Estimation device, program, and computer-readable recording medium having the program recorded thereon

Abstract

To provide an estimation device capable of estimating the cause of an abnormality in communication and the position of occurrence of the cause of the abnormality so as to be capable of dealing with the cause of the abnormality in communication.SOLUTION: Cause estimating means 51 estimates an abnormality cause which is a cause of abnormality in wireless communication when it is impossible to receive a packet transmitted by a communication device, based on the degree of deviation of observation data from learning data. Position estimating means 52 selects a position estimating method associated with the cause of abnormality estimated by cause estimating means 51, and estimates the position of occurrence of the abnormality cause according the selected position estimating method by using QoS / channel status information indicating communication quality when the packet is transmitted by wireless communication using a channel. The learning data includes first QoS / channel status information acquired in a normal state, and the observation data includes second QoS / channel status information acquired in a state where it is unclear whether a packet can be received or not.SELECTED DRAWING: Figure 2

Description

[Technical field]

[0001] The present invention relates to an estimation device, a program, and a computer-readable recording medium having a program recorded thereon. [Background technology]

[0002] In recent years, in order to increase productivity at manufacturing sites, various operations have been automated, such as monitoring the manufacturing status using various sensors and displays. When connecting these devices and equipment via a wired network, there is a risk of physical disconnection and the cost of rewiring due to layout changes, so there is a growing demand to connect them via a wireless network.

[0003] However, environments such as factories have many factors that can cause the wireless environment to fluctuate, such as changes in the radio wave propagation environment due to the movement of metal objects, radio wave interference due to the coexistence of multiple wireless systems and interference from installed equipment, as well as line changes due to the recent trend toward small-lot, multi-product production, making them environments where the quality of wireless communications can be unstable.

[0004] On the other hand, if a situation occurs in which the required communication quality (QoS: Quality of Service) cannot be ensured in an application that uses wireless communication (i.e., a communication anomaly), it can cause the line to stop, lowering factory productivity and directly leading to huge losses. To prevent this, when a communication anomaly occurs, it is necessary to identify the type and location of the cause of the anomaly, take prompt and appropriate action, and restore the application as quickly as possible.

[0005] Conventionally, a method for remotely diagnosing QoS abnormalities in wireless LANs (Local Area Networks) caused by interference, obstruction, and failures is known (Non-Patent Document 1). This method detects abnormalities by measuring response delays of ACK (Acknowledgement) in CAMA / CA (Carrier Sense Multiple Access with Collision Avoidance), and estimates the cause of the abnormality. More specifically, if ACK response delays occur in all communication terminals, it is estimated that the cause of the abnormality is near the access point, and if ACK response delays occur in only some communication terminals, it is estimated that the cause of the abnormality is near the communication terminal. [Prior art documents] [Non-patent literature]

[0006] [Non-Patent Document 1] H. Fujita et al.: Access Point Initiated Approach for Interfered Node Detection in 802.11 WLANs. in IEEE Vehicular Technology Conference (VTC2015-Spring), in press (2015). [Non-Patent Document 2] KP Murphy, Machine Learning: A Probabilistic Perspective, The MIT Press, 2012. (p.510 Sec.14.7.2 KDE ). [Non-Patent Document 3] O. Anava and K. Levy, “k*-nearest neighbors: From global to local,” in Proc. Int. Conf. Neural Inf. Process. Syst., 2016, pp. 4916-4924. [Non-Patent Document 4] Tsuyoshi Ide, Keisuke Inoue: Knowledge Discovery from Time Series Data Using Nonlinear Transformation, 4th Data Mining Workshop, Japan Society for Software Science and Technology Data Mining Study Group, Study Group Materials Series ISSN 1341-870X, No.29, pp.1-8(2004). [Non-Patent Document 5] https: / / www.itu.int / dms_pubrec / itu-r / rec / p / R-REC-P.1238-6-200910-S!!PDF-E.pdf. [Non-Patent Document 6] S. Tsutsui, M. Pelikan, and DE Goldberg. Evolutionary algorithm using marginal histogram models in continuous domain. In Proc. of the Optimization by Building and Using Probabilistic Models OBUPM Workshop at the 2001 Genetic and Evolutionary Computation Conf., pages 230-233, 2001. Summary of the Invention [Problem to be solved by the invention]

[0007] However, Non-Patent Document 1 does not deal with estimating the location where the cause of the communication anomaly occurred, so even if the cause of the communication anