval between the collection times of two adjacent connection information in the connection information sequence does not exceed the interval threshold; the any terminal is determined according to the connection information sequence Whether it is a nomadic terminal. Since the connection information sequence of a terminal can be obtained from the connection information collected by the AP, the nomadic terminal can be successfully determined based on the connection information sequence.
  • the start time of any terminal accessing the first AP, the stay time of any terminal in the first AP, and the disconnection of any terminal from the first AP are obtained according to the connection information sequence.
  • the disconnection time between the connections; the stay time does not exceed the first time threshold, and there is no second AP collected by any terminal within the preset time period before the start time and the preset time period after the disconnection time
  • the connection information determines that any terminal is a nomadic terminal, and the second AP is an AP other than the first AP in the WLAN to which the first AP belongs. Since the stay time of any terminal, the start time of access, and the disconnection time of disconnection are obtained, based on the stay time, start time and disconnection time, it can be successfully determined whether any terminal is nomadic terminal.
  • the present application provides a detection device for executing the first aspect or the method in any one of the possible implementation manners of the first aspect.
  • the device includes a unit for executing the method of the first aspect or any one of the possible implementation manners of the first aspect.
  • this application provides a detection device for executing the first aspect or the method in any one of the possible implementation manners of the first aspect.
  • the device includes a unit for executing the method of the first aspect or any one of the possible implementation manners of the first aspect.
  • an embodiment of the present application provides a detection device, which includes a processor and a memory.
  • the processor and the memory may be connected through a bus system.
  • the memory is used to store one or more programs, and the processor is used to execute one or more programs in the memory to complete the first aspect, the second aspect, any possible implementation manner of the first aspect, or the second aspect Any of the possible implementations of the method.
  • the present application provides a computer-readable storage medium.
  • the computer-readable storage medium stores instructions that, when run on a processor, cause the processor to execute the first, second, and first aspects described above. Any possible implementation of the aspect or a method in any possible implementation of the second aspect.
  • the present application provides a computer program product containing instructions, which when run on a processor, causes the processor to execute the first aspect, the second aspect, any possible implementation manner of the first aspect, or the second aspect.
  • the method in any possible implementation of the aspect.
  • the present application provides a detection system.
  • the system includes: a data analyzer CI and a first access point AP.
  • the first AP collects connection information of at least one terminal and sends the connection information to the CI.
  • the connection information includes the identification of the terminal and the collection time for collecting the connection information.
  • the at least one terminal is a terminal that accesses the first AP; the CI obtains at least one feature of the first AP according to the connection information, and the at least one feature is included when the time length is The number of target terminals requesting access to the first AP in the first time period, where the target terminals include at least one of nomadic terminals or terminals that are not in the whitelist of the wireless local area network WLAN where the first AP is located; detecting according to at least one characteristic Whether the first AP is a logical edge AP, the logical edge AP is an AP whose signal coverage reaches the edge of the signal coverage of the WLAN.
  • the acquired characteristics of the first AP include the number of target terminals requesting access to the first AP during the first time period, and the target terminals include nomadic terminals or terminals that are not in the whitelist of the WLAN where the first AP is located At least one of them, so that whether the first AP is a logical edge AP can be successfully detected based on the characteristics of the first AP.
  • the system further includes: an access controller AC.
  • the CI detects that the first AP is a logical edge AP, it sends an optimization request to the AC, and the optimization request includes the identity of the first AP; the AC Reduce the signal coverage of the first AP or control the first AP to prevent terminals whose downlink signal strength is less than the downlink signal strength threshold from accessing the first AP or start the delayed access function of the first AP. In this way, the performance of the logical edge AP can be optimized, and the impact caused by the logical edge AP can be reduced.
  • the present application provides a detection system.
  • the system includes a data analyzer CI and a first access point AP.
  • the first AP collects connection information of at least one terminal and sends the connection information to the CI.
  • the connection information includes the identification of the terminal and the collection time for collecting the connection information.
  • At least one terminal is a terminal connected to the first AP; for any one of the at least one terminals, the CI obtains the connection information sequence of any terminal, and the connection
  • the information sequence includes connection information arranged by collection time, each connection information in the connection information sequence includes the identification of any terminal, and the interval between the collection times of two adjacent connection information in the connection information sequence does not exceed the interval
  • Threshold Determine whether any terminal is a nomadic terminal according to the connection information sequence. In this way, since the connection information sequence of a terminal can be obtained from the connection information collected by the first AP, the nomadic terminal can be successfully determined based on the connection information sequence.
  • FIG. 1 is a schematic structural diagram of a WLAN provided by an embodiment of the present application.
  • FIG. 2 is a schematic diagram of a system architecture provided by an embodiment of the present application.
  • FIG. 3 is a schematic diagram of another system architecture provided by an embodiment of the present application.
  • FIG. 4 is a flowchart of a detection method provided in an embodiment of the present application.
  • FIG. 5 is a schematic structural diagram of a decision tree provided by an embodiment of the present application.
  • FIG. 6 is a schematic structural diagram of a detection device provided by an embodiment of the present application.
  • FIG. 7 is a schematic structural diagram of another detection device provided by an embodiment of the present application.
  • FIG. 8 is a schematic structural diagram of another detection device provided by an embodiment of the present application.
  • WLANs can be deployed in places such as buildings, so that terminals located in the places can access the WLAN.
  • the WLAN includes at least one AP, and a terminal located within the signal coverage area of the WLAN can access any one of the at least one AP, so as to achieve access to the WLAN.
  • a type of AP in the WLAN which may be a logical edge AP
  • a logical edge AP is the signal coverage area that reaches the edge of the WLAN signal coverage area.
  • AP When the AP is a logical edge AP, there may be a large number of target terminals requesting to access the AP during the time period when the duration is the first duration. For example, the number of target terminals requesting to access the AP within the first time period exceeds the first number threshold.
  • Target terminals include terminals with abnormal access status. For example, target terminals may include terminals that have not successfully accessed the AP, or have stayed in the WLAN for a time that does not exceed the first time threshold, or are not included in the whitelist of the WLAN. At least one.
  • the signal coverage area of the AP usually reaches the edge of the signal coverage area of the WLAN, and the signal coverage area of the AP may also exceed the range of the place, and may cover the pedestrian flow channel located outside the place.
  • the AP may be an edge AP located at the edge of the WLAN, or may not be an edge AP located at the edge of the WLAN. Referring to the WLAN deployed in the site 1 shown in FIG. 1, in the WLAN AP1, AP3, AP4, and AP5 are edge APs located at the edge of the WLAN, and AP2 is not the edge AP located at the edge of the WLAN.
  • the signal coverage area of AP2 reaches the edge of the coverage area of the WLAN signal, and exceeds the location 1 and covers the road 2 located outside the location 1.
  • AP4 is an edge AP located at the edge of the WLAN, and the signal coverage of AP4 also extends beyond the location 1, and covers another road 3 located outside the location 1. Therefore, AP2 and AP4 are APs belonging to this category.
  • the user's terminal Since there may be a passage through which there may be people in the signal coverage area of the AP, users walking on the passage will pass through the coverage area of the AP.
  • the user's terminal When a user passes through the coverage area of the AP, the user's terminal will request to access the AP to access the WLAN. If the user’s terminal successfully accesses the AP, the user may soon leave the AP’s signal coverage after accessing the AP, or even leave the WLAN signal coverage, the user’s terminal will disconnect from the AP. Connection, the user’s terminal stays in the WLAN for a short time, and the stay time may not exceed the first time threshold.
  • the terminal is also called a nomadic terminal.
  • Each AP in the WLAN may store a WLAN whitelist.
  • the WLAN whitelist may include the identities of the terminals of the employees of the company.
  • a large part of the people walking on the channel are not employees of the company, so that the terminals requesting to access the AP may include many terminals that do not belong to the whitelist.
  • the AP may deny the terminal access, resulting in the terminal not successfully accessing the AP.
  • the channel is located at the edge of the signal coverage of the WLAN, the signal strength of the WLAN is weak, and because the user may quickly pass through the signal coverage of the AP, the user's terminal may not successfully access the AP .
  • the signal coverage area of the AP reaches the edge of the signal coverage area of the WLAN, and the terminals requesting to access the AP may include a large number of nomadic terminals that have stayed in the WLAN for less than the first time threshold, and are unsuccessful.
  • the logical edge AP has a greater impact on the nomadic terminal.
  • the nomadic terminal Before the nomadic terminal enters the signal coverage of the logical edge AP, it may use a high-quality mobile network to provide Internet services for the running applications of the nomadic terminal. When it enters the signal coverage of the logical edge AP, it will automatically access the logical edge AP. , Access to WLAN. Since the nomadic terminal is located at the edge of the signal coverage of the logical edge AP, the signal of the logical edge AP at the location of the nomadic terminal is weak, and the Internet experience may be reduced after accessing the logical edge AP. When the nomadic terminal leaves the signal coverage area of the logical edge AP, the nomadic terminal will not reuse the mobile network to provide Internet services for the running applications of the nomadic terminal, which interrupts the Internet services provided for the running applications.
  • the logical edge AP has a greater impact on the resident terminals that access the logical edge AP or the terminals that are in the whitelist of the WLAN.
  • the resident terminal means that after accessing the logical edge AP, it is at the logical edge.
  • the terminal whose staying time in the AP exceeds the first time threshold. Since there may be a large number of nomadic terminals or terminals not belonging to the whitelist accessing the logical edge AP, a large amount of network resources of the logical edge AP may be consumed, thereby affecting the resident terminals located in the logical edge AP or the connected terminals belonging to the logical edge AP. Network resources used by whitelisted terminals.
  • logical edge APs will also have a greater impact on WLAN.
  • a large number of nomadic terminals request access to the logical edge AP, a large number of control plane messages will be generated, or a large number of nomadic terminals will leave the logical edge AP.
  • the large number of nomadic terminals disconnected from the logical edge APs will also generate a large number of control plane messages, which will increase the network burden on the control plane of the WLAN and may cause network failures.
  • the logical edge AP can be detected from the WLAN AP, and then the logical edge AP can be optimized to reduce or eliminate the above influence caused by the logical edge AP.
  • an embodiment of the present application provides a network architecture, including:
  • a data analyzer (campus insight, CI) and at least one AP in the WLAN may establish a network connection between the CI and the AP in each WLAN.
  • the CI may acquire at least one characteristic of the first AP; and detect whether the first AP is a logical edge AP according to the at least one characteristic.
  • the first AP is a logical edge AP.
  • the at least one feature includes the number of target terminals, and the number of target terminals may be used to reflect the access status of the terminals accessing the first AP.
  • the first condition includes that the number of target terminals is greater than a first number threshold.
  • the at least one feature may further include the total number of each connection event in the at least one connection event, and the total number of each connection event is used to reflect the performance of the first AP.
  • the at least one connection event is a connection event of at least one terminal accessing the first AP.
  • the first condition may further include that the total number of any connection event is greater than a threshold number corresponding to any connection event, or the first condition may also include The total number of any type of connection event is less than the number threshold corresponding to any type of connection event.
  • the at least one feature may further include at least one of a statistical value of signal-to-noise ratio, a statistical value of packet loss rate, a statistical value of retransmission rate, or a statistical value of channel utilization. Any one of the statistical value of the signal-to-noise ratio, the statistical value of the packet loss rate, the statistical value of the retransmission rate, or the statistical value of the channel utilization rate can be used to reflect the access situation of the terminal accessing the first AP.
  • the first condition may also include that the statistical value of the signal-to-noise ratio is less than the statistical threshold of the signal-to-noise ratio, the statistical value of the packet loss rate is greater than the statistical threshold of the packet loss rate, the statistical value of the retransmission rate is greater than the statistical threshold of the retransmission rate or the channel utilization statistics. The value is greater than at least one of the channel utilization thresholds.
  • the first AP may provide at least one frequency band for the terminal to access.
  • the at least one feature may further include the total number of terminals that access the frequency band.
  • the total number of terminals is used to reflect the load situation of the first AP.
  • the first condition may also include a condition that the total number of terminals is greater than a threshold value of the number of terminals.
  • the CI can be a computer, a server, or a server cluster, etc.
  • the network architecture may further include an AC.
  • a network connection may be established between the AC and the CI, and a network connection may be established between the AC and the AP in the at least one WLAN.
  • the AC is used to optimize the performance of any AP in the at least one WLAN.
  • the CI may send an optimization request to the AC, where the optimization request includes the identification of the first AP.
  • the AC receives the optimization request, and optimizes the performance of the first AP according to the identifier of the first AP included in the optimization request.
  • an embodiment of the present application provides a detection method, which is used to detect whether an AP in a WLAN is a logical edge AP.
  • the method can be applied to the network architecture shown in FIG. 2 or FIG. 3, including:
  • Step 201 CI receives connection information of at least one terminal collected by a first AP.
  • the connection information of any terminal includes the identification of any terminal and the connection information collected by the first AP The acquisition time.
  • CI may be used to detect APs in at least one WLAN, that is, CI is used to detect APs in one WLAN or detect APs in multiple WLANs, and the first AP is any AP in the at least one WLAN.
  • the at least one terminal is a terminal connected to the first AP.
  • the AP in the at least one WLAN detected by the CI may be configured by an administrator. For example, for each WLAN in the at least one WLAN, if the administrator of the WLAN needs CI to detect APs in the WLAN, the administrator can input the identity of each AP in the WLAN and the identity of the WLAN into the CI.
  • the CI can save the identity of each AP in the WLAN and the identity of the WLAN into the corresponding relationship between the identity of the AP and the identity of the WLAN, and for each AP in the WLAN, the CI can respectively store the identity of the AP according to the identity of the AP.
  • the first AP can collect the connection information of any terminal at different times, and send the collected connection information to the CI, and the first AP performs neighboring connection information of the any terminal The time interval between two acquisitions does not exceed the interval threshold.
  • the first AP may periodically collect the connection information of any terminal, and the length of the period during which the first terminal collects the connection information is less than or equal to the interval threshold.
  • connection information may also include the identification of the first AP, the collection time at which the first AP collects the connection information, the time at which any terminal accesses the first AP, and the time when any terminal has been connected to the first AP. At least the stay time, the identification of the frequency band that any terminal accesses, the uplink signal strength of any terminal, the data transmission delay, the signal-to-noise ratio, the packet loss rate, the retransmission rate, or the channel utilization rate, etc.
  • the identification of the frequency band that any terminal accesses the uplink signal strength of any terminal, the data transmission delay, the signal-to-noise ratio, the packet loss rate, the retransmission rate, or the channel utilization rate, etc.
  • the first AP can provide at least one frequency band, and a terminal accessing the first AP can select a frequency band from the at least one frequency band and access the selected frequency band, and then the terminal can use the accessed frequency band to communicate with the first AP.
  • the other AP also collects the connection information of the terminal connected to the AP like the first AP and sends the collected connection information of the terminal to the CI.
  • the CI will receive the connection information of the terminal sent by different APs, and the CI will store the connection information sent by the different APs.
  • the terminal After entering the signal coverage area of the first AP, the terminal may send an access request to the first AP to request access to the first AP. After the terminal requests to access the first AP, it may or may not successfully access the first AP.
  • the first AP When the terminal does not successfully access the first AP, the first AP will detect the terminal and generate an access failure event.
  • the access failure event may include the identity of the first AP and the start time when the terminal requests access, etc. Information, the access failure event is sent to the CI.
  • the other AP For any other AP in the at least one WLAN, the other AP generates an access failure event when detecting that a terminal that has not successfully accessed the other AP, like the first AP, and sends the generated access failure event to the CI.
  • the CI will receive the access failure events sent by different APs, and the CI saves the access failure events sent by different APs.
  • the identifier of the first AP may be the address of the first AP, for example, it may be an internet protocol (IP) address or a media access control address (MAC) used for interconnection between networks of the first AP. )address.
  • IP internet protocol
  • MAC media access control address
  • the identification of the terminal may be the address of the terminal, for example, the IP address or MAC address of the terminal.
  • the CI may periodically detect APs in the at least one WLAN to detect logical edge APs. For example, the CI may start to detect the AP in the at least one WLAN at a preset time point every day.
  • CI may divide a period into at least one time period, and the duration of each time period is equal to the first time period.
  • the AP in the WLAN is detected according to the connection information collected by the AP during the time period.
  • the CI detection process is as follows:
  • Step 202 CI obtains the number of target terminals of the first terminal according to the connection information of at least one terminal collected by the first AP during the time period.
  • the number of target terminals is a feature of the first AP, and the feature is used to reflect the access The access status of the terminal of the first AP.
  • the connection information includes the identification and collection time of the AP that collected the connection information.
  • the CI can obtain the connection information that includes the identification of the first AP and includes the collection time within the time period from the saved connection information, and the obtained connection information is the connection collected by the first AP during the time period Information, and then CI performs this step.
  • the target terminal may be determined from the at least one terminal according to the connection information of the at least one terminal collected by the first AP in the time period, and the number of target terminals may be counted.
  • the target terminal determined at this time includes at least one of a nomadic terminal or a terminal that does not belong to the target whitelist.
  • the target whitelist is the whitelist of the target WLAN, and the target WLAN is the WLAN to which the first AP belongs.
  • the number of target terminals may include At least one of the number of nomadic terminals or the number of terminals that do not belong to the target whitelist.
  • the target white list may be stored in the first AP, or the AC used to manage the target WLAN may be stored in the target white list.
  • the target whitelist can also be saved in the CI, and the target whitelist saved in the CI can be obtained from the AP belonging to the target WLAN, obtained from the AC, or received from the administrator of the target WLAN before performing this step. Sent.
  • the target whitelist of the target WLAN includes the terminal identification of each employee of the company.
  • the administrator of the target WLAN can save the target whitelist in the AC, and the administrator can also store the target whitelist in the target WLAN.
  • a whitelist of targets is saved in each AP.
  • the CI can obtain the target whitelist from the AP of the target WLAN, or receive the target whitelist input by the administrator.
  • a white list may be stored in the CI, and the white list is a white list of the WLAN.
  • the CI gets the whitelist of any WLAN, it can save the identity of each AP in the WLAN and the whitelist of the WLAN in the identity and whitelist of the AP. The corresponding relationship.
  • a terminal that does not belong to the target whitelist can be determined from the at least one terminal in the following manner.
  • the method can be:
  • the identification of each terminal in the at least one terminal and the identification of the first AP are acquired from the connection information of the at least one terminal.
  • use this whitelist as the target whitelist of the target WLAN to which the first AP belongs; or, when saving multiple whitelists in the CI, according to the identity of the first AP, compare the identity of the AP with The correspondence of the white list obtains the target white list of the WLAN to which the first AP belongs.
  • the terminals that do not belong to the target white list are determined from each terminal, and the number of terminals that do not belong to the target white list is counted.
  • the nomadic terminal may be determined from the at least one terminal in the following manner.
  • the method can be:
  • connection information sequence of any terminal is obtained from the connection information of the at least one terminal collected by the first AP during the time period, and the connection information sequence includes connections arranged by collection time Information, each connection information in the connection information sequence includes the identification of any terminal, and the interval between the collection times of two adjacent connection information in the connection information sequence does not exceed the interval threshold; the connection information sequence is determined according to the connection information sequence. Whether any terminal is a nomadic terminal. Repeat the above process to obtain the connection information sequence of each terminal in the at least one terminal, determine all nomadic terminals from the at least one terminal, and count the determined number of nomadic terminals.
  • the following operations from 2021 to 2022 may be used to determine whether any terminal is a nomadic terminal according to the connection information sequence.
  • the operations from 2021 to 2022 can be:
  • the collection time saved in the first piece of connection information in the connection information sequence can be used as the starting time for any terminal to access the first AP, and the last piece of connection information in the connection information sequence can be saved
  • the collection time is taken as the disconnection time for any terminal to disconnect from the first AP, and the stay time of the any terminal in the first AP is calculated according to the disconnection time and the start time.
  • each piece of connection information of any terminal collected by the first AP includes the start time of any terminal accessing the first AP and the stay time that any terminal has stayed at the first AP
  • the start time of any terminal accessing the first AP and the stay time of any terminal at the first AP can be obtained from the last piece of connection information in the connection information sequence, and the value in the last piece of connection information
  • the collection time is taken as the disconnection time for any terminal to disconnect from the first AP.
  • the first time period is the time period from the first time to the start time
  • the second time period is the time period from the disconnection time to the second time.
  • the first time is before the start time, and the second time is at the disconnection.
  • the time length of the first time period is equal to the preset time length
  • the time length of the second time period is equal to the preset time length
  • the preset time lengths are both greater than the interval threshold.
  • the second AP is the target WLAN except the first AP. Other APs outside.
  • connection information of any terminal collected by the second AP in the first time period which means that any terminal is not roaming from other APs of the target WLAN to the first AP.
  • connection information of any terminal collected by the second AP in the second time period which means that any terminal has not roamed from the first AP to other APs in the target WLAN.
  • the identities of other APs in the target WLAN can be obtained, and the connections collected by other APs can be obtained from the connection information saved by the CI information.
  • connection information collected by other APs it is detected whether there is connection information that includes the collection time in the first time period and the second time period and includes the identification of any terminal. If it does not exist, it is determined to be in the first time period If there is no connection information of any terminal collected by the second AP in the second time period, if it exists, it is determined that there is connection information of any terminal collected by the second AP in the first time period and the second time period.
  • connection information sequence may be used to determine whether any terminal is a nomadic terminal according to the connection information sequence.
  • connection information sequence of any terminal is used as the input of the terminal detection model, and whether the terminal is a nomadic terminal is detected through the terminal detection model.
  • the terminal detection model is obtained by training the first artificial intelligence (AI) model through the first training set.
  • the first training set includes multiple training sequences, and each training sequence is a connection of a nomadic terminal.
  • the message sequence or the connection message sequence of a non-nomadic terminal is obtained by training the first artificial intelligence (AI) model through the first training set.
  • the training sequence in the first training set includes two types.
  • One type of training sequence is a positive sample, and the training sequence of a positive sample is a connection information sequence of a nomadic terminal.
  • Another type of training sequence is a negative sample, and the training sequence of a negative sample is a non-nomadic connection information sequence.
  • connection information sequence of the terminal may be obtained first, and the connection information sequence is used as the training sequence, and then the above operations from 2021 to 2022 are used to determine whether the connection information sequence is the connection information sequence of the nomadic terminal, and if so, set the connection information sequence.
  • the category of the training sequence is a positive sample, if not, set the category of the training sequence to a negative sample.
  • the first AI model is trained using the training sequence to train the terminal detection model. Later, when the connection information sequence is detected, the connection information sequence can be detected without the above-mentioned 2021-2022 operation. Then, the terminal detection model is directly used to detect the connection information sequence.
  • the terminal detection model can be trained through the following operations 2121 to 2124.
  • the training process is:
  • each training sequence is referred to as the true category of each training sequence.
  • the first AI model used in this step may be a long short-term memory network (long short-term memory, lstm).
  • the first AI model predicts the category of each training sequence in the first training set.
  • the first AI model For each training sequence in the first training set, the first AI model extracts features from the training sequence, and predicts the first probability that the training sequence is a positive sample and the second probability that it is a negative sample based on the extracted features, the first probability The sum of the second probability and the second probability is equal to 1, and the category corresponding to the probability with the larger value of the two probabilities is output.
  • the first AI model uses the loss function to calculate the loss of the training sequence based on the true category (positive sample or negative sample) of the training sequence and the output category predicted by the first AI model for the training sequence Function value. Further, the network parameters of the first AI model are adjusted according to the loss function value of each training sequence.
  • the first AI module determines whether to continue training. When it is determined to continue training, it returns to execute 2122. When it determines to stop training, the first AI model is the terminal detection model at this time, and the return ends.
  • the loss function value of each training sequence is obtained after each training, and the obtained loss function value is curve-fitted. If the obtained curve gradually converges, And the loss function value after the last training is less than the preset threshold, it is determined to stop the training, otherwise, it is determined to continue the training.
  • the target terminal may also include a terminal that has not successfully accessed the first AP.
  • the number of target terminals may also include the number of terminals that have not successfully accessed the first AP. That is, in this step, the CI may also obtain the access failure events that include the identification of the first AP and include the access failure events within the time period from the access failure events stored in the CI, and count the number of access failure events obtained , Get the number of terminals that failed to access the first AP.
  • the number of target terminals of any AP may include at least one of the number of nomadic terminals, the number of terminals that do not belong to the target whitelist, or the number of terminals that have not successfully accessed the AP.
  • the first AP may also send the collected connection information to devices other than the CI, for example, to the AC.
  • the AC can obtain the connection information sequence of any terminal and determine whether the connection information sequence of any terminal is the connection information sequence of a nomadic terminal. If so, mark the connection information sequence of any terminal and send that any terminal to CI If not, send the connection information sequence of any terminal directly to CI.
  • the CI receives the connection information sequence of any terminal sent by the AC, and when the connection information sequence is marked, it determines that any terminal is a nomadic terminal. In this way, the detection process of the nomadic terminal is separated from the CI and implemented by other equipment, which can reduce the pressure of the calculation of the CI.
  • Step 203 the CI obtains at least one connection event of the at least one terminal, and for any one of the at least one connection event, counts the total number of any one of the connection events.
  • the total number of connection events is a feature of the first AP, and this feature is used to reflect the performance of the first AP.
  • This step is an optional step, that is, this step may not be executed, and step 204 is executed after step 202 is executed. Of course, this step may also be executed. After this step is executed, step 204 is executed again.
  • the connection information of the any terminal collected by the first AP includes at least one of the uplink signal strength of the any terminal or the data transmission delay of the any terminal One.
  • the CI has obtained the connection information sequence of any terminal.
  • the at least one connection event of any terminal may include at least one of at least one weak coverage event, at least one high latency event, at least one strong coverage event, or at least one low latency event, etc. .
  • the correspondence between the first intensity threshold and the weak coverage event may be stored in the CI, and the correspondence between the first intensity threshold and the weak coverage event stores at least one first intensity threshold and each first intensity threshold.
  • the first intensity threshold stored in the correspondence between the first intensity threshold and the weak coverage event is often small.
  • the first weak coverage event, the second weak coverage event, and the third weak coverage event shown in Table 1 below are respectively different types of connection events.
  • the correspondence between the first delay threshold and the high-latency event may be stored in the CI, and the correspondence between the first delay threshold and the high-latency event stores at least one first delay threshold and each first delay threshold.
  • the first delay threshold stored in the correspondence between the first delay threshold and the high delay event is often larger.
  • the first high-latency event, the second high-latency event, and the third-high-latency event shown in Table 2 below are respectively different types of connection events.
  • the correspondence between the second intensity threshold and the strong coverage event may be stored in the CI, and the correspondence between the second intensity threshold and the strong coverage event stores at least one second intensity threshold and each second intensity threshold.
  • the second intensity threshold stored in the correspondence between the second intensity threshold and the strong coverage event is often larger.
  • the first strong coverage event, the second strong coverage event, and the third strong coverage event shown in Table 3 below are respectively different types of connection events.
  • Second intensity threshold Strong coverage event 50db The first strong coverage event 52db The second strongest coverage event 67db The third strongest coverage event ... ...
  • the correspondence between the second delay threshold and the low-latency event may be stored in the CI, and the correspondence between the second delay threshold and the low-latency event may store at least one second delay threshold and each second delay threshold.
  • the second delay threshold stored in the correspondence between the second delay threshold and the low-latency event is often relatively small.
  • the first low-latency event, the second low-latency event, and the third low-latency event shown in Table 4 below are respectively different types of connection events.
  • Second delay threshold Weak coverage event 2
  • the first low-latency event 1.5 Second lowest latency event 1
  • the third lowest latency event ... ...
  • the uplink signal strength included in the last piece of connection information in the connection information sequence of any terminal from each of the first strength thresholds stored in the corresponding relationship between the first strength threshold and the weak coverage event, Select a first intensity threshold greater than the uplink signal intensity. According to each selected first intensity threshold, the corresponding weak coverage event is obtained from the corresponding relationship between the first intensity threshold and the weak coverage event, and the obtained weak coverage event is taken as the weak coverage event of any terminal. Or, from each of the second intensity thresholds stored in the corresponding relationship between the second intensity threshold and the strong coverage event, a second intensity threshold that is less than the uplink signal intensity is selected. According to each selected second intensity threshold, obtain the corresponding strong coverage event from the corresponding relationship between the second intensity threshold and the strong coverage event, and use the obtained strong coverage event as the strong coverage event of any terminal
  • the last piece of connection information in the connection information sequence of any terminal includes the uplink signal strength of 8, and the first strength greater than 8 is obtained from the correspondence between the first strength threshold and the weak coverage event as shown in Table 1.
  • the thresholds are 10 and 12, respectively. According to the first intensity thresholds 10 and 12, the first weak coverage event corresponding to the first intensity threshold 10 is obtained from the correspondence between the first intensity threshold and the weak coverage event as shown in Table 1, and The second weak coverage event corresponding to the second intensity threshold 12 is the first weak coverage event and the second weak coverage event of any terminal.
  • the weak coverage event of any terminal is used to indicate that any terminal may be located at the edge of the first AP, and when any terminal is far from the first AP, the first AP is at the position of any terminal
  • the coverage signal may be weak.
  • the strong coverage event of any terminal is used to indicate that any terminal may be close to the first AP, and the signal covered by the first AP at the location of any terminal may be relatively strong.
  • each first time delay is saved from the correspondence between the first delay threshold and the high-latency event.
  • the delay threshold the first delay threshold that is less than the data transmission delay is selected. According to each selected first delay threshold, obtain the corresponding high-latency event from the corresponding relationship between the first delay threshold and the high-latency event, and use the obtained high-latency event as the high-latency of any terminal event.
  • select a second delay threshold that is greater than the data transmission delay from each of the second delay thresholds stored in the corresponding relationship between the second delay threshold and the low-latency event. According to each selected second delay threshold, obtain the corresponding low-latency event from the corresponding relationship between the second delay threshold and the low-latency event, and use the obtained low-latency event as the low-latency of any terminal event.
  • the last piece of connection information in the connection information sequence of any terminal includes a data transmission delay of 6, and the corresponding relationship between the first delay threshold and the high delay event shown in Table 2 is obtained.
  • the delay thresholds are 3 and 5 respectively. According to the delay thresholds 3 and 5, the first high delay corresponding to the delay threshold 3 is obtained from the corresponding relationship between the first delay threshold and the high delay event shown in Table 2. Event, and the second high-latency event corresponding to the delay threshold of 5, that is, the first high-latency event and the second high-latency event of any terminal.
  • the high latency event of any terminal is used to indicate that any terminal may be located on the edge of the first AP, and the position of any terminal is far from the first AP, causing the first AP to be at the edge of any terminal.
  • the signal covered at the location may be weak, which in turn leads to a relatively long transmission delay required for the data sent by any terminal to the first AP.
  • the low-latency event of any terminal is used to indicate that any terminal may be closer to the first AP, and the signal covered by the first AP at the location of any terminal may be strong, which in turn causes any terminal to send to The transmission delay required for the data of the first AP is relatively small.
  • connection information of any terminal further includes at least one of the signal-to-noise ratio, packet loss rate, retransmission rate, channel utilization rate, and frequency band identifier accessed by any terminal;
  • the CI may also obtain the statistical value of the signal-to-noise ratio, the statistical value of the packet loss rate, the statistical value of the retransmission rate, and the statistical value of the channel utilization rate according to the connection information of the at least one terminal collected by the first AP during the time period. Or at least one of the total number of terminals that access the frequency band corresponding to the frequency band identifier.
  • any one of the statistical value of the signal-to-noise ratio, the statistical value of the packet loss rate, the statistical value of the retransmission rate, or the statistical value of the channel utilization is a feature of the first AP, and this feature is also used to reflect the terminal accessing the first AP Access situation.
  • the total number of terminals is also a feature of the first AP, and this feature is used to reflect the load situation of the first AP.
  • the average value of the signal-to-noise ratio may be calculated according to the signal-to-noise ratio of each of the at least one terminal, and the average value of the signal-to-noise ratio is used as the signal-to-noise ratio statistical value; Sort the signal-to-noise ratio of each of the at least one terminal, and use the signal-to-noise ratio in the middle position as the statistical value of the signal-to-noise ratio; or, from the signal-to-noise ratio of each of the at least one terminal Select the maximum signal-to-noise ratio or the minimum signal-to-noise ratio as the statistical value of the signal-to-noise ratio.
  • the average value of the packet loss rate may be calculated according to the packet loss rate of each of the at least one terminal, and the average value of the packet loss rate is used as the statistical value of the packet loss rate; or, Sort the packet loss rate of each terminal in the at least one terminal, and use the packet loss rate in the middle position as the packet loss rate statistical value; or, from the packet loss rate of each terminal in the at least one terminal Select the maximum packet loss rate or the minimum packet loss rate as the statistical value of the packet loss rate.
  • the average value of the retransmission rate may be calculated according to the retransmission rate of each of the at least one terminal, and the average value of the retransmission rate is used as the statistical value of the retransmission rate; or, Sort the retransmission rate of each terminal in the at least one terminal, and use the retransmission rate in the middle position as the retransmission rate statistical value; or, from the retransmission rate of each terminal in the at least one terminal Select the maximum retransmission rate or the minimum retransmission rate as the retransmission rate statistical value.
  • the average channel utilization rate may be calculated according to the channel utilization rate of each of the at least one terminal, and the average channel utilization rate may be used as the channel utilization statistical value; or, Sort the channel utilization rate of each terminal in the at least one terminal, and use the channel utilization rate in the middle position as the channel utilization rate statistical value; or, from the channel utilization rate of each terminal in the at least one terminal Select the maximum channel utilization rate or the minimum channel utilization rate as the channel utilization statistical value.
  • step 202 if it is determined that the category of each feature of the first AP is required before step 202 is performed, then only features belonging to the determined category need to be acquired in steps 202 and 203.
  • the determined categories include the number of nomadic terminals, the number of first high-latency events, and the number of first weak-coverage events. Then in step 202, the CI can obtain the number of nomadic terminals of the first terminal.
  • CI When acquiring the connection event of any terminal in step 203, for any terminal that accesses the first AP, CI reads the uplink signal strength and data transmission from the last connection information in the connection information sequence of any terminal Time; obtain the strength threshold corresponding to the first weak coverage event from the corresponding relationship between the strength threshold and the weak coverage event, and when the uplink signal strength is less than the obtained strength threshold, determine that the connection event of any terminal includes the first weak coverage event Covering events; obtain the delay threshold corresponding to the first high-latency event from the corresponding relationship between the delay threshold and the high-latency event, and determine the connection of any terminal when the data transmission delay is greater than the acquired delay threshold
  • the events include the first high-latency event; then CI then counts the total number of first weak coverage events and the total number of first high-latency events.
  • Step 204 CI judges whether the feature of the first AP satisfies the first condition, and detects that the first AP is a logical edge AP when the first condition is met.
  • the CI can use the following three methods to detect whether the first AP is a logical edge AP.
  • the three methods are:
  • CI detects that the first AP is a logical edge AP according to the number of target terminals of the first AP and the first condition.
  • the first condition includes that the number of target terminals is greater than the first number threshold, or the first AP is one of the first APs with the largest number of target terminals in the target WLAN, or the difference between the number of target terminals and the average number of target terminals.
  • the value is greater than the difference threshold, the difference threshold is equal to m times the variance value, and m is greater than 1.
  • the average number of target terminals and the variance value are obtained based on the number of target terminals of each AP in the target WLAN.
  • the difference between the number of target terminals and the average number of target terminals is equal to the number of target terminals minus the average number of target terminals.
  • step 203 above When the first method is adopted, the operation of step 203 above may not be performed, that is, after step 202 is performed, this step is directly performed.
  • the CI can determine whether the number of target terminals of the first AP exceeds the first number threshold, and if it exceeds, it is detected that the first AP is a logical edge AP. Or, the CI selects a preset number of APs with the largest number of target terminals from the target WLAN as logical edge APs. Or, CI calculates the average number of target terminals and variance based on the number of target terminals of each AP in the target WLAN, and subtracts the average number of target terminals from the number of target terminals of the first AP to obtain the difference, where the difference is greater than the variance value When m times of, it is detected that the first AP is a logical edge AP.
  • the first number threshold may be a value preset by CI or y times the average number of target terminals, and y is a value greater than 1.
  • the number of target terminals may be the number of nomadic terminals.
  • the CI determines whether the first AP meets the first condition according to each feature of the first AP, and when the first condition is satisfied, the first AP is determined to be a logical edge AP.
  • the first condition defines a judgment condition between any feature of the first AP and the feature threshold corresponding to the category of the feature.
  • the judgment condition between the feature and the feature threshold represents a size relationship, and the size relationship can be greater than the relationship or smaller than the relationship. If the first AP is a logical edge AP, the magnitude relationship between the feature and the feature threshold is consistent with the magnitude relationship indicated by the judgment condition.
  • each feature of the first AP is compared with the feature threshold corresponding to each feature category to obtain the comparison result of each feature; according to the feature threshold corresponding to each feature and the feature category
  • the judgment condition between and the comparison result of each feature detects whether the first AP is a logical edge AP.
  • the feature is compared with the feature threshold corresponding to the category of the feature, and the comparison result of the feature is obtained.
  • the comparison result is the magnitude relationship between the feature and the feature threshold. It is judged that the comparison result is consistent with the magnitude relationship indicated by the judgment condition corresponding to the category, indicating that the feature is a feature of the first AP conforming to the logical edge AP. According to the above method, it is concluded that each feature of the first AP is that the first AP conforms to the feature of a logical edge AP, and then the first AP is determined to be a logical edge AP.
  • the first condition includes that any feature is greater than the feature threshold corresponding to the category of any feature, or that any feature is less than the feature threshold corresponding to the category of any feature .
  • the first condition includes that the number of target terminals of the first AP is greater than the first number threshold.
  • the first condition may also include that the total number of any type of connection event counted by CI is greater than the number threshold corresponding to any type of connection event, or the total number of any type of connection event is less than any type of connection event. The threshold for the number of connection events.
  • the first condition may also include the condition that the statistical value of the signal-to-noise ratio is less than the statistical threshold of the signal-to-noise ratio, the condition that the statistical value of the packet loss rate is greater than the statistical threshold of the packet loss rate, the condition that the statistical value of the retransmission rate is greater than the statistical threshold of the retransmission rate, At least one of the condition that the statistical value of the channel utilization is greater than the channel utilization threshold or the condition that the total number of terminals in each frequency band connected to the first AP is greater than the threshold of the number of terminals corresponding to each frequency band.
  • any connection event may be a weak coverage event or a high latency event
  • the first condition includes that the weak coverage event is greater than the number threshold corresponding to the weak coverage event, or the high latency event The event is greater than the number threshold corresponding to the high-latency event.
  • any connection event may be a strong coverage event or a low latency event
  • the first condition includes that the strong coverage event is greater than the number threshold corresponding to the strong coverage event, or the low latency The event is greater than the number threshold corresponding to the low-latency event.
  • the first AP provides the first frequency band and the second frequency band for terminal access
  • the characteristics of the first AP include the number of target terminals 50, the number of first high-latency events 40, the number of second high-latency events 35, and the number of first high-latency events.
  • the number of weak coverage events is 29, the number of second weak coverage events is 36, the statistical value of signal-to-noise ratio is 58, the statistical value of packet loss rate is 0.35, the statistical value of retransmission rate is 0.48, the statistical value of channel utilization is 0.42, the first access
  • the total number of terminals in the frequency band is 48 and the total number of terminals accessing the second frequency band is 56.
  • the first number threshold is 30, the number threshold corresponding to the first high-latency event is 20, the number threshold corresponding to the second high-latency event is 22, the number threshold corresponding to the first weak coverage event is 18, and the number threshold corresponding to the second weak coverage event is 18.
  • the number threshold corresponding to the coverage event is 19, the signal-to-noise ratio statistical threshold is 60, the packet loss rate statistical threshold is 0.3, the retransmission rate statistical threshold is 0.4, the channel utilization statistical threshold is 0.35, and the number of terminals corresponding to the first frequency band is the threshold 30 and the threshold of the number of terminals corresponding to the second frequency band is 25.
  • the number of target terminals of the first AP 50 is greater than the first number threshold 30, the number of first high-latency events 40 is greater than the number threshold corresponding to the first high-latency event 15, and the number of second high-latency events 35 is greater than the second high-latency event.
  • the number threshold corresponding to extended events is 22, the number of first weak coverage events 29 is greater than the number threshold corresponding to the first weak coverage events 18, the number of second weak coverage events 36 is greater than the number threshold corresponding to the second weak coverage events 19, SNR statistics
  • the value 58 is less than the signal-to-noise ratio statistical threshold 60
  • the packet loss rate statistical value 0.35 is greater than the packet loss rate statistical threshold 0.3
  • the retransmission rate statistical value 0.48 is greater than the retransmission rate true statistical threshold 0.4
  • the channel utilization statistical value 0.42 is greater than the channel utilization statistical threshold 0.35
  • the total number of terminals accessing the first frequency band is 48 greater than the threshold number of terminals 30 corresponding to the first frequency band
  • the total number of terminals accessing the second frequency band 56 is greater than the threshold number of terminals 25 corresponding to the second frequency band. Therefore, it can be concluded that the characteristic of the first AP satisfies the first condition, thereby determining that the first AP is a logical edge AP.
  • the second method is compared with the first method.
  • the second method can improve the detection The precision.
  • the random forest model can be trained before performing the above step 202.
  • the trained random forest model has the function of detecting whether the AP is a logical edge AP, and the random forest model is used to determine the first AP that needs to be acquired.
  • the category and feature threshold corresponding to each feature. In this way, in the foregoing steps 202 and 203, only the characteristics of the first AP belonging to the determined category need to be acquired, and then the second method is used in this step to detect whether the first AP is a logical edge AP.
  • the second training set includes multiple training samples and the category corresponding to each training sample.
  • Each training sample is at least one feature of an AP, where the AP is a logical edge
  • the category of the training sample in AP is a positive sample, and the category of the training sample is a negative sample when the AP is a non-logical edge AP.
  • At least one feature of the logical edge AP is obtained through the above steps 201 to 203, the at least one feature is used as a training sample, and the type of the training sample is set to Positive sample.
  • at least one feature of the non-logical edge AP is obtained through the above steps 201 to 203, the at least one feature is used as a training sample, and the type of the training sample is set to negative sample.
  • the constructed second training set may have fewer training samples of positive samples, for example, the number of positive samples in the second training set is less than the third number threshold. This is mainly because there are usually fewer logical edge APs. In any WLAN, there may or may not be logical edge APs in the WLAN. In the case that the WLAN has logical edge APs, the number of logical edge APs in the WLAN is also small, usually only a few. Therefore, technicians may not have enough known logical edge APs to form training samples of positive samples, resulting in a small number of training samples of positive samples in the constructed second training set.
  • the second training set constructed has more training samples of positive samples.
  • the second training set can be used to train in addition to the random forest.
  • the random forest model can be trained through the following operations from 2041 to 2044.
  • the operations from 2041 to 2044 are:
  • each training sample is referred to as the true category of each training sample.
  • the random forest model predicts the category of each training sample in the second training set.
  • the random forest model For each training sample in the second training set, the random forest model extracts features from the training sample, and predicts the first probability that the training sample is a positive sample and the second probability that the training sample is a negative sample based on the extracted features.
  • the sum of the first probability and the second probability is equal to 1, and the category corresponding to the probability with the larger value of the two probabilities is output.
  • the random forest model uses the loss function to calculate the loss function value of the training sample according to the true category of the training sample and the output category predicted by the random forest model for the training sample. Further, the parameters of the random forest model are adjusted according to the loss function values of all training samples in the second training set.
  • the random forest model determines whether to continue training. When it is determined to continue training, it returns to execution 2042, and when it determines to stop training.
  • the loss function value of each training sample is obtained after each training, and the obtained loss function value is curve-fitted. If the obtained curve gradually converges, And the loss function value after the last training is less than the preset threshold, it is determined to stop the training, otherwise, it is determined to continue the training.
  • the random forest model obtained after training includes at least one decision tree, and each path in the decision tree is used to detect whether an AP in the second training set is a logical edge AP.
  • the leaf node of the path is used to save the detection result of the path.
  • the nodes in the path other than the leaf node correspond to a category and feature threshold. The node is used to determine whether the feature belonging to the category exceeds the feature threshold and according to the judgment result Select the next level node belonging to the path.
  • a decision tree of the random forest model is shown.
  • every node in the decision tree except the leaf node stores a feature category and feature threshold.
  • the category saved by root node 1 is the number of nomadic terminals and a stored feature threshold of 50. Root node 1 is used to determine whether the number of nomadic terminals exceeds the feature threshold of 50, and then select the path according to the result of the judgment.
  • the next level of node is node 2 or node 3.
  • the category saved by node 2 is the number of first high-latency events and the saved feature threshold is 30.
  • Node 2 is used to determine whether the number of first high-latency events exceeds the feature threshold of 30, and then select the next node of the path according to the result of the judgment ,
  • the next layer node is node 4 or node 5.
  • the category saved by node 4 is the number of first weak coverage events and the saved feature threshold is 20.
  • Node 4 is used to determine whether the number of first weak coverage events exceeds the feature threshold of 20, and then select the next node of the path according to the judgment result.
  • the next layer node is a leaf node 6 or a leaf node 7.
  • the leaf node 6 or the leaf node 7 is used to store the detection result of the AP, and the detection result can be a logical edge AP or a non-logical edge AP.
  • the trained random forest model is overfitted and the generalization ability of the model is poor. In this way, for the first AP to be detected, the trained random forest model is directly used to detect whether the first AP is a logical edge AP, and the detection accuracy is low. Therefore, in the second method, the trained random forest model is not used directly to detect the first AP.
  • the trained random forest model determine the category corresponding to each feature of the first AP that needs to be acquired, the feature threshold, and the judgment condition between the feature and the feature threshold, and then the CI can only be used in the above steps 202 to 203
  • the feature corresponding to the determined category is acquired, so that the number of features for acquiring the first AP can be reduced, and the amount of calculation can be reduced.
  • the following operations (1)-(5) can be used to determine the category and feature threshold corresponding to each feature of the first AP that needs to be acquired.
  • the operations of (1)-(5) are:
  • the target category is a category to which any feature in the second training set belongs.
  • the target node judges the magnitude relationship between the features belonging to the target category and the feature threshold corresponding to the target node, and obtains the judgment result.
  • the judgment condition corresponding to the judgment result is the magnitude relation, that is, the judgment result corresponds to
  • the judgment condition may be greater than the relationship or less than the relationship.
  • the judgment condition corresponding to the judgment result is used to indicate that the feature is greater than the feature threshold or the feature is less than the feature threshold.
  • the number of acquired target nodes may be greater than or equal to 1, and there may be a judgment condition corresponding to a part of the target nodes that is greater than the relationship, and the judgment conditions corresponding to the remaining target nodes are less than the relationship.
  • the target category is determined as the category corresponding to the feature of the first AP that needs to be acquired.
  • an average value is calculated according to the acquired feature threshold value corresponding to each node, and the average value is used as the feature threshold value corresponding to the target category.
  • the acquired feature thresholds corresponding to each target node are sorted, and the feature thresholds ranked in the middle position are used as the feature thresholds corresponding to the target category.
  • the logical edge AP detection model is used to detect whether the first AP is a logical edge AP. That is, in the third method, the logical edge AP judges whether at least one feature of the first AP satisfies the first condition, and when it is judged that the first condition is satisfied, the result that the first AP is the logical edge AP is output.
  • the second training set can be used to train the second AI model to obtain Logical edge AP detection model, and in the third method, the logical edge AP detection model can use support vector machine (SVM), linear regression (LR) or convolutional neural network, CNN) and other AI models.
  • SVM support vector machine
  • LR linear regression
  • CNN convolutional neural network
  • the second AI model can be trained through the following operations 2141 to 2144.
  • the operations from 2141 to 2144 are:
  • each training sample is referred to as the true category of each training sample.
  • the second AI model predicts the category of each training sample in the second training set.
  • the second AI model For each training sample in the second training set, the second AI model extracts features from the training sample, and predicts the first probability of the training sample being a positive sample and the second probability of being a negative sample based on the extracted features, the first probability The sum of the second probability and the second probability is equal to 1, and the category corresponding to the probability with the larger value of the two probabilities is output.
  • the second AI model uses the loss function to calculate the loss function value of the training sample according to the true category of the training sample and the output category predicted by the second AI model for the training sample. Further, the network parameters of the second AI model are adjusted according to the loss function values of all training samples in the second training set.
  • the second AI module determines whether to continue training. When it is determined to continue training, it returns to execute 2142. When it determines to stop training, the second AI module at this time is a logical edge AP detection model, and the return ends.
  • the CI repeats the above process 201 to 204 to detect each logical edge AP from the AP of the target WLAN.
  • Step 205 The CI sends an optimization request to the AC.
  • the optimization request includes the identification of each logical edge AP.
  • the optimization request may also include information such as the number of target terminals of each logical edge AP.
  • Step 206 The AC may receive the optimization request, and optimize the performance of the logical edge AP according to the identifier of the logical edge AP included in the optimization request.
  • the following three optimization methods can be used to optimize the logical edge AP.
  • the three optimization methods are:
  • AC can reduce the signal coverage of logical edge APs.
  • the signal coverage of the logical edge AP usually exceeds the location where the target WLAN is located.
  • the signal coverage of the logical edge AP may include pedestrian channels outside the location. Therefore, by reducing the signal coverage of the logical edge AP, you can make the logical edge
  • the signal coverage of the AP does not include the people flow channel located outside the location, which can reduce the number of target terminals requesting access to the logical edge AP.
  • the AC can reduce the signal coverage of the logical edge AP by reducing the transmit power of the logical edge AP.
  • the AC can reduce the transmit power of the logical edge AP multiple times to gradually reduce the signal coverage of the logical edge AP. After reducing the transmit power of the logical edge AP each time, it is determined whether to continue to reduce the transmit power of the logical edge AP.
  • the AC can reduce the transmit power of the logical edge AP with a fixed step.
  • the fixed step size is 2db, that is, the AC reduces the transmit power of the logical edge AP by 2db each time.
  • the AC when the AC needs to reduce the transmit power of the logical edge AP, it may send an instruction to the logical edge AP, and the instruction may include the fixed step size.
  • the logical edge AP receives the instruction and reduces its own transmission power according to the fixed step in the instruction.
  • the AC can also notify the administrator to inform the administrator that the transmit power of the logical edge AP needs to be reduced and how much power needs to be reduced, and request the administrator to confirm .
  • the AC reduces the transmit power of the logical edge AP after receiving the confirmation from the administrator.
  • the AC After reducing the transmit power of the logical edge AP, the AC needs to determine whether to continue to reduce the transmit power of the logical edge AP. In implementation, after each time the transmit power of the logical edge AP is reduced, the AC may request the CI to count the number of target terminals requesting access to the logical edge AP during the first period of time according to the operations of steps 201 to 202 above. When the number of target terminals requesting to access the logical edge AP during this time period is less than the second number threshold, the reduction of the transmission power of the logical edge AP may be stopped.
  • the second number threshold may be a preset threshold, or may be obtained according to the maximum number of target terminals, and the maximum number of target terminals is the maximum value among the number of target terminals of logical edge APs that have been counted.
  • the second number threshold may be equal to x times the maximum number of target terminals, and x is a value less than 1, and may be 0.1, 0.2, or 0.3. or,
  • the AC Since the AC reduces the transmit power of the logical edge AP each time, it can obtain the number of target terminals requesting access to the logical edge AP. Regarding the number of target terminals obtained by the AC for multiple consecutive times, the AC finds that the difference between the number of any two target terminals in the number of target terminals is less than the difference threshold, which means that after the AC reduces the transmit power of the logical edge AP, it requests The number of target terminals that access logical edge APs will no longer continue to drop significantly. At this time, the AC also determines to stop reducing the transmit power of the logical edge AP. or,
  • the AC obtains the actual transmit power of the logical edge AP every time after reducing the transmit power of the logical edge AP.
  • the actual transmit power of the logical edge AP is lower than the power threshold, it determines to stop reducing the transmit power of the logical edge AP. or,
  • the AC reduces the transmit power of the logical edge AP each time the AC reduces the transmit power of the logical edge AP, it detects whether the target WLAN has a coverage hole, and when a coverage hole occurs, it determines to stop reducing the transmit power of the logical edge AP.
  • a monitor is set in the WLAN, and the AC can monitor whether there is a coverage hole in the WLAN through the monitor, and when a coverage hole occurs, it stops reducing the transmit power of the logical edge AP.
  • the AC can also increase the transmit power of the logical edge AP, and stop increasing the transmit power of the logical edge AP after the monitor detects that the coverage hole in the WLAN disappears.
  • the AC can control the logical edge AP to block the access of terminals whose downlink signal strength is lower than the downlink signal strength threshold.
  • Terminals whose downlink signal is lower than the downlink signal strength threshold are often located at the edge of the signal coverage of the logical edge AP, and the edge of the signal coverage of the logical edge AP may be outside the location of the target WLAN, so the terminal is often far away from the logical edge AP. As a result, the downlink signal strength is low, which is lower than the downlink signal strength threshold.
  • the AC can send a control instruction to the logical edge AP.
  • the logical edge AP receives the control instruction, when a terminal requests to access the logical edge AP, it obtains the downlink signal strength of the terminal, and the downlink signal When the strength is lower than the downlink signal strength threshold, the terminal is denied access, which can reduce the number of target terminals that access the logical edge AP.
  • the AC can activate the delayed access function of the logical edge AP to reduce the number of target terminals connected to the logical edge AP.
  • the logical edge AP includes a delayed access function. After the delayed access function is activated, when a terminal requests to access the logical edge AP, the logical edge AP will not immediately execute the access process for the terminal to access, but wait For a period of time, when the waiting time exceeds the first time threshold, the access flow is started to allow the terminal to access.
  • the time that the nomadic terminal stays in the signal coverage area of the logical edge AP is often short, less than the first time threshold. In this way, when the nomadic terminal enters the signal coverage area of the logical edge AP and requests access, because The logical edge AP will not immediately allow nomadic terminals to access, but will start the access process after the waiting time exceeds the first time threshold. In this way, after the logical edge AP starts the access process, the nomadic terminal has left the signal coverage of the logical edge AP, so the logical edge AP will stop continuing to execute the access process, which can reduce the number of nomadic terminals accessing the logical edge AP. number.
  • the AP in the WLAN collects the connection information of at least one terminal, and sends the collected connection information of the at least one terminal to the CI.
  • the CI obtains at least one characteristic of the AP according to the connection information of the at least one terminal.
  • the at least one characteristic may include the number of target terminals of the AP or the total number of various connection events generated by the at least one terminal, etc., and then according to the at least one characteristic of the AP.
  • the feature detects whether the AP is a logical edge AP. In this way, the logical edge AP can be accurately detected from the WLAN AP, which can facilitate processing of the logical edge AP, for example, optimize the performance of the logical edge AP to reduce or eliminate the impact of the logical edge AP.
  • an embodiment of the present application provides a detection device 300, which can be deployed in the CI in any of the foregoing embodiments, including:
  • the processing unit 301 is configured to obtain at least one feature of the first access point AP, the at least one feature includes the number of target terminals requesting to access the first AP within the first time period, and the target terminal is in an abnormal access state Terminal
  • the detecting unit 302 is configured to detect whether the first AP is a logical edge AP according to at least one characteristic, and the logical edge AP is an AP whose signal coverage reaches the edge of the signal coverage of the wireless local area network WLAN.
  • the target terminal includes at least one of a nomadic terminal, a terminal that has not successfully accessed, or a terminal that does not belong to the whitelist of the WLAN to which the first AP belongs.
  • the nomadic terminal stays in the accessed AP for a time that does not exceed a first time threshold, and the nomadic terminal is within a preset time period before accessing the accessed AP and is disconnected and disconnected. After the accessed AP is connected, any AP that does not access the WLAN is not accessed within a preset period of time.
  • the detection unit 302 is configured to:
  • the first AP is a logical edge AP
  • the first condition includes that the number of target terminals is greater than a first number threshold, or the first AP is One of the first APs with the largest number of target terminals in the WLAN, or the difference between the number of target terminals and the average number of target terminals is greater than a difference threshold, and the difference threshold is equal to the variance value M times of, m is greater than 1, the average number of target terminals and the variance value are obtained based on the number of target terminals of each AP in the WLAN.
  • the target terminal includes at least one of the nomadic terminal or a terminal that does not belong to the whitelist, and
  • the device 300 further includes: a receiving unit 303,
  • the receiving unit 303 is configured to receive connection information of at least one terminal collected by a first AP, where the connection information includes an identifier of the terminal, and the at least one terminal is a terminal that accesses the first AP;
  • the processing unit 301 is configured to determine a target terminal from the at least one terminal according to the connection information of the at least one terminal; and to count the determined number of target terminals.
  • the time interval for the first AP to collect connection information from the same terminal twice does not exceed an interval threshold, and the connection information further includes the collection time;
  • the processing unit 301 is configured to:
  • connection information sequence of the any terminal is acquired, the connection information sequence includes connection information arranged by collection time, and each connection information in the connection information sequence includes all the connection information.
  • the interval between the collection times of two adjacent connection information in the connection information sequence does not exceed the interval threshold;
  • processing unit 301 is configured to:
  • connection information sequence the start time of any terminal accessing the first AP, the stay time of any terminal in the first AP, and the disconnection of any terminal from the first AP The disconnection time of the connection between the first AP;
  • connection information determines that the any terminal is a nomadic terminal, the second AP is an AP other than the first AP in the WLAN, and the preset duration is greater than the interval threshold.
  • processing unit 301 is configured to:
  • connection information sequence is used as an input of a terminal detection model, and whether any terminal is a nomadic terminal is detected through the terminal detection model.
  • the at least one feature further includes the total number of each connection event in the at least one connection event of the at least one terminal.
  • the first A condition further includes that the total number of any connection event is greater than the number threshold corresponding to any connection event, or the first condition further includes that the total number of any connection event is less than the any connection event.
  • the number threshold corresponding to a connection event is the number threshold corresponding to a connection event.
  • connection information of any terminal further includes at least one of uplink signal strength of any terminal or data transmission delay of any terminal;
  • the processing unit 301 is further configured to:
  • the first condition includes that the total number of weak coverage events is greater than a threshold number corresponding to the weak coverage events
  • the processing unit 301 is configured to:
  • a corresponding weak coverage event is acquired from the corresponding relationship between the intensity threshold and the weak coverage event as the connection event of any terminal.
  • the first condition includes that the total number of high-latency events is greater than the number threshold corresponding to the high-latency events
  • the processing unit 301 is configured to:
  • a corresponding high-latency event is acquired from the corresponding relationship between the delay threshold and a high-latency event as the connection event of any terminal.
  • connection information further includes at least one of a signal-to-noise ratio, a packet loss rate, a retransmission rate, a channel utilization rate of the terminal, and an identifier of the frequency band accessed by the terminal;
  • the at least one feature further includes at least one of statistical value of signal-to-noise ratio, statistical value of packet loss rate, statistical value of retransmission rate, statistical value of channel utilization, or the total number of terminals accessing the frequency band corresponding to the frequency band identifier, so
  • the statistical value of the signal-to-noise ratio is obtained based on the signal-to-noise ratio of the at least one terminal
  • the statistical value of the packet loss rate is obtained based on the packet loss rate of the at least one terminal
  • the statistical value of the retransmission rate is based on Obtained by the retransmission rate of the at least one terminal
  • the channel utilization statistical value is obtained based on the channel utilization rate of the at least one terminal;
  • the first condition further includes that the statistical value of the signal-to-noise ratio is less than the statistical threshold of the signal-to-noise ratio, the statistical value of the packet loss rate is greater than the statistical threshold of the packet loss rate, the statistical value of the retransmission rate is greater than the statistical threshold of the retransmission rate, and At least one of the channel utilization statistical value is greater than the channel utilization threshold or the total terminal number is greater than the terminal number threshold.
  • processing unit 301 is further configured to:
  • the training set includes a plurality of training samples and a category corresponding to each training sample.
  • the training sample whose category is a positive sample includes at least one feature of the logical edge AP, and the training sample whose category is a negative sample includes At least one feature of a non-logical edge AP
  • the random forest model after training includes at least one decision tree, and each path in the decision tree is used to detect whether any AP in the training set is a logical edge AP, so
  • the leaf nodes of the path are used to save the detection results of the path, the nodes in the path other than the leaf nodes correspond to a category and feature threshold, and the node is used to determine whether the first feature exceeds the feature corresponding to the node Threshold and selecting the next layer node belonging to the path according to the judgment result, the first feature is the feature of any AP belonging to the corresponding category of the node;
  • processing unit 301 is configured to:
  • Selecting a target path from the at least one decision tree, and the detection result of the target path is a logical edge AP
  • the detection unit 302 is configured to:
  • the at least one feature is used as an input of a logical edge AP identification model, and whether the first AP is a logical edge AP is detected through the logical edge AP identification model.
  • processing unit 301 is further configured to:
  • a threshold terminal accesses the first AP or instructs the AC to start the delayed access function of the first AP.
  • the processing unit obtains at least one characteristic of the first access point AP, and the at least one characteristic includes the number of target terminals requesting access to the first AP within a time period of the first time period, where the target terminal is A terminal with an abnormal access state; the detection unit detects whether the first AP is a logical edge AP according to at least one feature. Since the characteristics of the first AP acquired by the acquiring unit include the number of target terminals requesting access to the first AP within the first time period, and the target terminals are terminals with abnormal access status, the detection unit can be based on the first The characteristics of the AP can successfully detect whether the first AP is a logical edge AP.
  • an embodiment of the present application provides a detection device 400, which can be deployed in the CI in any of the foregoing embodiments, and includes:
  • the receiving unit 401 is configured to receive connection information of at least one terminal collected by the first access point AP, where the connection information includes the identification of the terminal and the collection time for collecting the connection information, and the at least one terminal is an access point. Terminal to the first AP;
  • the processing unit 402 is configured to obtain, for any one of the at least one terminal, a connection information sequence of the any terminal, the connection information sequence includes connection information arranged by collection time, and in the connection information sequence Each connection information includes the identifier of any one of the terminals, and the interval between collection times of two adjacent connection information in the connection information sequence does not exceed the interval threshold;
  • the processing unit 402 is further configured to determine whether the any terminal is a nomadic terminal according to the connection information sequence.
  • processing unit 402 is configured to:
  • connection information sequence the start time of any terminal accessing the first AP, the stay time of any terminal in the first AP, and the disconnection of any terminal from the first AP are obtained according to the connection information sequence.
  • connection information determines that the any terminal is a nomadic terminal, and the second AP is an AP other than the first AP in the WLAN to which the first AP belongs.
  • connection information of at least one terminal collected by the first AP received by the receiving unit Since the connection information includes the identification of the terminal and the collection time of collecting the connection information, the processing unit can obtain the connection information to any terminal.
  • a connection information sequence the connection information sequence includes connection information arranged by collection time, each connection information in the connection information sequence includes the identification of any terminal, and the collection time of two adjacent connection information in the connection information sequence The interval does not exceed the interval threshold; in this way, the processing unit can successfully determine whether any terminal is a nomadic terminal according to the connection information sequence.
  • FIG. 8 is a schematic diagram of a detection device 500 provided by an embodiment of the application.
  • the device 500 includes at least one processor 501, a bus system 502, a memory 503, and at least one transceiver 504.
  • the device 800 is a device with a hardware structure, and can be used to implement the functional modules in the device 300 described in FIG. 6 or the device 400 described in FIG. 7.
  • the processing unit 301 and the detection unit 302 in the device 300 shown in FIG. 6, or the processing unit 402 in the device 400 shown in FIG. 7 can call the memory 503 through the at least one processor 501.
  • the receiving unit 303 in the device 300 shown in FIG. 6 or the receiving unit 401 in the device 400 shown in FIG. 7 may be implemented by the transceiver 504.
  • the device 500 can also be used to implement the CI function in any of the foregoing embodiments.
  • the above-mentioned processor 501 may be a general-purpose central processing unit (central processing unit, CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more for controlling the computer Apply for integrated circuits for program execution.
  • CPU central processing unit
  • ASIC application-specific integrated circuit
  • the above-mentioned bus system 502 may include a path for transferring information between the above-mentioned components.
  • the aforementioned transceiver 504 is used to communicate with other devices or a communication network.
  • the above-mentioned memory 503 may be a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, random access memory (RAM), or other types that can store information and instructions.
  • the type of dynamic storage device can also be electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM), or other optical disk storage, CD-ROM Storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or can be used to carry or store desired program codes in the form of instructions or data structures and can be used by Any other medium accessed by the computer, but not limited to this.
  • the memory can exist independently and is connected to the processor through a bus.
  • the memory can also be integrated with the processor.
  • the memory 503 is used to store application program codes for executing the solutions of the present application, and the processor 501 controls the execution.
  • the processor 501 is configured to execute the application program code stored in the memory 503, so as to realize the functions in the method of the present patent.
  • the processor 501 may include one or more CPUs, such as CPU0 and CPU1 in FIG. 8.
  • the apparatus 500 may include multiple processors, such as the processor 501 and the processor 507 in FIG. 8. Each of these processors can be a single-CPU (single-CPU) processor or a multi-core (multi-CPU) processor.
  • the processor here may refer to one or more devices, circuits, and/or processing cores for processing data (for example, computer program instructions).
  • the apparatus 500 may further include an output device 505 and an input device 506.
  • the output device 505 communicates with the processor 501 and can display information in a variety of ways.
  • the output device 505 may be a liquid crystal display (LCD) or the like.
  • the input device 506 communicates with the processor 501 and can accept user input in a variety of ways.
  • the input device 506 may be a touch screen device or a sensor device or the like.
  • the program can be stored in a computer-readable storage medium.
  • the storage medium mentioned can be a read-only memory, a magnetic disk or an optical disk, etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Electromagnetism (AREA)
  • Quality & Reliability (AREA)
  • Computer Security & Cryptography (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Small-Scale Networks (AREA)

Abstract

本申请实施例公开了一种检测方法、装置及系统,属于通信领域。所述方法包括:获取第一接入点AP的至少一个特征,所述至少一个特征包括在时长为第一时长的时间段内请求接入所述第一AP的目标终端数目,所述目标终端为接入状态异常的终端;根据所述至少一个特征,检测所述第一AP是否为逻辑边缘AP,逻辑边缘AP是信号覆盖范围抵达所在无线局域网WLAN的信号覆盖范围边缘的AP。采用本申请,可以成功检测出第一AP是否为逻辑边缘AP。

Description

检测方法、装置及系统
本申请要求于2019年9月2日提交的申请号为201910824288.7、发明名称为“识别边缘接入点的方法、装置和系统”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及通信领域,特别涉及一种检测方法、装置及系统。
背景技术
无线局域网(wireless local area network,WLAN)通常包括至少一个接入点(access point,AP),用户可以接入到WLAN中的一个AP,以实现接入到该WLAN中。AP可以为用户提供服务,为了能够使AP更好地为用户提供服务,目前还设有接入控制器(access point controller,AC),AC可以对AP进行管理,例如,AC可以对AP进行优化,以使该AP可以更好地提供服务。
WLAN中可能存在一类AP,该AP的信号覆盖范围抵达WLAN的信号覆盖范围的边缘,且该AP的信号覆盖范围内可能有人流通行的通道。这类AP的存在可能会给WLAN产生较大的影响,因此急需要检测出这类AP,以对这类AP进行优化,减小对WLAN产生的影响,所以如何检测这类AP是目前急需要解决的问题。
发明内容
本申请实施例提供了一种检测方法、装置及系统,能够检测出逻辑边缘AP,所述技术方案如下:
第一方面,本申请提供了一种检测方法,在所述方法中:获取第一接入点AP的至少一个特征,该至少一个特征包括在时长为第一时长的时间段内请求接入第一AP的目标终端数目,目标终端为接入状态异常的终端;根据至少一个特征,检测第一AP是否为逻辑边缘AP,逻辑边缘AP是信号覆盖范围抵达所在无线局域网WLAN的信号覆盖范围边缘的AP。由于获取的第一AP的特征包括在时长为第一时长的时间段内请求接入第一AP的目标终端数目,而目标终端为接入状态异常的终端,从而基于第一AP的特征能够成功检测第一AP是否为逻辑边缘AP。
在一种可能的实现方式中,目标终端包括游牧终端、没有成功接入的终端或不属于第一AP所属WLAN的白名单中的终端中的至少一种。这几种类别的终端通常会出现在逻辑边缘AP的信号覆盖范围中,所以根据目标终端数目可以成功检测第一AP是否为逻辑边缘AP。
在另一种可能的实现方式中,在至少一个特征满足第一条件的情况下,确定第一AP为逻辑边缘AP,第一条件包括目标终端数目大于第一数目阈值,或者,第一AP是WLAN中 的目标终端数目最大的第一个数个AP中的一个AP,或者,目标终端数目与目标终端平均数目的差值大于差值阈值,该差值阈值等于方差值的m倍,m大于1,目标终端平均数目和方差值是基于WLAN中的各AP的目标终端数目得到的。从而通过第一条件可以成功检测出第一AP是否为逻辑边缘AP。
在另一种可能的实现方式中,目标终端包括游牧终端或不属于该白名单中的终端中的至少一种,接收第一AP采集的至少一个终端的连接信息,该连接信息包括该终端的标识,至少一个终端为接入到第一AP的终端;根据该至少一个终端的连接信息,从至少一个终端中确定出目标终端;统计确定出的目标终端的数目。由于第一AP采集终端的连接信息,这样可以基于该连接信息成功确定出目标终端,进而能够成功检测第一AP是否为逻辑边缘AP。
在另一种可能的实现方式中,第一AP用于对同一终端进行相邻两次采集连接信息的时间间隔不超过间隔阈值,该连接信息还包括采集时间;对于至少一个终端中的任一终端,获取该任一终端的连接信息序列,该连接信息序列包括按采集时间排列的连接信息,该连接信息序列中的每个连接信息包括该任一终端的标识,该连接信息序列中的相邻两个连接信息的采集时间之间的间隔不超过所述间隔阈值;根据该连接信息序列确定该任一终端是否为游牧终端。如此,由于能够从AP采集的连接信息中得到一个终端的连接信息序列,这样基于该连接信息序列可以成功地确定出游牧终端。
在另一种可能的实现方式中,根据该连接信息序列获取该任一终端接入第一AP的起始时间、该任一终端在所述第一AP中的停留时间和该任一终端断开与第一AP之间连接的断开时间;在停留时间未超过第一时间阈值,以及在起始时间之前的预设时长内和在断开时间之后的预设时长内没有第二AP采集的该任一终端的连接信息,确定该任一终端为游牧终端,第二AP为WLAN中的除第一AP之外的其他AP,预设时长大于所述间隔阈值。由于获取该任一终端的停留时间、接入的起始时间和断开连接的断开时间,从而基于该停留时间、起始时间和断开时间,能够成功地确定该任一终端是否为游牧终端。
在另一种可能的实现方式中,该至少一个特征还包括所述至少一个终端的至少一种连接事件中的每种连接事件的总数目,对每种连接事件中的任一种连接事件,第一条件还包括该任一种连接事件的总数目大于该任一种连接事件对应的数目阈值,或者,第一条件还包括该任一种连接事件的总数目小于该任一种连接事件对应的数目阈值。由于该至少一个特征还包括每种连接事件的总数目,如此可以提高检测逻辑边缘AP的准确性。
在另一种可能的实现方式中,该任一终端的连接信息还包括该任一终端的上行信号强度或该任一终端的数据传输时延中的至少一个;根据该任一终端的连接信息序列中的最后一条连接信息包括的上行信号强度或数据传输时延中的至少一个,获取该任一终端的连接事件;统计该任一种连接事件的总数目。如此可以得到终端的连接事件,进而可以统计连接事件的总数目,基于该连接事件的总数目可以提高检测逻辑边缘AP的准确性。
在另一种可能的实现方式中,该连接信息还包括终端的信噪比、丢包率、重传率、信道利用率和终端接入的频段标识中的至少一个;至少一个特征还包括信噪比统计值、丢包率统计值、重传率统计值、信道利用率统计值或接入该频段标识对应的频段的总终端数目中的至少一个,信噪比统计值是基于至少一个终端的信噪比得到的,丢包率统计值是基于至少一个终端的丢包率得到的,重传率统计值是基于至少一个终端的重传率得到的,信道利用率统计值是基于至少一个终端的信道利用率得到的;第一条件还包括信噪比统计值小于信噪比统计阈值、丢包率统计值大于丢包率统计阈值、重传率统计值大于重传率统计阈值、信道利用率统计值大于信道利用率阈值或总终端数目大于终端数目阈值中的至少一个。如此丰富了第一AP的特征,可以提高检测逻辑边缘AP的准确性。
在另一种可能的实现方式中,使用训练集训练随机森林模型,训练集包括多个训练样本和每个训练样本对应的类别,类别为正样本的训练样本包括逻辑边缘AP的至少一个特征,类别为负样本的训练样本包括非逻辑边缘AP的至少一个特征,训练后的所述随机森林模型包括至少一个决策树,决策树中的每条路径用于检测所述训练集中的任一AP是否为逻辑边缘AP,路径的叶子节点用于保存该路径的检测结果,该路径中除叶子节点之外的节点对应一个类别和特征阈值,该节点用于判断第一特征是否超过该节点对应的特征阈值以及根据判断结果选择属于该路径的下一层节点,第一特征是该任一AP的属于该节点对应类别的特征;根据至少一个决策树确定需要获取的第一AP的各特征对应的类别、特征阈值以及特征与所述特征阈值之间的判断条件。如此可以根据训练后的随机森林模型,确定需要获取的第一AP的各特征对应的类别,这样可以确定出用于反应逻辑边缘AP特性的特征,基于该特征可以成功检测出逻辑边缘AP,另外还可减小获取的特征数目,减少运算的数据量。
在另一种可能的实现方式中,在检测出第一AP为逻辑边缘AP时,指示接入控制器AC减小第一AP的信号覆盖范围或指示所述AC控制第一AP阻止下行信号强度小于下行信号强度阈值的终端接入第一AP或指示AC启动第一AP的延迟接入功能。如此可以实现对逻辑边缘AP的性能进行优化,减小逻辑边缘AP带来的影响。
第二方面,本申请提供了一种检测方法,在所述方法中:接收第一接入点AP采集的至少一个终端的连接信息,该连接信息包括终端的标识和采集该连接信息的采集时间,至少一个终端为接入到第一AP的终端;对于至少一个终端中的任一终端,获取该任一终端的连接信息序列,该连接信息序列包括按采集时间排列的连接信息,该连接信息序列中的每个连接信息包括该任一终端的标识,该连接信息序列中的相邻两个连接信息的采集时间之间的间隔不超过该间隔阈值;根据该连接信息序列确定该任一终端是否为游牧终端。由于能够从AP采集的连接信息中得到一个终端的连接信息序列,这样基于该连接信息序列可以成功地确定出游牧终端。
在另一种可能的实现方式中,根据该连接信息序列获取任一终端接入第一AP的起始时间、任一终端在第一AP中的停留时间和任一终端断开与第一AP之间连接的断开时间;在停留时间未超过第一时间阈值,以及在起始时间之前的预设时长内和在断开时间之后的预设时 长内没有第二AP采集的任一终端的连接信息,确定任一终端为游牧终端,第二AP为第一AP所属WLAN中的除第一AP之外的其他AP。由于获取该任一终端的停留时间、接入的起始时间和断开连接的断开时间,从而基于该停留时间、起始时间和断开时间,能够成功地确定该任一终端是否为游牧终端。
第三方面,本申请提供了一种检测装置,用于执行第一方面或第一方面的任意一种可能实现方式中的方法。具体地,所述装置包括用于执行第一方面或第一方面的任意一种可能实现方式的方法的单元。
第四方面,本申请提供了一种检测装置,用于执行第一方面或第一方面的任意一种可能实现方式中的方法。具体地,所述装置包括用于执行第一方面或第一方面的任意一种可能实现方式的方法的单元。
第五方面,本申请实施例提供了一种检测装置,所述装置包括:处理器和存储器。其中,所述处理器以及所述存储器之间可以通过总线系统相连。所述存储器用于存储一个或多个程序,所述处理器用于执行所述存储器中的一个或多个程序,完成第一方面、第二方面、第一方面的任意可能实现方式或第二方面的任意可能实现方式中的方法。
第六方面,本申请提供了一种计算机可读存储介质,计算机可读存储介质中存储有指令,当其在处理器上运行时,使得处理器执行上述第一方面、第二方面、第一方面的任意可能实现方式或第二方面的任意可能实现方式中的方法。
第七方面,本申请提供了一种包含指令的计算机程序产品,当其在处理器上运行时,使得处理器执行上述第一方面、第二方面、第一方面的任意可能实现方式或第二方面的任意可能实现方式中的方法。
第八方面,本申请提供了一种检测系统,所述系统包括:数据分析器CI和第一接入点AP,第一AP采集的至少一个终端的连接信息,向CI发送该连接信息,该连接信息包括终端的标识和采集该连接信息的采集时间,该至少一个终端为接入到第一AP的终端;CI根据该连接信息获取第一AP的至少一个特征,至少一个特征包括在时长为第一时长的时间段内请求接入第一AP的目标终端数目,目标终端包括游牧终端或不属于第一AP所在无线局域网WLAN的白名单中的终端中至少一种;根据至少一个特征,检测第一AP是否为逻辑边缘AP,逻辑边缘AP是信号覆盖范围抵达所在WLAN的信号覆盖范围边缘的AP。由于获取的第一AP的特征包括在时长为第一时长的时间段内请求接入第一AP的目标终端数目,而目标终端包括游牧终端或不属于第一AP所在WLAN的白名单中的终端中至少一种,从而基于第一AP的特征能够成功检测第一AP是否为逻辑边缘AP。
在一种可能的实现方式中,所述系统还包括:接入控制器AC,CI在检测出第一AP为逻辑边缘AP时,向AC发送优化请求,优化请求包括第一AP的标识;AC减小第一AP的 信号覆盖范围或控制第一AP阻止下行信号强度小于下行信号强度阈值的终端接入第一AP或启动所述第一AP的延迟接入功能。如此可以实现对逻辑边缘AP的性能进行优化,减小逻辑边缘AP带来的影响。
第九方面,本申请提供了一种检测系统,所述系统包括:数据分析器CI和第一接入点AP,第一AP采集的至少一个终端的连接信息,向CI发送该连接信息,该连接信息包括终端的标识和采集该连接信息的采集时间,至少一个终端为接入到第一AP的终端;CI对于至少一个终端中的任一终端,获取任一终端的连接信息序列,该连接信息序列包括按采集时间排列的连接信息,该连接信息序列中的每个连接信息包括任一终端的标识,该连接信息序列中的相邻两个连接信息的采集时间之间的间隔不超过间隔阈值;根据该连接信息序列确定该任一终端是否为游牧终端。如此,由于能够从第一AP采集的连接信息中得到一个终端的连接信息序列,这样基于该连接信息序列可以成功地确定出游牧终端。
附图说明
图1是本申请实施例提供的一种WLAN的结构示意图;
图2是本申请实施例提供的一种系统架构示意图;
图3是本申请实施例提供了另一种系统架构示意图;
图4是本申请实施例提供了一种检测方法流程图;
图5是本申请实施例提供的一种决策树的结构示意图;
图6是本申请实施例提供的一种检测装置结构示意图;
图7是本申请实施例提供的另一种检测装置结构示意图;
图8是本申请实施例提供的另一种检测装置结构示意图。
具体实施方式
目前可以在建筑物等场所部署WLAN,以便位于该场所内的终端接入该WLAN。WLAN包括至少一个AP,位于WLAN的信号覆盖范围内的终端可以接入到该至少一个AP中的任一个AP,以实现接入WLAN。
对于像在食堂、办公区或咖啡厅等场所中部署的WLAN,在该WLAN中可能存在一类AP,该AP可能是逻辑边缘AP,逻辑边缘AP是信号覆盖范围抵达所在WLAN信号覆盖范围边缘的AP。在该AP是逻辑边缘AP时,在时长为第一时长的时间段内可能存在大量目标终端请求接入该AP。例如在时长为第一时长的时间段内请求接入该AP的目标终端数目超过第一数目阈值。目标终端包括接入状态异常的终端,例如目标终端可以包括没有成功接入该AP的终端或在WLAN中停留的时间未超过第一时间阈值的终端或不属于WLAN的白名单中的终端中的至少一种。
对于该AP,该AP的信号覆盖范围通常抵达WLAN的信号覆盖范围的边缘,该AP的信号覆盖范围可能还超出了该场所的范围,并可能覆盖到位于该场所外的人流通道。该AP可能是位于WLAN边缘的边缘AP,也可能不是位于该WLAN边缘的边缘AP。参见图1所示的在场所1内部署的WLAN,在该WLAN中AP1、AP3、AP4和AP5为位于WLAN边缘的边缘AP,而AP2不是位于WLAN边缘的边缘AP。AP2的信号覆盖范围抵达WLAN信号覆 盖范围的边缘,且超出了该场所1,覆盖了位于该场所1外的道路2上。AP4为位于WLAN的缘的边缘AP,且AP4的信号覆盖范围也超出了该场所1,并覆盖了位于该场所1外的另一条道路3上。所以AP2和AP4为属于这一类的AP。
由于该AP的信号覆盖范围内可能有人流通行的通道,这样行走在该通道上的用户就会从该AP覆盖的范围穿过。当用户在穿过该AP覆盖的范围时,该用户的终端会请求接入该AP,以接入WLAN。如果该用户的终端成功接入该AP,在接入该AP后该用户又可能很快离开该AP的信号覆盖范围,甚至离开WLAN的信号覆盖范围,该用户的终端就会断开与该AP的连接,导致该用户的终端在WLAN中停留的时间较短,停留的时间可能未超过第一时间阈值。其中,该终端在WLAN中停留的时间未超过第一时间阈值,且该终端在接入该AP之前的预设时长内以及在断开与该AP连接之后的预设时长内没有接入该WLAN的任一AP,该终端又被称为游牧终端。
WLAN中的每个AP中可能保存有WLAN的白名单,例如,假设该场所是一公司的办公区,该WLAN的白名单可以包括该公司的各员工的终端的标识。然而在通道上行走的人流当中有很大一部分的人员不是该公司的员工,这样导致请求接入该AP的终端中可能包括很多不属于该白名单的终端。
不属于该白名单的终端在请求接入该AP时,该AP可能会拒绝该终端接入,导致该终端没有成功接入到该AP。或者,通常该通道位于WLAN的信号覆盖范围的边缘,WLAN的信号强度较弱,又由于用户可能很快穿过该AP的信号覆盖范围,就可能导致该用户的终端没有成功接入到该AP。
所以对于该AP,该AP的信号覆盖范围抵达WLAN的信号覆盖范围的边缘,还有请求接入该AP的终端中可能包括大量在WLAN停留的时间未超过第一时间阈值的游牧终端、没有成功接入该AP的终端或不属于WLAN的白名单的终端,所以该AP为逻辑边缘AP。
逻辑边缘AP的存在可能会产生如下几方面的影响,分别为:
第一方面,逻辑边缘AP对游牧终端的有较大的影响。游牧终端在进入逻辑边缘AP的信号覆盖范围之前,可能使用优质的移动网络为游牧终端正在运行的应用提供上网服务,当进入到逻辑边缘AP的信号覆盖范围中,会自动通过接入逻辑边缘AP,接入到WLAN。由于游牧终端位于逻辑边缘AP的信号覆盖范围的边缘,在游牧终端的位置处逻辑边缘AP的信号较弱,在接入到逻辑边缘AP后可能降低了上网体验。在游牧终端离开逻辑边缘AP的信号覆盖范围,游牧终端不会重新使用移动网络为游牧终端正在运行的应用提供上网服务,中断了为该运行的应用提供的上网服务。
第二方面,逻辑边缘AP对接入逻辑边缘AP的常驻终端或接入到属于WLAN的白名单中的终端有较大的影响,常驻终端是指接入逻辑边缘AP后,在逻辑边缘AP中停留的时间超过第一时间阈值的终端。由于可能有大量游牧终端或不属于该白名单的终端接入到逻辑边缘AP,消耗了大量逻辑边缘AP的网络资源,从而影响了位于逻辑边缘AP中的常驻终端或已接入的属于该白名单的终端使用的网络资源。
第三方面,逻辑边缘AP对WLAN也会产生较大的影响,在有大量的游牧终端请求接入逻辑边缘AP时会产生大量控制面消息,或在有大量的游牧终端离开逻辑边缘AP的信号覆盖范围时该大量的游牧终端断开与逻辑边缘AP的连接,也会产生大量的控制面消息,都会增加WLAN的控制面的网络负担,可能会引起网络故障。
为了消除逻辑边缘AP带来的影响,可以从WLAN的AP中检测出逻辑边缘AP,然后对逻辑边缘AP进行优化,以减小或消除逻辑边缘AP产生上述影响。
参见图2,本申请实施例提供了一种网络架构,包括:
数据分析器(campus insight,CI)和至少一个WLAN中的AP,CI可以与每个WLAN中的AP之间建立有网络连接。
对于该至少一个WLAN中的AP中的任一AP,为了便于说明该AP为第一AP,CI可以获取第一AP的至少一个特征;根据该至少一个特征检测第一AP是否为逻辑边缘AP。
CI获取该至少一个特征的详细过程以及检测第一AP是否为逻辑边缘AP的详细过程,将在后续图4所示的实施例进行详细说明,在此先不介绍。
可选的,在该至少一个特征满足第一条件的情况下,确定第一AP为逻辑边缘AP。
可选的,该至少一个特征包括目标终端数目,该目标终端数目可以用于反应接入第一AP的终端的接入情况。第一条件包括该目标终端数目大于第一数目阈值。
可选的,该至少一个特征还可以包括至少一种连接事件中的每种连接事件的总数目,每种连接事件的总数目用于反应第一AP的性能情况。该至少一种连接事件是接入到第一AP的至少一个终端的连接事件。对于该至少一种连接事件中的任一种连接事件,第一条件还可以包括该任一种连接事件的总数目大于该任一种连接事件对应的数目阈值,或者,第一条件还可以包括该任一种连接事件的总数目小于该任一种连接事件对应的数目阈值。
可选的,该至少一个特征还可以包括信噪比统计值、丢包率统计值、重传率统计值或信道利用率统计值等中的至少一个。其中信噪比统计值、丢包率统计值、重传率统计值或信道利用率统计值中的任一者可用于反应接入第一AP的终端的接入情况。第一条件还可以包括该信噪比统计值小于信噪比统计阈值、该丢包率统计值大于丢包率统计阈值、该重传率统计值大于重传率统计阈值或该信道利用率统计值大于信道利用率阈值中的至少一个。
可选的,第一AP可提供至少一个频段供终端接入,对于该至少一个频段中的任一个频段,该至少一个特征还可以包括接入到该频段的总终端数目。该总终端数目用于反应第一AP的负载情况。第一条件还可以包括该总终端数目大于终端数目阈值的条件。
可选的,CI可以是计算机、服务器或服务器集群等。
可选的,参见图3,该网络架构还可以包括AC,AC可以与CI之间建立有网络连接,以及AC可以与该至少一个WLAN中的AP之间建立有网络连接。
AC用于对该至少一个WLAN中的任一个AP的性能进行优化。
可选的,在CI检测出第一AP为逻辑边缘AP时,CI可以向AC发送优化请求,该优化请求包括第一AP的标识。AC接收该优化请求,根据该优化请求包括的第一AP的标识对第一AP的性能进行优化。
AC对第一AP的性能进行优化的详细实现过程,将在后续图4所示的实施例中进行详细说明,在此先不详细介绍。
参见图4,本申请实施例提供了一种检测方法,该检测方法用于检测WLAN中的AP是否为逻辑边缘AP,该方法可以应用于图2或图3所示的网络架构,包括:
步骤201:CI接收第一AP采集的至少一个终端的连接信息,对于该至少一个终端中的 任一终端,该任一终端的连接信息包括该任一终端的标识和第一AP采集该连接信息的采集时间。
CI可以用于检测至少一个WLAN中的AP,即CI用于检测一个WLAN中的AP或检测多个WLAN中的AP,第一AP是该至少一个WLAN中的任一AP。该至少一个终端是接入到第一AP中的终端。
CI检测的该至少一个WLAN中的AP可以是管理员配置的。例如,对于该至少一个WLAN中的每个WLAN,如果该WLAN的管理员需要CI检测该WLAN中的AP,该管理员可以向CI输入该WLAN中的每个AP的标识和该WLAN的标识。
CI可以将该WLAN中的每个AP的标识和该WLAN的标识对应保存到AP的标识与WLAN的标识的对应关系中,以及,对于该WLAN中的每个AP,CI根据该AP的标识分别向该AP发送采集请求,以触发该AP采集接入到该AP的终端的连接信息。
对于该至少一个终端中的任一终端,第一AP可以在不同时间采集该任一终端的连接信息,向CI发送该采集的连接信息,第一AP对该任一终端的连接信息进行相邻两次采集的时间间隔不超过间隔阈值。
可选的,第一AP可以周期性地采集该任一终端的连接信息,第一终端采集连接信息的周期长度小于或等于该间隔阈值。
可选的,该连接信息还可以包括第一AP的标识,第一AP采集该连接信息的采集时间、该任一终端接入第一AP的起始时间、该任一终端在第一AP已停留的停留时间、该任一终端接入的频段的标识、该任一终端的上行信号强度、数据传输时延、信噪比、丢包率、重传率或信道利用率等信息中的至少一个。
第一AP可以提供至少一个频段,接入第一AP的终端可以从该至少一个频段中选择一个频段,接入该选择的频段,然后该终端可以使用接入的频段与第一AP进行通信。
对于该至少一个WLAN中的任一其他AP,该其他AP也同第一AP一样采集接入到该AP的终端的连接信息并向CI发送采集的终端的连接信息。也就是说,CI会接收到不同的AP发送的终端的连接信息,且CI保存不同AP发送的连接信息。
在进入第一AP信号覆盖范围的终端,该终端可能会向第一AP发送接入请求,以请求接入第一AP。该终端请求接入第一AP后,可能会成功接入到第一AP,也可能没有成功接入第一AP。
在该终端没有成功接入第一AP时,第一AP会检测到该终端并产生接入失败事件,该接入失败事件可以包括第一AP的标识和该终端请求接入的起始时间等信息,向CI发送该接入失败事件。
对于该至少一个WLAN中的任一其他AP,该其他AP也同第一AP一样在检测到没有成功接入该其他AP的终端时产生接入失败事件并向CI发送产生的接入失败事件。也就是说,CI会接收到不同的AP发送的接入失败事件,且CI保存不同AP发送的接入失败事件。
可选的,第一AP的标识可以是第一AP的地址,例如可以是第一AP的网络之间互连的协议(internet protocol,IP)地址或媒体访问控制地址(media access control address,MAC)地址。
可选的,终端的标识可以是终端的地址,例如可以是终端的IP地址或MAC地址。
CI可以周期性地检测该至少一个WLAN中的AP,以检测出逻辑边缘AP。例如,CI可 以在每天的预设时间点开始检测该至少一个WLAN中的AP。
可选的,CI可以将一个周期分成至少一个时间段,每个时间段的时长等于第一时长。对于该至少一个时间段中的任一个时间段,根据在该时间段内AP采集的连接信息,检测该WLAN中的AP。CI的检测过程如下:
步骤202:CI根据在该时间段内第一AP采集的至少一个终端的连接信息,获取第一终端的目标终端数目,该目标终端数目是第一AP的一个特征,该特征用于反应接入第一AP的终端的接入情况。
对于CI中保存的任一个连接信息,该连接信息中包括采集该连接信息的AP的标识和采集时间。对于第一AP,CI可以从保存的连接信息中获取包括第一AP的标识且包括的采集时间位于该时间段内的连接信息,获取的连接信息为在该时间段内第一AP采集的连接信息,然后CI执行本步骤。
在本步骤中,可以根据在该时间段内第一AP采集的该至少一个终端的连接信息,从该至少一个终端中确定出目标终端,统计目标终端数目。此时确定出的目标终端包括游牧终端或不属于目标白名单的终端中的至少一种,目标白名单是目标WLAN的白名单,目标WLAN是第一AP所属的WLAN,该目标终端数目可以包括游牧终端数目或不属于目标白名单的终端数目中的至少一个。
第一AP中可以保存有目标白名单,或者,用于管理目标WLAN的AC中可以保存有目标白名单。CI中也可以保存有目标白名单,且CI中保存的目标白名单可以是在执行本步骤之前CI从属于目标WLAN中的AP中获取的、从该AC中获取的或者接收目标WLAN的管理员发送的。
例如,假设部署目标WLAN的场所是一个公司,目标WLAN的目标白名单包括该公司的各员工的终端的标识,目标WLAN的管理员可以在AC中保存目标白名单,管理员还可以在目标WLAN中的每个AP中保存目标白名单。CI在可以从目标WLAN的AP中获取目标白名单,或接收该管理员输入的目标白名单。
可选的,在CI用于检测一个WLAN中的AP时,CI中可能保存一个白名单,该白名单为该WLAN的白名单。在CI用于检测多个WLAN中的AP时,CI在得到任一个WLAN的白名单时,可以将该WLAN中的每个AP的标识和该WLAN的白名单对应保存在AP的标识与白名单的对应关系中。
可选的,可以通过如下方式从该至少一个终端中确定出不属于目标白名单的终端。该方式可以为:
从该至少一个终端的连接信息中获取该至少一个终端中的每个终端的标识和第一AP的标识。在CI中保存一个白名单时,将该白名单作为第一AP所属的目标WLAN的目标白名单;或者,在CI中保存多个白名单时,根据第一AP的标识,从AP的标识与白名单的对应关系获取第一AP所属的WLAN的目标白名单。根据每个终端的标识和目标白名单,从每个终端中确定出不属于目标白名单的终端,统计不属于目标白名单的终端数目。
可选的,可以通过如下方式从该至少一个终端中确定出游牧终端。该方式可以为:
对于该任一个终端中的任一个终端,在该时间段内第一AP采集的该至少一个终端的连接信息中获取该任一终端的连接信息序列,该连接信息序列包括按采集时间排列的连接信息,该连接信息序列中的每个连接信息包括该任一终端的标识,该连接信息序列中的相邻两个连 接信息的采集时间之间的间隔不超过间隔阈值;根据连接信息序列确定该任一终端是否为游牧终端。重复上述过程,从而得到该至少一个终端中的每个终端的连接信息序列,以及从该至少一个终端中确定出所有的游牧终端,统计确定出的游牧终端数目。
可选的,在本步骤中,可以通过如下2021至2022的操作,实现根据连接信息序列确定该任一终端是否为游牧终端。该2021至2022的操作可以为:
2021:根据该任一终端的连接信息序列获取该任一终端接入第一AP的起始时间、该任一终端在第一AP中的停留时间和该任一终端断开与第一AP之间连接的断开时间。
可选的,可以将该连接信息序列中的第一条连接信息中保存的采集时间作为该任一终端接入第一AP的起始时间,将该连接信息序列中的最后一条连接信息中保存的采集时间作为该任一终端断开与第一AP之间连接的断开时间,根据该断开时间和该起始时间计算该任一终端在第一AP中的停留时间。或者,
可选的,在第一AP采集的该任一终端的每条连接信息中包括该任一终端接入第一AP的起始时间和该任一终端在第一AP已停留的停留时间的情况下,可以从该连接信息序列的最后一条连接信息中获取该任一终端接入第一AP的起始时间和该任一终端在第一AP的停留时间,以及将该最后一条连接信息中的采集时间作为该任一终端断开与第一AP之间连接的断开时间。
2022:在该停留时间未超第一时间阈值,在第一时间段和第二时间段内没有第二AP采集的该任一终端的连接信息,确定该任一终端为游牧终端。
第一时间段是第一时间到该起始时间的时间段,第二时间段是该断开时间到第二时间的时间段,第一时间位于该起始时间之前,第二时间位于该断开时间之后,第一时间段的时间长度等于预设时长,第二时间段的时间长度等于预设时长,预设时长均大于该间隔阈值,第二AP为目标WLAN中的除第一AP之外的其他AP。
在第一时间段没有第二AP采集的该任一终端的连接信息,表示该任一终端不是从目标WLAN的其他AP漫游到第一AP中。在第二时间段没有第二AP采集的该任一终端的连接信息,表示该任一终端没有从第一AP漫游到目标WLAN的其他AP中。
在本步骤中,可以根据第一AP的标识,以及AP的标识与WLAN的标识的对应关系,获取目标WLAN中的其他各AP的标识,从CI保存的连接信息中获取其他各AP采集的连接信息。从其他各AP采集的连接信息中检测是否存在包括的采集时间位于第一时间段和第二时间段内且包括该任一终端的标识的连接信息,如果不存在,则确定在第一时间段和第二时间段内没有第二AP采集的该任一终端的连接信息,如果存在,则确定在第一时间段和第二时间段内有第二AP采集的该任一终端的连接信息。
可选的,在本步骤中,还可以通过其他方式,实现根据连接信息序列确定该任一终端是否为游牧终端。该其他方式为,
将该任一终端的连接信息序列作为终端检测模型的输入,并通过终端检测模型检测该任一终端是否为游牧终端。
可选的,终端检测模型是通过第一训练集对第一人工智能(artificial intelligence,AI)模型进行训练得到的,第一训练集包括多个训练序列,每个训练序列是一个游牧终端的连接信息序列或是一个非游牧终端的连接信息序列。
第一训练集中的训练序列包括两类,其中一类的训练序列的类别为正样本,正样本的训 练序列为游牧终端的连接信息序列。另一类的训练序列为负样本,负样本的训练序列为非游牧连接信息序列。
可选的,可以先获取到终端的连接信息序列,将该连接信息序列作为训练序列,然后采用上述2021至2022的操作确定该连接信息序列是否为游牧终端的连接信息序列,如果是,设置该训练序列的类别为正样本,如果不是,设置该训练序列的类别为负样本。这样在得到大量的训练序列后,使用训练序列训练第一AI模型,以训练出终端检测模型,之后在检测连接信息序列时,就可以不用通过上述2021至2022的操作来检测该连接信息序列,而时直接使用该终端检测模型来检测该连接信息序列。
可选的,在本步骤中可以通过如下2121至2124的操作来训练终端检测模型。该训练过程为:
2121:将第一训练集中的训练序列输入到第一AI模型中,每个训练序列的类别为正样本或负样本。
为了便于说明,将每个训练序列的类别分别称为每个训练序列的真实类别。
可选的,本步骤中采用的第一AI模型可以为长短期记忆网络(long short-term memory,lstm)。
2122:第一AI模型预测第一训练集中的每个训练序列的类别。
对于第一训练集中的每个训练序列,第一AI模型从该训练序列中提取特征,基于提取的特征预测该训练序列为正样本的第一概率以及为负样本的第二概率,第一概率和第二概率之和等于1,输出两个概率中数值较大的概率对应的类别。
2123:对于每个训练序列,第一AI模型根据该训练序列的真实类别(正样本或者负样本)和第一AI模型对该训练序列预测后输出的类别,利用损失函数计算该训练序列的损失函数值。进一步,根据每个训练序列的损失函数值调整第一AI模型的网络参数。
2124:第一AI模块确定是否继续训练,在确定继续训练时,返回执行2122,在确定停止训练时,此时第一AI模型为终端检测模型,结束返回。
可以根据损失函数值确定是否继续训练,具体地,在训练过程中,每次训练后得到每个训练序列的损失函数值,将得到的损失函数值进行曲线拟合,若得到的曲线逐渐收敛,且最后一次训练后的损失函数值小于预设定的阈值,则确定停止训练,否则,确定继续训练。
可选的,目标终端还可以包括没有成功接入第一AP的终端。目标终端数目还可以包括没有成功接入第一AP的终端数目。即在本步骤中,CI还可以从CI保存的接入失败事件中获取包括第一AP的标识且包括的起始时间位于该时间段内的接入失败事件,统计获取的接入失败事件数目,得到没有成功接入第一AP的终端数目。
对于目标WLAN中的其他每个AP,对其他每个AP执行本步骤的操作,得到其他每个AP的目标终端数目,即在本步骤可以得到目标WLAN中的每个AP的目标终端数据,对于任一AP的目标终端数目可以包括游牧终端数目、不属于目标白名单的终端数目或没有成功接入该AP的终端数目中至少一个。
可选的,第一AP也可以将采集的连接信息发送除CI之外的其他设备,例如发送给AC。这样AC可以获取任一终端的连接信息序列和确定该任一终端的连接信息序列是否为游牧终端的连接信息序列,如果是,标记该任一终端的连接信息序列,向CI发送该任一终端的连接信息序列,如果不是,直接向CI发送该任一终端的连接信息序列。这样CI接收AC发送该 任一终端的连接信息序列,在该连接信息序列被标记时,确定该任一终端为游牧终端。这样游牧终端的检测流程从CI中分离,由其他设备来实现,可以减轻CI的运算压力。
步骤203:可选的,CI获取该至少一终端的至少一种连接事件,对于该至少一种连接事件中的任一种连接事件,统计该任一种连接事件的总数目,该任一种连接事件的总数目是第一AP的一个特征,该特征用于反应第一AP的性能情况。
本步骤是一个可选的步骤,即可以不执行本步骤,在执行完步骤202后就执行步骤204,当然也可以执行本步骤,在执行完本步骤后,再执行步骤204。
可选的,对于该至少一个终端中的任一终端,第一AP采集的该任一终端的连接信息包括该任一终端中的上行信号强度或该任一终端的数据传输时延中的至少一个。在步骤202中,CI已获取到该任一终端的连接信息序列。
可选的,该任一终端的至少一种连接事件可以包括至少一种弱覆盖事件、至少一种高时延事件、至少一种强覆盖事件或至少一种低时延事件等中至少一种。
可选的,CI中可以保存有第一强度阈值与弱覆盖事件的对应关系,第一强度阈值与弱覆盖事件的对应关系中保存有至少一个第一强度阈值和每个第一强度阈值对应的弱覆盖事件,第一强度阈值与弱覆盖事件的对应关系中保存的第一强度阈值往往较小。例如,参见下表1所示的第一强度阈值与弱覆盖事件的对应关系。下表1所示的第一弱覆盖事件、第二弱覆盖事件和第三弱覆盖事件分别为不同种类的连接事件。
表1
第一强度阈值 弱覆盖事件
10db 第一弱覆盖事件
12db 第二弱覆盖事件
7db 第三弱覆盖事件
…… ……
可选的,CI中可以保存有第一时延阈值与高时延事件的对应关系,第一时延阈值与高时延事件的对应关系中保存有至少一个第一时延阈值和每个第一时延阈值对应的高时延事件,第一时延阈值与高时延事件的对应关系中保存的第一时延阈值往往较大。例如,参见下表2所示的第一时延阈值与高时延事件的对应关系。下表2所示的第一高时延事件、第二高时延事件和第三高时延事件分别为不同种类的连接事件。
表2
第一时延阈值 弱覆盖事件
3 第一高时延事件
5 第二高时延事件
7 第三高时延事件
…… ……
可选的,CI中可以保存有第二强度阈值与强覆盖事件的对应关系,第二强度阈值与强覆盖事件的对应关系中保存有至少一个第二强度阈值和每个第二强度阈值对应的强覆盖事件,第二强度阈值与强覆盖事件的对应关系中保存的第二强度阈值往往较大。例如,参见下表3 所示的第二强度阈值与强覆盖事件的对应关系。下表3所示的第一强覆盖事件、第二强覆盖事件和第三强覆盖事件分别为不同种类的连接事件。
表3
第二强度阈值 强覆盖事件
50db 第一强覆盖事件
52db 第二强覆盖事件
67db 第三强覆盖事件
…… ……
可选的,CI中可以保存有第二时延阈值与低时延事件的对应关系,第二时延阈值与低时延事件的对应关系中保存有至少一个第二时延阈值和每个第二时延阈值对应的低时延事件,第二时延阈值与低时延事件的对应关系中保存的第二时延阈值往往较小。例如,参见下表4所示的第二时延阈值与低时延事件的对应关系。下表4所示的第一低时延事件、第二低时延事件和第三低时延事件分别为不同种类的连接事件。
表4
第二时延阈值 弱覆盖事件
2 第一低时延事件
1.5 第二低时延事件
1 第三低时延事件
…… ……
在本步骤中,可以通过如下2031至2032的操作来实现,该2031至2032的操作分别为:
2031:根据该任一终端的连接信息序列中的最后一条连接信息包括的上行信号强度或数据传输时延中的至少一个,获取任一终端的至少一种连接事件。
可选的,在该任一终端的连接信息序列中的最后一条连接信息包括的上行信号强度的情况下,从第一强度阈值与弱覆盖事件的对应关系中保存的各第一强度阈值中,选择大于该上行信号强度的第一强度阈值。根据选择的每个第一强度阈值,从第一强度阈值与弱覆盖事件的对应关系中获取对应的弱覆盖事件,将获取的弱覆盖事件作为该任一终端的弱覆盖事件。或者,从第二强度阈值与强覆盖事件的对应关系中保存的各第二强度阈值中,选择小于该上行信号强度的第二强度阈值。根据选择的每个第二强度阈值,从第二强度阈值与强覆盖事件的对应关系中获取对应的强覆盖事件,将获取的强覆盖事件作为该任一终端的强覆盖事件
例如,该任一终端的连接信息序列中的最后一条连接信息包括的上行信号强度为8,从如表1所示的第一强度阈值与弱覆盖事件的对应关系中获取大于8的第一强度阈值分别为10和12,根据第一强度阈值10和12,从如表1所示的第一强度阈值与弱覆盖事件的对应关系中获取第一强度阈值10对应的第一弱覆盖事件,以及第二强度阈值12对应的第二弱覆盖事件,即得到了该任一终端的第一弱覆盖事件和第二弱覆盖事件。
其中,该任一终端的弱覆盖事件用于表示该任一终端可能位于第一AP的边缘,在该任一终端距离第一AP的位置较远,第一AP在该任一终端的位置处覆盖的信号可能较弱。该任一终端的强覆盖事件用于表示该任一终端可能距离第一AP低近,第一AP在该任一终端的位 置处覆盖的信号可能较强。
可选的,在该任一终端的连接信息序列中的最后一条连接信息包括的数据传输时延的情况下,从第一时延阈值与高时延事件的对应关系中保存的各第一时延阈值中,选择小于该数据传输时延的第一时延阈值。根据选择的每个第一时延阈值,从第一时延阈值与高时延事件的对应关系中获取对应的高时延事件,将获取的高时延事件作为该任一终端的高时延事件。或者,从第二时延阈值与低时延事件的对应关系中保存的各第二时延阈值中,选择大于该数据传输时延的第二时延阈值。根据选择的每个第二时延阈值,从第二时延阈值与低时延事件的对应关系中获取对应的低时延事件,将获取的低时延事件作为该任一终端的低时延事件。
例如,该任一终端的连接信息序列中的最后一条连接信息包括的数据传输时延为6,从如表2所示的第一时延阈值与高时延事件的对应关系中获取小于6的时延阈值分别为3和5,根据时延阈值3和5,从如表2所示的第一时延阈值与高时延事件的对应关系中获取时延阈值3对应的第一高时延事件,以及时延阈值5对应的第二高时延事件,即该任一终端的第一高时延事件和第二高时延事件。
其中,该任一终端的高时延事件用于表示该任一终端可能位于第一AP的边缘,在该任一终端距离第一AP的位置较远,导致第一AP在该任一终端的位置处覆盖的信号可能较弱,进而导致该任一终端发送给第一AP的数据所需要的传输时延较大。该任一终端的低时延事件用于表示该任一终端可能距离第一AP较近,第一AP在该任一终端的位置处覆盖的信号可能较强,进而导致该任一终端发送给第一AP的数据所需要的传输时延较小。
2032:对于该至少一个终端的至少一个连接事件中的任一种连接事件,统计该任一种连接事件的总数目。
可选的,该任一终端的连接信息还包括该任一终端的信噪比、丢包率、重传率、信道利用率和该任一终端接入的频段标识等信息中的至少一个;
可选的,CI还可以根据在该时间段内第一AP采集的该至少一个终端的连接信息,获取信噪比统计值、丢包率统计值、重传率统计值、信道利用率统计值或接入该频段标识对应的频段的总终端数目等中的至少一个。
其中,信噪比统计值、丢包率统计值、重传率统计值或信道利用率统计值中的任一者是第一AP的特征,该特征也用于反应接入第一AP的终端的接入情况。该总终端数目也是第一AP的特征,该特征用于反应第一AP的负载情况。
可选的,对于上述信噪比统计值,可以根据该至少一个终端中的每个终端的信噪比计算信噪比平均值,将该信噪比平均值作为信噪比统计值;或者,对该至少一个终端中的每个终端的信噪比进行排序,将排在中间位置的信噪比作为信噪比统计值;或者,从该至少一个终端中的每个终端的信噪比中选择最大信噪比或最小信噪比作为信噪比统计值。
可选的,对于上述丢包率统计值,可以根据该至少一个终端中的每个终端的丢包率计算丢包率平均值,将该丢包率平均值作为丢包率统计值;或者,对该至少一个终端中的每个终端的丢包率进行排序,将排在中间位置的丢包率作为丢包率统计值;或者,从该至少一个终端中的每个终端的丢包率中选择最大丢包率或最小丢包率作为丢包率统计值。
可选的,对于上述重传率统计值,可以根据该至少一个终端中的每个终端的重传率计算重传率平均值,将该重传率平均值作为重传率统计值;或者,对该至少一个终端中的每个终端的重传率进行排序,将排在中间位置的重传率作为重传率统计值;或者,从该至少一个终 端中的每个终端的重传率中选择最大重传率或最小重传率作为重传率统计值。
可选的,对于上述信道利用率统计值,可以根据该至少一个终端中的每个终端的信道利用率计算信道利用率平均值,将该信道利用率平均值作为信道利用率统计值;或者,对该至少一个终端中的每个终端的信道利用率进行排序,将排在中间位置的信道利用率作为信道利用率统计值;或者,从该至少一个终端中的每个终端的信道利用率中选择最大信道利用率或最小信道利用率作为信道利用率统计值。
可选的,如果在执行步骤202之前,已确定需要第一AP的各特征的类别,则在步骤202和203中只需要获取属于确定的类别的特征。例如,假设确定的类别包括游牧终端数目、第一高时延事件数目、第一弱覆盖事件数目。则在步骤202中,CI可以获取第一终端的游牧终端数目。在步骤203中获取该任一终端的连接事件时,对于接入到第一AP的任一终端,CI从该任一终端的连接信息序列的最后一个连接信息中读取上行信号强度和数据传输时间;从强度阈值与弱覆盖事件的对应关系中获取第一弱覆盖事件对应的强度阈值,在该上行信号强度小于获取的强度阈值时,确定该任一终端的连接事件中包括该第一弱覆盖事件;从时延阈值与高时延事件的对应关系中获取第一高时延事件对应的时延阈值,在该数据传输时延大于获取的时延阈值时,确定该任一终端的连接事件中包括该第一高时延事件;然后CI再统计第一弱覆盖事件的总数目和第一高时延事件的总数目。
步骤204:CI判断第一AP的特征是否满足第一条件,在满足第一条件时检测第一AP为逻辑边缘AP。
在本步骤中,CI可以采用如下三种方式检测第一AP是否为逻辑边缘AP。该三种方式分别为:
第一种方式,CI根据第一AP的目标终端数目和第一条件检测第一AP为逻辑边缘AP。第一条件包括目标终端数目大于第一数目阈值,或者,第一AP是目标WLAN中的目标终端数目最大的第一个数个AP中的一个,或者,目标终端数目与目标终端平均数目的差值大于差值阈值,该差值阈值等于方差值的m倍,m大于1,目标终端平均数目和该方差值是基于目标WLAN中的各AP的目标终端数目得到的。
目标终端数目与目标终端平均数目的差值等于目标终端数目减去目标终端平均数目。
当采用第一种方式时,可以不执行上述步骤203的操作,即在执行完步骤202之后,直接执行本步骤。
在第一种方式中,CI可以判断第一AP的目标终端数目是否超过第一数目阈值,如果超过,则检测出第一AP为逻辑边缘AP。或者,CI从目标WLAN中选择目标终端数目最大的预设个数个AP作为逻辑边缘AP。或者,CI根据目标WLAN中的各AP的目标终端数目计算目标终端平均数目和方差,将第一AP的目标终端数目减去目标终端平均数目,得到差值,在该差值大于该方差值的m倍时,检测出第一AP为逻辑边缘AP。
该第一数目阈值可以是CI预设的值或是目标终端平均数目的y倍,y为大于1的数值。
在第一种方式中,目标终端数目可以为游牧终端数目。
第二种方式,CI根据第一AP的每个特征是否满足第一条件,在满足第一条件时,确定第一AP为逻辑边缘AP。
第一条件定义了第一AP的任一个特征与该特征的类别对应的特征阈值之间的判断条件。该特征与该特征阈值之间的判断条件表示一种大小关系,该大小关系可以大于关系或小于关 系。如果第一AP是逻辑边缘AP,则该特征与该特征阈值之间的大小关系和该判断条件表示的大小关系相符。
在第二种方式中,将第一AP的每个特征分别与每个特征的类别对应的特征阈值进行比较,得到每个特征的比较结果;根据每个特征与该特征的类别对应的特征阈值之间的判断条件,以及每个特征的比较结果检测第一AP是否为逻辑边缘AP。
即在第二种方式中,对于任一个特征,该特征与该特征的类别对应的特征阈值进行比较,得到该特征的比较结果,该比较结果是该特征与该特征阈值之间的大小关系,判断该比较结果与该类别对应的判断条件表示的大小关系相符,表明该特征是第一AP符合逻辑边缘AP的一个特征。按上述方式,得出第一AP的每个特征均是第一AP符合逻辑边缘AP的特征,则确定第一AP为逻辑边缘AP。
可选的,对于第一AP的任一个特征,第一条件包括该任一个特征大于该任一个特征的类别对应的特征阈值,或者,该任一个特征小于该任一个特征的类别对应的特征阈值。具体的,第一条件包括第一AP的目标终端数目大于第一数目阈值。且第一条件除了包括该条件外,还可以包括CI统计的任一种连接事件的总数目大于任一种连接事件对应的数目阈值,或,该任一种连接事件的总数目小于该任一种连接事件对应的数目阈值。或者,第一条件还可以包括信噪比统计值小于信噪比统计阈值的条件、丢包率统计值大于丢包率统计阈值的条件、重传率统计值大于重传率统计阈值的条件、信道利用率统计值大于信道利用率阈值的条件或接入到第一AP的每个频段的总终端数目分别大于每个频段对应的终端数目阈值的条件中的至少一个。
可选的,该任一种连接事件可能是一种弱覆盖事件或是一种高时延事件,第一条件包括该弱覆盖事件大于该弱覆盖事件对应的数目阈值,或者,该高时延事件大于该高时延事件对应的数目阈值。
可选的,该任一种连接事件可能是一种强覆盖事件或是一种低时延事件,第一条件包括该强覆盖事件大于该强覆盖事件对应的数目阈值,或者,该低时延事件大于该低时延事件对应的数目阈值。
例如,假设第一AP提供第一频段和第二频段供终端接入,第一AP的特征包括目标终端数目50、第一高时延事件数目40、第二高时延事件数目35、第一弱覆盖事件数目29、第二弱覆盖事件数目36、信噪比统计值为58、丢包率统计值为0.35、重传率统计值为0.48、信道利用率统计值为0.42、接入第一频段的总终端数目为48以及接入第二频段的总终端数目为56。
假设,第一数目阈值为30,第一高时延事件对应的数目阈值为20、第二高时延事件对应的数目阈值为22、第一弱覆盖事件对应的数目阈值为18、第二弱覆盖事件对应的数目阈值为19、信噪比统计阈值为60、丢包率统计阈值为0.3、重传率统计阈值为0.4、信道利用率统计阈值为0.35、第一频段对应的终端数目阈值为30以及第二频段对应的终端数目阈值为25。
其中第一AP的目标终端数目50大于第一数目阈值30,第一高时延事件数目40大于第一高时延事件对应的数目阈值15,第二高时延事件数目35大于第二高时延事件对应的数目阈值22,第一弱覆盖事件数目29大于第一弱覆盖事件对应的数目阈值18,第二弱覆盖事件数目36大于第二弱覆盖事件对应的数目阈值19,信噪比统计值58小于信噪比统计阈值60,丢包率统计值0.35大于丢包率统计阈值0.3,重传率统计值0.48大于重传率真统计阈值0.4, 信道利用率统计值0.42大于信道利用率统计阈值0.35,接入第一频段的总终端数目48大第一频段对应的终端数目阈值30以及接入第二频段的总终端数目56大于第二频段对应的终端数目阈值25。因此可以得出第一AP的特征满足第一条件,从而确定第一AP为逻辑边缘AP。
第二种方式相比较第一种方式,除了使用目标终端数目,结合第一AP的其他特征来检测第一AP是否为逻辑边缘AP,所以相比第一种方式,第二种方式可以提高检测的精度。
在第二种方式中,可以执行上述步骤202之前可以先训练出随机森林模型,训练出的随机森林模型具有检测AP是否为逻辑边缘AP的功能,使用随机森林模型确定需要获取的第一AP的各特征对应的类别和特征阈值。这样在上述步骤202和203中只需要获取第一AP的属于确定出的类别的特征,然后在本步骤中使用第二种方式检测第一AP是否为逻辑边缘AP。
在训练随机森林模型之前,先构造第二训练集,第二训练集包括多个训练样本和每个训练样本对应的类别,每个训练样本为一个AP的至少一个特征,在该AP是逻辑边缘AP时该训练样本的类别为正样本,在该AP是非逻辑边缘AP时该训练样本的类别为负样本。
在本步骤中,对于WLAN中已知的逻辑边缘AP,通过上述201至203的步骤得到该逻辑边缘AP的至少一个特征,将该至少一个特征作为一个训练样本并将该训练样本的类别设置为正样本。以及,对于WLAN中已知的非逻辑边缘AP,通过上述201至203的步骤得到该非逻辑边缘AP的至少一个特征,将该至少一个特征作为一个训练样本并将该训练样本的类别设置为负样本。
需要说明的是:构造的第二训练集中的正样本的训练样本可能较少,例如第二训练集中的正样本数目少于第三数目阈值。这主要是由于逻辑边缘AP通常较少,在任一个WLAN中,该WLAN中可能有逻辑边缘AP,也可能没有逻辑边缘AP。在该WLAN有逻辑边缘AP的情况下,该WLAN中的逻辑边缘AP的数目也较少,通常只有几个。所以技术人员可能没有得到足够多的已知逻辑边缘AP来形成正样本的训练样本,导致构造的第二训练集中的正样本的训练样本数目往往较少。当然技术人员也可以对大量的WLAN进行分析得到较多的逻辑边缘AP,使得构造的第二训练集中有较多的正样本的训练样本,此时可以使用第二训练集训练除随机森林之外的其他AI模块得到一个用于检测逻辑边缘AP的智能模型,该实现方式将在后续第三种方式进行详细说明。
可选的,可以通过如下2041至2044的操作,来训练随机森林模型。该2041至2044的操作分别为:
2041:将第二训练集中的训练样本输入到随机森林模型中,每个训练样本的类别为正样本或负样本。
为了便于说明,将每个训练样本的类别分别称为每个训练样本的真实类别。
2042:随机森林模型预测第二训练集中的每个训练样本的类别。
对于第二训练集中的每个训练样本,随机森林模型从该训练样本中提取特征,基于提取的特征预测该训练样本为正样本的第一概率以及预测该训练样本为负样本的第二概率,第一概率和第二概率之和等于1,输出两个概率中数值较大的概率对应的类别。
2043:对于每个训练样本,该随机森林模型根据该训练样本的真实类别和随机森林模型对该训练样本预测后输出的类别,利用损失函数计算该训练样本的损失函数值。进一步,根据第二训练集所有训练样本的损失函数值调整该随机森林模型的参数。
2044:随机森林模型确定是否继续训练,在确定继续训练时,返回执行2042,在确定停 止训练时。
可以根据损失函数值确定是否继续训练,具体地,在训练过程中,每次训练后得到每个训练样本的损失函数值,将得到的损失函数值进行曲线拟合,若得到的曲线逐渐收敛,且最后一次训练后的损失函数值小于预设定的阈值,则确定停止训练,否则,确定继续训练。
训练后得到的随机森林模型包括至少一个决策树,该决策树中的每条路径用于检测第二训练集中的一个AP是否为逻辑边缘AP。该路径的叶子节点用于保存该路径的检测结果,该路径中除叶子节点之外的节点对应一个类别和特征阈值,该节点用于判断属于该类别的特征是否超过该特征阈值以及根据判断结果选择属于该路径的下一层节点。
例如,参见图5所示的是随机森林模型的一个决策树,对任一个决策树,该决策树中除叶子节点之外的其他每个节点,该节点保存一个特征的类别和特征阈值。例如,参见图5所示的一个决策树,根节点1保存的类别为游牧终端数目以及保存特征阈值50,根节点1用于判断游牧终端数目是否超过特征阈值50,然后根据判断的结果选择路径的下一层节点,该下一层节点为节点2或节点3。节点2保存的类别为第一高时延事件数目以及保存特征阈值为30,节点2用于判断第一高时延事件数目是否超过特征阈值30,然后根据判断的结果选择路径的下一层节点,该下一层节点为节点4或节点5。节点4保存的类别为第一弱覆盖事件数目以及保存特征阈值为20,节点4用于判断第一弱覆盖事件数目是否超过特征阈值20,然后根据判断的结果选择路径的下一层节点,该下一层节点为叶子节点6或叶子节点7,叶子节点6或叶子节点7用于保存对AP的检测结果,该检测结果可以是逻辑边缘AP,或非逻辑边缘AP。
需要说明的是:在第二训练集中的正样本的训练样本数目较少的情况下,训练出的随机森林模型过拟合且模型泛化能力差。这样对于待检测的第一AP,直接使用训练出的随机森林模型检测第一AP是否为逻辑边缘AP,检测的精度较低。所以在第二种方式中并不是直接使用训练出的随机森林模型来对第一AP进行检测。而是根据训练出的随机森林模型,确定需要获取的第一AP的各特征对应的类别、特征阈值以及该特征与该特征阈值之间的判断条件,然后CI在上述步骤202至203中可以仅获取确定的类别对应的特征,从而可以减少获取第一AP的特征数目,可以减小运算量。
可选的,可以通过如下(1)-(5)的操作,来确定需要获取的第一AP的各特征对应的类别和特征阈值。该(1)-(5)的操作分别为:
(1):从该至少一个决策树中选择目标路径,目标路径的检测结果是逻辑边缘AP。
(2):从选择的目标路径包括的节点中获取目标类别对应的目标节点和该目标节点的判断结果对应的判断条件,该目标类别是第二训练集中的任一个特征属于的类别。
在目标路径中,目标节点判断属于目标类别的特征与目标节点对应的特征阈值之间的大小关系,得到判断结果,该判断结果对应的判断条件就是该大小关系,也就是说该判断结果对应的判断条件可能是大于关系或小于关系。该判断结果对应的判断条件用于表示该特征大于该特征阈值,或是该特征小于该特征阈值。
在本步骤中,获取的目标节点的数目可能大于或等于1,可能有一部分目标节点对应的判断条件是大于关系,而剩余的目标节点对应的判断条件是小于关系。
(3):在该获取的目标节点数目超过第二数目阈值时,将该目标类别确定为需要获取的第一AP的特征对应的类别。
(4):根据获取的目标节点对应的特征阈值获取目标类别对应的特征阈值。
可选的,根据获取的每个节点对应的特征阈值,计算出平均值,将该平均值作为目标类别对应的特征阈值。或者,对获取的每个目标节点对应的特征阈值进行排序,将排在中间位置的特征阈值作为目标类别对应的特征阈值。
(5):从目标节点的判断结果对应的判断条件中统计每种判断条件对应的目标节点数目,选择目标节点数目最大的判断条件作为目标类别的特征与目标类别对应的特征阈值之间的判断条件。
第三种方式,将第一AP的至少一个特征作为逻辑边缘AP检测模型的输入,并通过该逻辑边缘AP检测模型检测第一AP的是否为逻辑边缘AP。即在第三种方式中,由该逻辑边缘AP判断第一AP的至少一个特征是否满足第一条件,且在判断出满足第一条件时输出第一AP为逻辑边缘AP的结果。
在构造出的第二训练集中包括的正样本的训练样本较多的情况下,例如第二训练集中的正样本数目超过第三数目阈值,可以使用第二训练集训练出第二AI模型,得到逻辑边缘AP检测模型,且在第三种方式中逻辑边缘AP检测模型可以采用支持向量机(support vector machine,SVM)、线性回归算法(linear regression,LR)或卷积神经网络(convolutional neural network,CNN)等AI模型。
可选的,可以通过如下2141至2144的操作,来训练第二AI模型。该2141至2144的操作分别为:
2141:将第二训练集中的训练样本输入到第二AI模型中,每个训练样本的类别为正样本或负样本。
为了便于说明,将每个训练样本的类别分别称为每个训练样本的真实类别。
2142:第二AI模型预测第二训练集中的每个训练样本的类别。
对于第二训练集中的每个训练样本,第二AI模型从该训练样本中提取特征,基于提取的特征预测该训练样本为正样本的第一概率以及为负样本的第二概率,第一概率和第二概率之和等于1,输出两个概率中数值较大的概率对应的类别。
2143:对于每个训练样本,第二AI模型根据该训练样本的真实类别和第二AI模型对该训练样本预测后输出的类别,利用损失函数计算该训练样本的损失函数值。进一步,根据第二训练集中的所有训练样本的损失函数值调整第二AI模型的网络参数。
2144:第二AI模块确定是否继续训练,在确定继续训练时,返回执行2142,在确定停止训练时,此时的第二AI模块为逻辑边缘AP检测模型,结束返回。
具体地,在训练过程中,将得到的损失函数值进行曲线拟合,若得到的曲线逐渐收敛,且最后一次训练后的损失函数值小于预设定的阈值,则确定停止训练,否则,确定继续训练。
可选的,CI重复上述201至204的过程可以从目标WLAN的AP检测出各逻辑边缘AP。
步骤205:CI向AC发送优化请求,该优化请求包括各逻辑边缘AP的标识。
可选的,该优化请求还可以包括各逻辑边缘AP的目标终端数目等信息。
步骤206:AC可以接收该优化请求,根据该优化请求包括的逻辑边缘AP的标识,对逻辑边缘AP的性能进行优化。
对逻辑边缘AP的优化可以采用如下三种优化方式。该三种优化方式分别为:
第一优化方式,AC可以减小逻辑边缘AP的信号覆盖范围。
逻辑边缘AP的信号覆盖范围通常超出目标WLAN所在的场所,逻辑边缘AP的信号覆盖范围内可能包括位于该场所外的人流通道,所以通过减小逻辑边缘AP的信号覆盖范围,可以使逻辑边缘AP的信号覆盖范围不包括位于该场所外的人流通道,从而可以减小请求接入逻辑边缘AP的目标终端数目。
AC可以通过减小逻辑边缘AP的发射功率,来减小逻辑边缘AP的信号覆盖范围。AC可以分多次减小逻辑边缘AP的发射功率,以逐渐减小逻辑边缘AP的信号覆盖范围。在每次减小逻辑边缘AP的发射功率之后,确定是否继续减小逻辑边缘AP的发射功率。
可选的,AC可以以固定步长来减小逻辑边缘AP的发射功率。例如,假设固定步长为2db,即AC每次将逻辑边缘AP的发射功率减小2db。
可选的,AC需要减小逻辑边缘AP的发射功率时,可以向逻辑边缘AP发送指令,该指令可以包括该固定步长。逻辑边缘AP接收该指令,根据该指令中的该固定步长减小自身发射功率。
可选的,AC在每次减小逻辑边缘AP的发射功率之前,还可以通知管理员,以告知管理员需要减小逻辑边缘AP的发射功率以及需要减小多大的功率,并请求管理员确认。AC在接收到管理员的确认后减小逻辑边缘AP的发射功率。
在减小逻辑边缘AP的发射功率后,AC需要确定是否继续减小逻辑边缘AP的发射功率。在实现时,每次减小逻辑边缘AP的发射功率后,AC可以请求CI按上述步骤201至202的操作统计在时长为第一时长的时间段内请求接入逻辑边缘AP的目标终端数目,在该时间段内请求接入逻辑边缘AP的目标终端数目小于第二数目阈值时可以停止继续减小逻辑边缘AP的发射功率。该第二数目阈值可以是预设的阈值,也可以是根据最大目标终端数目得到,最大目标终端数目是已统计的逻辑边缘AP的目标终端数目中的最大值。例如,第二数目阈值可能等于x倍的最大目标终端数目,x为小于1的数值,可以为0.1、0.2或0.3等值。或者,
由于AC在每次减小逻辑边缘AP的发射功率后,便可以得到请求接入逻辑边缘AP的目标终端数目。对于AC连续多次得到的目标终端数目,AC发现该多个目标终端数目中任意两个目标终端数目之间的差值小于差值阈值,表示在AC减小逻辑边缘AP的发射功率后,请求接入逻辑边缘AP的目标终端数目不再继续大幅降低。此时,AC也确定停止减小逻辑边缘AP的发射功率。或者,
AC在每次减小逻辑边缘AP的发射功率后,获取逻辑边缘AP的实际发射功率,在逻辑边缘AP的实际发射功率低于功率阈值时,确定停止减小逻辑边缘AP的发射功率。或者,
AC在每次减小逻辑边缘AP的发射功率后,检测目标WLAN是否出现覆盖漏洞,在出现覆盖漏洞时,确定停止减小逻辑边缘AP的发射功率。
可选的,在WLAN中设置监控器,AC可以通过监控器监测WLAN中是否现覆盖漏洞,在出现覆盖漏洞时,就停止减小逻辑边缘AP的发射功率。可选的,AC还可以增加该逻辑边缘AP的发射功率,在通过监控器监测出WLAN中的覆盖漏洞消失后,停止增加逻辑边缘AP的发射功率。
第二种优化方式,AC可以控制逻辑边缘AP阻止下行信号强度低于下行信号强度阈值的终端接入。
下行信号低于下行信号强度阈值的终端往往位于逻辑边缘AP信号覆盖范围的边缘,而逻辑边缘AP的信号覆盖范围的边缘可能位于目标WLAN所在场所的外部,所以终端往往距 离逻辑边缘AP较远,导致其下行信号强度较低,低于下行信号强度阈值。
在第二种优化方式中,AC可以向逻辑边缘AP发送控制指令,逻辑边缘AP接收该控制指令后,当有终端请求接入逻辑边缘AP时,获取该终端的下行信号强度,在该下行信号强度低于下行信号强度阈值时,拒绝该终端接入,这样可以减小接入到逻辑边缘AP中的目标终端数目。
第三种优化方式,AC可以启动逻辑边缘AP的延迟接入功能,以减小接入到逻辑边缘AP中的目标终端数目。
逻辑边缘AP中包括延迟接入功能,在该延迟接入功能启动后,当有终端请求接入逻辑边缘AP时,逻辑边缘AP不会立即执行接入流程以让该终端接入,而是等待一段时间,当等待的时间超过第一时间阈值时,再启动接入流以让该终端接入。
然而,对于游牧终端,该游牧终端在逻辑边缘AP的信号覆盖范围内停留的时间往往较短,小于第一时间阈值,这样在游牧终端进入逻辑边缘AP的信号覆盖范围内请求接入时,由于逻辑边缘AP不会立即允许游牧终端接入,而是等待时间超过第一时间阈值后才启动接入流程。这样在逻辑边缘AP启动接入流程后,游牧终端已离开逻辑边缘AP的信号覆盖范围,因此逻辑边缘AP会停止继续执行该接入流程,从而可以减小接入到逻辑边缘AP中的游牧终端数目。
当然除了该三种优化方式外,还可以采用其他方式,例如通过提高逻辑边缘AP接入信号强度门限来减小接入逻辑边缘AP的目标终端数目,或通过在逻辑边缘AP的速率集中删除低速率选项来减小接入逻辑边缘AP的目标终端数目,对于其他方式在本步骤中不再一一详细说明。
在本申请实施例中,WLAN中的AP采集至少一个终端的连接信息,并向CI发送采集的至少一个终端的连接信息。CI根据该至少一个终端的连接信息,获取AP的至少一个特征,该至少一个特征可以包括AP的目标终端数目或该至少一个终端产生的各种连接事件总数目等,然后根据该AP的至少一个特征检测该AP是否为逻辑边缘AP。从而可以从WLAN的AP中准确地检测出逻辑边缘AP,这样可以便于对逻辑边缘AP进行处理,例如对逻辑边缘AP的性能进行优化,以减小或消除逻辑边缘AP产生影响。
参见图6,本申请实施例提供了一种检测装置300,所述装置300可以部署在上述任一实施例中的CI中,包括:
处理单元301,用于获取第一接入点AP的至少一个特征,至少一个特征包括在时长为第一时长的时间段内请求接入第一AP的目标终端数目,目标终端为接入状态异常的终端;
检测单元302,用于根据至少一个特征,检测述第一AP是否为逻辑边缘AP,逻辑边缘AP是信号覆盖范围抵达所在无线局域网WLAN的信号覆盖范围边缘的AP。
可选的,目标终端包括游牧终端、没有成功接入的终端或不属于所述第一AP所属WLAN的白名单中的终端中的至少一种。
可选的,所述游牧终端为在所接入的AP中停留的时间不超过第一时间阈值,且所述游牧终端在接入所接入的AP之前的预设时长内和在断开与所接入的AP连接之后的预设时长内没有接入所述WLAN的任一AP。
可选的,所述检测单元302,用于:
在所述至少一个特征满足第一条件的情况下,确定所述第一AP为逻辑边缘AP,所述第一条件包括所述目标终端数目大于第一数目阈值,或者,所述第一AP是所述WLAN中的目标终端数目最大的第一个数个AP中的一个AP,或者,所述目标终端数目与目标终端平均数目的差值大于差值阈值,所述差值阈值等于方差值的m倍,m大于1,所述目标终端平均数目和所述方差值是基于所述WLAN中的各AP的目标终端数目得到的。
可选的,所述目标终端包括所述游牧终端或不属于所述白名单中的终端中至少一种,
所述装置300还包括:接收单元303,
所述接收单元303,用于接收第一AP采集的至少一个终端的连接信息,所述连接信息包括所述终端的标识,所述至少一个终端为接入到所述第一AP的终端;
所述处理单元301,用于根据所述至少一个终端的连接信息,从所述至少一个终端中确定出目标终端;统计所述确定出的目标终端的数目。
可选的,所述第一AP用于对同一终端进行相邻两次采集连接信息的时间间隔不超过间隔阈值,所述连接信息还包括采集时间;
所述处理单元301,用于:
对于所述至少一个终端中的任一终端,获取所述任一终端的连接信息序列,所述连接信息序列包括按采集时间排列的连接信息,所述连接信息序列中的每个连接信息包括所述任一终端的标识,所述连接信息序列中的相邻两个连接信息的采集时间之间的间隔不超过所述间隔阈值;
根据所述连接信息序列确定所述任一终端是否为游牧终端。
可选的,所述处理单元301,用于:
根据所述连接信息序列获取所述任一终端接入所述第一AP的起始时间、所述任一终端在所述第一AP中的停留时间和所述任一终端断开与所述第一AP之间连接的断开时间;
在所述停留时间未超过第一时间阈值,以及在所述起始时间之前的预设时长内和在所述断开时间之后的预设时长内没有第二AP采集的所述任一终端的连接信息,确定所述任一终端为游牧终端,所述第二AP为所述WLAN中的除所述第一AP之外的其他AP,所述预设时长大于所述间隔阈值。
可选的,所述处理单元301,用于:
将所述连接信息序列作为终端检测模型的输入,并通过所述终端检测模型检测所述任一终端是否为游牧终端。
可选的,所述至少一个特征还包括所述至少一个终端的至少一种连接事件中的每种连接事件的总数目,对所述每种连接事件中的任一种连接事件,所述第一条件还包括所述任一种连接事件的总数目大于所述任一种连接事件对应的数目阈值,或者,所述第一条件还包括所述任一种连接事件的总数目小于所述任一种连接事件对应的数目阈值。
可选的,所述任一终端的连接信息还包括所述任一终端的上行信号强度或所述任一终端的数据传输时延中的至少一个;
所述处理单元301,还用于:
根据所述任一终端的连接信息序列中的最后一条连接信息包括的上行信号强度或数据传输时延中的至少一个,获取所述任一终端的连接事件;
统计所述任一种连接事件的总数目。
可选的,所述第一条件包括弱覆盖事件的总数目大于所述弱覆盖事件对应的数目阈值,
所述处理单元301,用于:
从强度阈值与弱覆盖事件的对应关系中的各强度阈值中,获取大于所述最后一条连接信息包括的上行信号强度的强度阈值;
根据所述获取的强度阈值,从所述强度阈值与弱覆盖事件的对应关系获取对应的弱覆盖事件作为所述任一终端的连接事件。
可选的,所述第一条件包括高时延事件的总数目大于所述高时延事件对应的数目阈值,
所述处理单元301,用于:
从时延阈值与高时延事件的对应关系中的各时延阈值中,获取大于所述最后一条连接信息包括的数据传输时延的时延阈值;
根据所述获取的时延阈值,从所述时延阈值与高时延事件的对应关系获取对应的高时延事件作为所述任一终端的连接事件。
可选的,所述连接信息还包括所述终端的信噪比、丢包率、重传率、信道利用率和所述终端接入的频段标识中的至少一个;
所述至少一个特征还包括信噪比统计值、丢包率统计值、重传率统计值、信道利用率统计值或接入所述频段标识对应的频段的总终端数目中的至少一个,所述信噪比统计值是基于所述至少一个终端的信噪比得到的,所述丢包率统计值是基于所述至少一个终端的丢包率得到的,所述重传率统计值是基于所述至少一个终端的重传率得到的,所述信道利用率统计值是基于所述至少一个终端的信道利用率得到的;
所述第一条件还包括所述信噪比统计值小于信噪比统计阈值、所述丢包率统计值大于丢包率统计阈值、所述重传率统计值大于重传率统计阈值、所述信道利用率统计值大于信道利用率阈值或所述总终端数目大于终端数目阈值中的至少一个。
可选的,所述处理单元301,还用于:
使用训练集训练随机森林模型,所述训练集包括多个训练样本和每个训练样本对应的类别,类别为正样本的训练样本包括逻辑边缘AP的至少一个特征,类别为负样本的训练样本包括非逻辑边缘AP的至少一个特征,训练后的所述随机森林模型包括至少一个决策树,所述决策树中的每条路径用于检测所述训练集中的任一AP是否为逻辑边缘AP,所述路径的叶子节点用于保存所述路径的检测结果,所述路径中除叶子节点之外的节点对应一个类别和特征阈值,所述节点用于判断第一特征是否超过所述节点对应的特征阈值以及根据判断结果选择属于所述路径的下一层节点,所述第一特征是所述任一AP的属于所述节点对应类别的特征;
根据所述至少一个决策树确定需要获取的所述第一AP的各特征对应的类别、特征阈值以及所述特征与所述特征阈值之间的判断条件。
可选的,所述处理单元301,用于:
从所述至少一个决策树中选择目标路径,所述目标路径的检测结果是逻辑边缘AP;
从所述选择的目标路径包括的节点中获取目标类别对应的目标节点和所述目标节点的判断结果对应的判断条件,所述目标类别是所述训练集中的任一个特征属于的类别;
在所述获取的目标节点数目超过第二数目阈值时,将所述目标类别确定为需要获取的所述第一AP的特征对应的类别;
根据所述获取的目标节点对应的特征阈值获取所述目标类别对应的特征阈值,从所述目标节点的判断结果对应的判断条件中统计每种判断条件对应的目标节点数目,选择目标节点数目最大的判断条件作为所述目标类别的特征与所述目标类别对应的特征阈值之间的判断条件。
可选的,所述检测单元302,用于:
将所述至少一个特征作为逻辑边缘AP识别模型的输入,并通过所述逻辑边缘AP识别模型检测所述第一AP是否为逻辑边缘AP。
可选的,所述处理单元301,还用于:
在检测出所述第一AP为逻辑边缘AP时,指示接入控制器AC减小所述第一AP的信号覆盖范围或指示所述AC控制所述第一AP阻止下行信号强度小于下行信号强度阈值的终端接入所述第一AP或指示所述AC启动所述第一AP的延迟接入功能。
在本申请实施例中,处理单元获取第一接入点AP的至少一个特征,该至少一个特征包括在时长为第一时长的时间段内请求接入第一AP的目标终端数目,目标终端为接入状态异常的终端;检测单元根据至少一个特征,检测第一AP是否为逻辑边缘AP。由于获取单元获取的第一AP的特征包括在时长为第一时长的时间段内请求接入第一AP的目标终端数目,而目标终端为接入状态异常的终端,从而检测单元可以基于第一AP的特征能够成功检测第一AP是否为逻辑边缘AP。
参见图7,本申请实施例提供了一种检测装置400,所述装置400可以部署在上述任一实施例中的CI中,包括:
接收单元401,用于接收第一接入点AP采集的至少一个终端的连接信息,所述连接信息包括所述终端的标识和采集所述连接信息的采集时间,所述至少一个终端为接入到所述第一AP的终端;
处理单元402,用于对于所述至少一个终端中的任一终端,获取所述任一终端的连接信息序列,所述连接信息序列包括按采集时间排列的连接信息,所述连接信息序列中的每个连接信息包括所述任一终端的标识,所述连接信息序列中的相邻两个连接信息的采集时间之间的间隔不超过所述间隔阈值;
所述处理单元402,还用于根据所述连接信息序列确定所述任一终端是否为游牧终端。
可选的,所述处理单元402,用于:
根据所述连接信息序列获取所述任一终端接入所述第一AP的起始时间、所述任一终端在所述第一AP中的停留时间和所述任一终端断开与所述第一AP之间连接的断开时间;
在所述停留时间未超过第一时间阈值,以及在所述起始时间之前的预设时长内和在所述断开时间之后的预设时长内没有第二AP采集的所述任一终端的连接信息,确定所述任一终端为游牧终端,所述第二AP为所述第一AP所属WLAN中的除所述第一AP之外的其他AP。
在本申请实施例,通过接收单元接收的第一AP采集的至少一个终端的连接信息,由于该连接信息包括终端的标识和采集该连接信息的采集时间,这样处理单元可以获取至任一终端的连接信息序列,该连接信息序列包括按采集时间排列的连接信息,该连接信息序列中的每个连接信息包括该任一终端的标识,该连接信息序列中的相邻两个连接信息的采集时间之间的间隔不超过该间隔阈值;如此处理单元根据该连接信息序列可以成功地确定该任一终端 是否为游牧终端。
参见图8,图8所示为本申请实施例提供的一种检测装置500示意图。该装置500包括至少一个处理器501,总线系统502,存储器503以及至少一个收发器504。
该装置800是一种硬件结构的装置,可以用于实现图6所述的装置300或图7所述的装置400中的功能模块。例如,本领域技术人员可以想到图6所示的装置300中的处理单元301和检测单元302,或者,图7所述的装置400中的处理单元402可以通过该至少一个处理器501调用存储器503中的代码来实现,图6所示的装置300中的接收单元303或图7所示的装置400中的接收单元401可以通过该收发器504来实现。
可选的,该装置500还可用于实现上述任一实施例中CI的功能。
可选的,上述处理器501可以是一个通用中央处理器(central processing unit,CPU),微处理器,特定应用集成电路(application-specific integrated circuit,ASIC),或一个或多个用于控制本申请方案程序执行的集成电路。
上述总线系统502可包括一通路,在上述组件之间传送信息。
上述收发器504,用于与其他设备或通信网络通信。
上述存储器503可以是只读存储器(read-only memory,ROM)或可存储静态信息和指令的其他类型的静态存储设备,随机存取存储器(random access memory,RAM)或者可存储信息和指令的其他类型的动态存储设备,也可以是电可擦可编程只读存储器(electrically erasable programmable read-only memory,EEPROM)、只读光盘(compact disc read-only memory,CD-ROM)或其他光盘存储、光碟存储(包括压缩光碟、激光碟、光碟、数字通用光碟、蓝光光碟等)、磁盘存储介质或者其他磁存储设备、或者能够用于携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其他介质,但不限于此。存储器可以是独立存在,通过总线与处理器相连接。存储器也可以和处理器集成在一起。
其中,存储器503用于存储执行本申请方案的应用程序代码,并由处理器501来控制执行。处理器501用于执行存储器503中存储的应用程序代码,从而实现本专利方法中的功能。
在具体实现中,作为一种实施例,处理器501可以包括一个或多个CPU,例如图8中的CPU0和CPU1。
在具体实现中,作为一种实施例,该装置500可以包括多个处理器,例如图8中的处理器501和处理器507。这些处理器中的每一个可以是一个单核(single-CPU)处理器,也可以是一个多核(multi-CPU)处理器。这里的处理器可以指一个或多个设备、电路、和/或用于处理数据(例如计算机程序指令)的处理核。
在具体实现中,作为一种实施例,该装置500还可以包括输出设备505和输入设备506。输出设备505和处理器501通信,可以以多种方式来显示信息。例如,输出设备505可以是液晶显示器(liquid crystal display,LCD)等。输入设备506和处理器501通信,可以以多种方式接受用户的输入。例如,输入设备506可以是触摸屏设备或传感设备等。
本领域普通技术人员可以理解实现上述实施例的全部或部分步骤可以通过硬件来完成,也可以通过程序来指令相关的硬件完成,所述的程序可以存储于一种计算机可读存储介质中,上述提到的存储介质可以是只读存储器,磁盘或光盘等。
以上所述仅为本申请一个实施例,并不用以限制本申请,凡在本申请的精神和原则之内,所作的任何修改、等同替换、改进等,均应包含在本申请的保护范围之内。

Claims (43)

  1. 一种检测方法,其特征在于,所述方法包括:
    获取第一接入点AP的至少一个特征,所述至少一个特征包括在时长为第一时长的时间段内请求接入所述第一AP的目标终端数目,所述目标终端为接入状态异常的终端;
    根据所述至少一个特征,检测所述第一AP是否为逻辑边缘AP,逻辑边缘AP是信号覆盖范围抵达所在无线局域网WLAN的信号覆盖范围边缘的AP。
  2. 如权利要求1所述的方法,其特征在于,所述目标终端包括游牧终端、没有成功接入的终端或不属于所述第一AP所属WLAN的白名单中的终端中的至少一种。
  3. 如权利要求2所述的方法,其特征在于,所述游牧终端为在所接入的AP中停留的时间不超过第一时间阈值,且在接入所接入的AP之前的预设时长内和在断开与所接入的AP连接之后的预设时长内没有接入所述WLAN的任一AP的终端。
  4. 如权利要求1至3任一项所述的方法,其特征在于,所述根据所述至少一个特征,检测所述第一AP是否为逻辑边缘AP,包括:
    在所述至少一个特征满足第一条件的情况下,确定所述第一AP为逻辑边缘AP,所述第一条件包括所述目标终端数目大于第一数目阈值,或者,所述第一AP是所述WLAN中的目标终端数目最大的第一个数个AP中的一个AP,或者,所述目标终端数目与目标终端平均数目的差值大于差值阈值,所述差值阈值等于方差值的m倍,m大于1,所述目标终端平均数目和所述方差值是基于所述WLAN中的各AP的目标终端数目得到的。
  5. 如权利要求1至4任一项所述的方法,其特征在于,所述目标终端包括所述游牧终端或不属于所述白名单中的终端中的至少一种,
    所述获取第一接入点AP的至少一个特征,包括:
    接收第一AP采集的至少一个终端的连接信息,所述连接信息包括所述终端的标识,所述至少一个终端为接入到所述第一AP的终端;
    根据所述至少一个终端的连接信息,从所述至少一个终端中确定出目标终端;
    统计确定出的目标终端的数目。
  6. 如权利要求5所述的方法,其特征在于,所述第一AP用于对同一终端进行相邻两次采集连接信息的时间间隔不超过间隔阈值,所述连接信息还包括采集时间;
    所述根据所述至少一个终端的连接信息,从所述至少一个终端中确定出目标终端,包括:
    对于所述至少一个终端中的任一终端,获取所述任一终端的连接信息序列,所述连接信息序列包括按采集时间排列的连接信息,所述连接信息序列中的每个连接信息包括所述任一终端的标识,所述连接信息序列中的相邻两个连接信息的采集时间之间的间隔不超过所述间隔阈值;
    根据所述连接信息序列确定所述任一终端是否为游牧终端。
  7. 如权利要求6所述的方法,其特征在于,所述根据所述连接信息序列确定所述任一终端是否为游牧终端,包括:
    根据所述连接信息序列获取所述任一终端接入所述第一AP的起始时间、所述任一终端在所述第一AP中的停留时间和所述任一终端断开与所述第一AP之间连接的断开时间;
    在所述停留时间未超过第一时间阈值,以及在所述起始时间之前的预设时长内和在所述断开时间之后的预设时长内没有第二AP采集的所述任一终端的连接信息,确定所述任一终端为游牧终端,所述第二AP为所述WLAN中的除所述第一AP之外的其他AP,所述预设时长大于所述间隔阈值。
  8. 如权利要求6所述的方法,其特征在于,所述根据所述连接信息序列确定所述任一终端是否为游牧终端,包括:
    将所述连接信息序列作为终端检测模型的输入,并通过所述终端检测模型检测所述任一终端是否为游牧终端。
  9. 如权利要求6至8任一项所述的方法,其特征在于,所述至少一个特征还包括所述至少一个终端的至少一种连接事件中的每种连接事件的总数目,对所述每种连接事件中的任一种连接事件,所述第一条件还包括所述任一种连接事件的总数目大于所述任一种连接事件对应的数目阈值,或者,所述第一条件还包括所述任一种连接事件的总数目小于所述任一种连接事件对应的数目阈值。
  10. 如权利要求9所述的方法,其特征在于,所述任一终端的连接信息还包括所述任一终端的上行信号强度或所述任一终端的数据传输时延中的至少一个;
    所述获取第一接入点AP的至少一个特征之前,还包括:
    根据所述任一终端的连接信息序列中的最后一条连接信息包括的上行信号强度或数据传输时延中的至少一个,获取所述任一终端的连接事件;
    统计所述任一种连接事件的总数目。
  11. 如权利要求10所述的方法,其特征在于,所述第一条件包括弱覆盖事件的总数目大于所述弱覆盖事件对应的数目阈值,
    所述获取所述任一终端的连接事件,包括:
    从强度阈值与弱覆盖事件的对应关系中的各强度阈值中,获取大于所述最后一条连接信息包括的上行信号强度的强度阈值;
    根据所述获取的强度阈值,从所述强度阈值与弱覆盖事件的对应关系获取对应的弱覆盖事件作为所述任一终端的连接事件。
  12. 如权利要求10所述的方法,其特征在于,所述第一条件包括高时延事件的总数目大于所述高时延事件对应的数目阈值,
    所述获取所述任一终端的连接事件,包括:
    从时延阈值与高时延事件的对应关系中的各时延阈值中,获取大于所述最后一条连接信息包括的数据传输时延的时延阈值;
    根据所述获取的时延阈值,从所述时延阈值与高时延事件的对应关系获取对应的高时延事件作为所述任一终端的连接事件。
  13. 如权利要求5至12任一所述的方法,其特征在于,所述连接信息还包括所述终端的信噪比、丢包率、重传率、信道利用率和所述终端接入的频段标识中的至少一个;
    所述至少一个特征还包括信噪比统计值、丢包率统计值、重传率统计值、信道利用率统计值或接入所述频段标识对应的频段的总终端数目中的至少一个,所述信噪比统计值是基于所述至少一个终端的信噪比得到的,所述丢包率统计值是基于所述至少一个终端的丢包率得到的,所述重传率统计值是基于所述至少一个终端的重传率得到的,所述信道利用率统计值是基于所述至少一个终端的信道利用率得到的;
    所述第一条件还包括所述信噪比统计值小于信噪比统计阈值、所述丢包率统计值大于丢包率统计阈值、所述重传率统计值大于重传率统计阈值、所述信道利用率统计值大于信道利用率阈值或所述总终端数目大于终端数目阈值中的至少一个。
  14. 如权利要求1至13任一项所述的方法,其特征在于,所述获取第一接入点AP的至少一个特征之前,还包括:
    使用训练集训练随机森林模型,所述训练集包括多个训练样本和每个训练样本对应的类别,类别为正样本的训练样本包括逻辑边缘AP的至少一个特征,类别为负样本的训练样本包括非逻辑边缘AP的至少一个特征,训练后的所述随机森林模型包括至少一个决策树,所述决策树中的每条路径用于检测所述训练集中的任一AP是否为逻辑边缘AP,所述路径的叶子节点用于保存所述路径的检测结果,所述路径中除叶子节点之外的节点对应一个类别和特征阈值,所述节点用于判断第一特征是否超过所述节点对应的特征阈值以及根据判断结果选择属于所述路径的下一层节点,所述第一特征是所述任一AP的属于所述节点对应类别的特征;
    根据所述至少一个决策树确定需要获取的所述第一AP的各特征对应的类别、特征阈值以及所述特征与所述特征阈值之间的判断条件。
  15. 如权利要求14所述的方法,其特征在于,所述根据所述至少一个决策树确定需要获取的所述第一AP的各特征对应的类别、特征阈值以及所述特征与所述特征阈值之间的判断条件,包括:
    从所述至少一个决策树中选择目标路径,所述目标路径的检测结果是逻辑边缘AP;
    从所述选择的目标路径包括的节点中获取目标类别对应的目标节点和所述目标节点的判断结果对应的判断条件,所述目标类别是所述训练集中的任一个特征属于的类别;
    在所述获取的目标节点数目超过第二数目阈值时,将所述目标类别确定为需要获取的所述第一AP的特征对应的类别;
    根据所述获取的目标节点对应的特征阈值获取所述目标类别对应的特征阈值,从所述目 标节点的判断结果对应的判断条件中统计每种判断条件对应的目标节点数目,选择目标节点数目最大的判断条件作为所述目标类别的特征与所述目标类别对应的特征阈值之间的判断条件。
  16. 如权利要求1至15任一项所述的方法,其特征在于,所述根据所述至少一个特征,检测所述第一AP是否为逻辑边缘AP,包括:
    将所述至少一个特征作为逻辑边缘AP识别模型的输入,并通过所述逻辑边缘AP识别模型检测所述第一AP是否为逻辑边缘AP。
  17. 如权利要求1至16任一项所述的方法,其特征在于,所述方法还包括:
    在检测出所述第一AP为逻辑边缘AP时,指示接入控制器AC减小所述第一AP的信号覆盖范围或指示所述AC控制所述第一AP阻止下行信号强度小于下行信号强度阈值的终端接入所述第一AP或指示所述AC启动所述第一AP的延迟接入功能。
  18. 一种检测方法,其特征在于,所述方法包括:
    接收第一接入点AP采集的至少一个终端的连接信息,所述连接信息包括所述终端的标识和采集所述连接信息的采集时间,所述至少一个终端为接入到所述第一AP的终端;
    对于所述至少一个终端中的任一终端,获取所述任一终端的连接信息序列,所述连接信息序列包括按采集时间排列的连接信息,所述连接信息序列中的每个连接信息包括所述任一终端的标识,所述连接信息序列中的相邻两个连接信息的采集时间之间的间隔不超过所述间隔阈值;
    根据所述连接信息序列确定所述任一终端是否为游牧终端。
  19. 如权利要求18所述的方法,其特征在于,所述根据所述连接信息序列确定所述任一终端是否为游牧终端,包括:
    根据所述连接信息序列获取所述任一终端接入所述第一AP的起始时间、所述任一终端在所述第一AP中的停留时间和所述任一终端断开与所述第一AP之间连接的断开时间;
    在所述停留时间未超过第一时间阈值,以及在所述起始时间之前的预设时长内和在所述断开时间之后的预设时长内没有第二AP采集的所述任一终端的连接信息,确定所述任一终端为游牧终端,所述第二AP为所述第一AP所属WLAN中的除所述第一AP之外的其他AP。
  20. 一种检测装置,其特征在于,所述装置包括:
    处理单元,用于获取第一接入点AP的至少一个特征,所述至少一个特征包括在时长为第一时长的时间段内请求接入所述第一AP的目标终端数目,所述目标终端为接入状态异常的终端;
    检测单元,用于根据所述至少一个特征,检测所述第一AP是否为逻辑边缘AP,逻辑边缘AP是信号覆盖范围抵达所在无线局域网WLAN的信号覆盖范围边缘的AP。
  21. 如权利要求20所述的装置,其特征在于,所述目标终端包括游牧终端、没有成功接 入的终端或不属于所述第一AP所属WLAN的白名单中的终端中的至少一种。
  22. 如权利要求21所述的装置,其特征在于,所述游牧终端为在所接入的AP中停留的时间不超过第一时间阈值,且所述游牧终端在接入所接入的AP之前的预设时长内和在断开与所接入的AP连接之后的预设时长内没有接入所述WLAN的任一AP。
  23. 如权利要求20至22任一项所述的装置,其特征在于,所述检测单元,用于:
    在所述至少一个特征满足第一条件的情况下,确定所述第一AP为逻辑边缘AP,所述第一条件包括所述目标终端数目大于第一数目阈值,或者,所述第一AP是所述WLAN中的目标终端数目最大的第一个数个AP中的一个AP,或者,所述目标终端数目与目标终端平均数目的差值大于差值阈值,所述差值阈值等于方差值的m倍,m大于1,所述目标终端平均数目和所述方差值是基于所述WLAN中的各AP的目标终端数目得到的。
  24. 如权利要求20至23任一项所述的装置,其特征在于,所述目标终端包括所述游牧终端或不属于所述白名单中的终端中至少一种,
    所述装置还包括:接收单元,
    所述接收单元,用于接收第一AP采集的至少一个终端的连接信息,所述连接信息包括所述终端的标识,所述至少一个终端为接入到所述第一AP的终端;
    所述处理单元,用于根据所述至少一个终端的连接信息,从所述至少一个终端中确定出目标终端;统计所述确定出的目标终端的数目。
  25. 如权利要求24所述的装置,其特征在于,所述第一AP用于对同一终端进行相邻两次采集连接信息的时间间隔不超过间隔阈值,所述连接信息还包括采集时间;
    所述处理单元,用于:
    对于所述至少一个终端中的任一终端,获取所述任一终端的连接信息序列,所述连接信息序列包括按采集时间排列的连接信息,所述连接信息序列中的每个连接信息包括所述任一终端的标识,所述连接信息序列中的相邻两个连接信息的采集时间之间的间隔不超过所述间隔阈值;
    根据所述连接信息序列确定所述任一终端是否为游牧终端。
  26. 如权利要求25所述的装置,其特征在于,所述处理单元,用于:
    根据所述连接信息序列获取所述任一终端接入所述第一AP的起始时间、所述任一终端在所述第一AP中的停留时间和所述任一终端断开与所述第一AP之间连接的断开时间;
    在所述停留时间未超过第一时间阈值,以及在所述起始时间之前的预设时长内和在所述断开时间之后的预设时长内没有第二AP采集的所述任一终端的连接信息,确定所述任一终端为游牧终端,所述第二AP为所述WLAN中的除所述第一AP之外的其他AP,所述预设时长大于所述间隔阈值。
  27. 如权利要求25所述的装置,其特征在于,所述处理单元,用于:
    将所述连接信息序列作为终端检测模型的输入,并通过所述终端检测模型检测所述任一终端是否为游牧终端。
  28. 如权利要求25至27任一项所述的装置,其特征在于,所述至少一个特征还包括所述至少一个终端的至少一种连接事件中的每种连接事件的总数目,对所述每种连接事件中的任一种连接事件,所述第一条件还包括所述任一种连接事件的总数目大于所述任一种连接事件对应的数目阈值,或者,所述第一条件还包括所述任一种连接事件的总数目小于所述任一种连接事件对应的数目阈值。
  29. 如权利要求28所述的装置,其特征在于,所述任一终端的连接信息还包括所述任一终端的上行信号强度或所述任一终端的数据传输时延中的至少一个;
    所述处理单元,还用于:
    根据所述任一终端的连接信息序列中的最后一条连接信息包括的上行信号强度或数据传输时延中的至少一个,获取所述任一终端的连接事件;
    统计所述任一种连接事件的总数目。
  30. 如权利要求29所述的装置,其特征在于,所述第一条件包括弱覆盖事件的总数目大于所述弱覆盖事件对应的数目阈值,
    所述处理单元,用于:
    从强度阈值与弱覆盖事件的对应关系中的各强度阈值中,获取大于所述最后一条连接信息包括的上行信号强度的强度阈值;
    根据所述获取的强度阈值,从所述强度阈值与弱覆盖事件的对应关系获取对应的弱覆盖事件作为所述任一终端的连接事件。
  31. 如权利要求29所述的装置,其特征在于,所述第一条件包括高时延事件的总数目大于所述高时延事件对应的数目阈值,
    所述处理单元,用于:
    从时延阈值与高时延事件的对应关系中的各时延阈值中,获取大于所述最后一条连接信息包括的数据传输时延的时延阈值;
    根据所述获取的时延阈值,从所述时延阈值与高时延事件的对应关系获取对应的高时延事件作为所述任一终端的连接事件。
  32. 如权利要求24至31任一所述的装置,其特征在于,所述连接信息还包括所述终端的信噪比、丢包率、重传率、信道利用率和所述终端接入的频段标识中的至少一个;
    所述至少一个特征还包括信噪比统计值、丢包率统计值、重传率统计值、信道利用率统计值或接入所述频段标识对应的频段的总终端数目中的至少一个,所述信噪比统计值是基于所述至少一个终端的信噪比得到的,所述丢包率统计值是基于所述至少一个终端的丢包率得到的,所述重传率统计值是基于所述至少一个终端的重传率得到的,所述信道利用率统计值是基于所述至少一个终端的信道利用率得到的;
    所述第一条件还包括所述信噪比统计值小于信噪比统计阈值、所述丢包率统计值大于丢包率统计阈值、所述重传率统计值大于重传率统计阈值、所述信道利用率统计值大于信道利用率阈值或所述总终端数目大于终端数目阈值中的至少一个。
  33. 如权利要求20至31任一项所述的装置,其特征在于,所述处理单元,还用于:
    使用训练集训练随机森林模型,所述训练集包括多个训练样本和每个训练样本对应的类别,类别为正样本的训练样本包括逻辑边缘AP的至少一个特征,类别为负样本的训练样本包括非逻辑边缘AP的至少一个特征,训练后的所述随机森林模型包括至少一个决策树,所述决策树中的每条路径用于检测所述训练集中的任一AP是否为逻辑边缘AP,所述路径的叶子节点用于保存所述路径的检测结果,所述路径中除叶子节点之外的节点对应一个类别和特征阈值,所述节点用于判断第一特征是否超过所述节点对应的特征阈值以及根据判断结果选择属于所述路径的下一层节点,所述第一特征是所述任一AP的属于所述节点对应类别的特征;
    根据所述至少一个决策树确定需要获取的所述第一AP的各特征对应的类别、特征阈值以及所述特征与所述特征阈值之间的判断条件。
  34. 如权利要求33所述的装置,其特征在于,所述处理单元,用于:
    从所述至少一个决策树中选择目标路径,所述目标路径的检测结果是逻辑边缘AP;
    从所述选择的目标路径包括的节点中获取目标类别对应的目标节点和所述目标节点的判断结果对应的判断条件,所述目标类别是所述训练集中的任一个特征属于的类别;
    在所述获取的目标节点数目超过第二数目阈值时,将所述目标类别确定为需要获取的所述第一AP的特征对应的类别;
    根据所述获取的目标节点对应的特征阈值获取所述目标类别对应的特征阈值,从所述目标节点的判断结果对应的判断条件中统计每种判断条件对应的目标节点数目,选择目标节点数目最大的判断条件作为所述目标类别的特征与所述目标类别对应的特征阈值之间的判断条件。
  35. 如权利要求20至34任一项所述的方法,其特征在于,所述检测单元,用于:
    将所述至少一个特征作为逻辑边缘AP识别模型的输入,并通过所述逻辑边缘AP识别模型检测所述第一AP是否为逻辑边缘AP。
  36. 如权利要求20至35任一项所述的装置,其特征在于,所述处理单元,还用于:
    在检测出所述第一AP为逻辑边缘AP时,指示接入控制器AC减小所述第一AP的信号覆盖范围或指示所述AC控制所述第一AP阻止下行信号强度小于下行信号强度阈值的终端接入所述第一AP或指示所述AC启动所述第一AP的延迟接入功能。
  37. 一种检测装置,其特征在于,所述装置包括:
    接收单元,用于接收第一接入点AP采集的至少一个终端的连接信息,所述连接信息包括所述终端的标识和采集所述连接信息的采集时间,所述至少一个终端为接入到所述第一AP 的终端;
    处理单元,用于对于所述至少一个终端中的任一终端,获取所述任一终端的连接信息序列,所述连接信息序列包括按采集时间排列的连接信息,所述连接信息序列中的每个连接信息包括所述任一终端的标识,所述连接信息序列中的相邻两个连接信息的采集时间之间的间隔不超过所述间隔阈值;
    所述处理单元,还用于根据所述连接信息序列确定所述任一终端是否为游牧终端。
  38. 如权利要求37所述的装置,其特征在于,所述处理单元,用于:
    根据所述连接信息序列获取所述任一终端接入所述第一AP的起始时间、所述任一终端在所述第一AP中的停留时间和所述任一终端断开与所述第一AP之间连接的断开时间;
    在所述停留时间未超过第一时间阈值,以及在所述起始时间之前的预设时长内和在所述断开时间之后的预设时长内没有第二AP采集的所述任一终端的连接信息,确定所述任一终端为游牧终端,所述第二AP为所述第一AP所属WLAN中的除所述第一AP之外的其他AP。
  39. 一种检测的装置,其特征在于,所述装置包括:
    至少一个处理器和至少一个存储器,所述至少一个存储器存储有一个或多个程序,所述一个或多个程序被配置成由所述至少一个处理器执行,所述一个或多个程序包含用于实现如权利要求1至19任一项所述的方法的指令。
  40. 一种计算机可读存储介质,其特征在于,所述计算机可读存储介质中存储有指令,当所述指令被处理器运行时,使得所述处理器执行如权利要求1至19任一项所述的方法。
  41. 一种检测系统,其特征在于,所述系统包括:数据分析器CI和第一接入点AP,
    所述第一AP,用于采集的至少一个终端的连接信息,向所述CI发送所述连接信息,所述连接信息包括所述终端的标识和采集所述连接信息的采集时间,所述至少一个终端为接入到所述第一AP的终端;
    所述CI,用于根据所述连接信息获取所述第一AP的至少一个特征,所述至少一个特征包括在时长为第一时长的时间段内请求接入所述第一AP的目标终端数目,所述目标终端包括游牧终端或不属于所述第一AP所在无线局域网WLAN的白名单中的终端中至少一种;根据所述至少一个特征,检测所述第一AP是否为逻辑边缘AP,逻辑边缘AP是信号覆盖范围抵达所在WLAN的信号覆盖范围边缘的AP。
  42. 如权利要求41所述的系统,其特征在于,所述系统还包括:接入控制器AC,
    所述CI,还用于在检测出所述第一AP为逻辑边缘AP时,向所述AC发送优化请求,所述优化请求包括所述第一AP的标识;
    所述AC,用于减小所述第一AP的信号覆盖范围或控制所述第一AP阻止下行信号强度小于下行信号强度阈值的终端接入所述第一AP或启动所述第一AP的延迟接入功能。
  43. 一种检测系统,其特征在于,所述系统包括:数据分析器CI和第一接入点AP,
    所述第一AP,用于采集的至少一个终端的连接信息,向所述CI发送所述连接信息,所述连接信息包括所述终端的标识和采集所述连接信息的采集时间,所述至少一个终端为接入到所述第一AP的终端;
    所述CI,用于对于所述至少一个终端中的任一终端,获取所述任一终端的连接信息序列,所述连接信息序列包括按采集时间排列的连接信息,所述连接信息序列中的每个连接信息包括所述任一终端的标识,所述连接信息序列中的相邻两个连接信息的采集时间之间的间隔不超过所述间隔阈值;根据所述连接信息序列确定所述任一终端是否为游牧终端。
PCT/CN2020/112902 2019-09-02 2020-09-01 检测方法、装置及系统 Ceased WO2021043146A1 (zh)

Priority Applications (4)

Application Number Priority Date Filing Date Title
JP2022513994A JP7293502B2 (ja) 2019-09-02 2020-09-01 検出方法、装置及びシステム
KR1020227008650A KR20220048014A (ko) 2019-09-02 2020-09-01 검출 방법, 장치 및 시스템
EP20861088.1A EP4021060B1 (en) 2019-09-02 2020-09-01 Detection method, apparatus and system
US17/683,995 US12279138B2 (en) 2019-09-02 2022-03-01 Detection method, apparatus, and system

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN201910824288.7 2019-09-02
CN201910824288 2019-09-02
CN201910866845.1 2019-09-13
CN201910866845.1A CN112449361B (zh) 2019-09-02 2019-09-13 检测方法、装置及系统

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US17/683,995 Continuation US12279138B2 (en) 2019-09-02 2022-03-01 Detection method, apparatus, and system

Publications (1)

Publication Number Publication Date
WO2021043146A1 true WO2021043146A1 (zh) 2021-03-11

Family

ID=74733380

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2020/112902 Ceased WO2021043146A1 (zh) 2019-09-02 2020-09-01 检测方法、装置及系统

Country Status (6)

Country Link
US (1) US12279138B2 (zh)
EP (1) EP4021060B1 (zh)
JP (1) JP7293502B2 (zh)
KR (1) KR20220048014A (zh)
CN (1) CN112449361B (zh)
WO (1) WO2021043146A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP4344283A4 (en) * 2021-06-30 2024-10-09 Huawei Technologies Co., Ltd. DEVICE IDENTIFICATION METHOD, AND TERMINAL DEVICE PROCESSING METHOD, APPARATUS AND SYSTEM
EP4489464A4 (en) * 2022-03-25 2025-10-01 Huawei Tech Co Ltd COVER HOLE DETECTION METHOD AND ASSOCIATED DEVICE

Families Citing this family (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2023075949A1 (en) * 2021-10-29 2023-05-04 Qualcomm Incorporated Assistance data selection
CN118175562A (zh) * 2022-12-01 2024-06-11 锐捷网络股份有限公司 一种异常链路确定方法、装置、电子设备及存储介质
CN117119503B (zh) * 2023-10-18 2024-01-23 山西腾炎科技有限公司 基于5g工业互联网的智慧矿山数据采集方法
CN117295099B (zh) * 2023-11-24 2024-03-12 深圳市盾魂科技有限公司 一种rtk高精度定位终端维护方法及系统

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101640889A (zh) * 2008-07-29 2010-02-03 华为技术有限公司 一种固定或游牧用户终端接入网络的方法及装置
US20140348165A1 (en) * 2009-07-14 2014-11-27 Saguna Networks Ltd. System and method for efficient delivery of multi-unicast communication traffic
CN107113639A (zh) * 2014-11-17 2017-08-29 三星电子株式会社 用于在无线通信系统中控制接入点操作的方法

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US7313113B1 (en) * 2003-04-04 2007-12-25 Airespace, Inc. Dynamic transmit power configuration system for wireless network environments
US9432848B2 (en) * 2004-03-23 2016-08-30 Aruba Networks, Inc. Band steering for multi-band wireless clients
US8005030B2 (en) * 2005-12-30 2011-08-23 Cisco Technology, Inc. WLAN diagnostics using traffic stream metrics
CN103916816B (zh) * 2014-03-25 2017-07-28 深圳市宏电技术股份有限公司 一种无线接入方法和接入控制器
US10313957B2 (en) * 2014-08-11 2019-06-04 Telefonaktiebolaget Lm Ericsson (Publ) Method and apparatus for access controlling
US10419438B2 (en) * 2015-06-09 2019-09-17 Intel Corporation System, apparatus and method for auto-optimization of access control policy and key management in a network authoring tool
US10129765B2 (en) 2015-08-24 2018-11-13 Kodacloud Inc. Adaptive wireless coverage configurations
DE102016102983A1 (de) * 2016-02-19 2017-08-24 Intel IP Corporation Verfahren und Vorrichtungen zum Verwalten von Paketdatennetzverbindungen

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101640889A (zh) * 2008-07-29 2010-02-03 华为技术有限公司 一种固定或游牧用户终端接入网络的方法及装置
US20140348165A1 (en) * 2009-07-14 2014-11-27 Saguna Networks Ltd. System and method for efficient delivery of multi-unicast communication traffic
CN107113639A (zh) * 2014-11-17 2017-08-29 三星电子株式会社 用于在无线通信系统中控制接入点操作的方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4021060A4

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP4344283A4 (en) * 2021-06-30 2024-10-09 Huawei Technologies Co., Ltd. DEVICE IDENTIFICATION METHOD, AND TERMINAL DEVICE PROCESSING METHOD, APPARATUS AND SYSTEM
EP4489464A4 (en) * 2022-03-25 2025-10-01 Huawei Tech Co Ltd COVER HOLE DETECTION METHOD AND ASSOCIATED DEVICE

Also Published As

Publication number Publication date
EP4021060A1 (en) 2022-06-29
US12279138B2 (en) 2025-04-15
CN112449361B (zh) 2022-10-18
CN112449361A (zh) 2021-03-05
JP7293502B2 (ja) 2023-06-19
KR20220048014A (ko) 2022-04-19
EP4021060A4 (en) 2022-09-14
US20220182864A1 (en) 2022-06-09
EP4021060B1 (en) 2025-03-26
JP2022546818A (ja) 2022-11-09

Similar Documents

Publication Publication Date Title
WO2021043146A1 (zh) 检测方法、装置及系统
CN112188533B (zh) 一种网络性能的上报方法及装置
US9071998B2 (en) Optimizing performance information collection
EP3817443A1 (en) Method executed by user equipment, and user equipment
US10813045B2 (en) Access controller, station, access point, and control method
US11757748B2 (en) Policy determining method, system, and apparatus
US20130344874A1 (en) Load management
KR101712899B1 (ko) 무선 베어러 제어 방법, 기기, 및 시스템
AU2019370757A1 (en) Method for controlling connection between terminal and network, and apparatus
WO2005086420A1 (en) A method for achieving load balance between access point devices in the wireless local area network
WO2021258767A1 (zh) 通信方法、装置及系统
US11172432B2 (en) Cell bar method and apparatus
WO2018227867A1 (zh) 一种5g大连接物联网中基于用户分类的差异化退避方法
CN112752319A (zh) 网络切换方法、装置、终端及存储介质
WO2011106941A1 (en) Network entity and method for providing a service for a user entity in a communication network
CN111050356B (zh) 数据传输的处理方法及装置、存储介质、电子装置
CN100455060C (zh) 小区下行准入方法、系统及网络侧设备
WO2011095128A1 (zh) 一种prach资源密度调整的方法及基站
CN111448595A (zh) 防止共享车辆停车拥塞的方法和装置
CN114079999A (zh) 通信控制方法、网元及存储介质
CN112995936B (zh) 230MHz与1800MHz电力无线专网融合组网方法和系统
CN109246760B (zh) 基于功率控制的无线局域网负载均衡方法及无线接入装置、无线接入设备
CN108449767B (zh) 一种应用于WiFi漫游区域中的网络负载均衡方法
EP4642102A1 (en) Method and apparatus for selecting ap
CA3104929C (en) Method and user equipment (ue) for generating report of random access channel(rach) performance of the ue for enhanced coverage level

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 20861088

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2022513994

Country of ref document: JP

Kind code of ref document: A

NENP Non-entry into the national phase

Ref country code: DE

ENP Entry into the national phase

Ref document number: 20227008650

Country of ref document: KR

Kind code of ref document: A

ENP Entry into the national phase

Ref document number: 2020861088

Country of ref document: EP

Effective date: 20220324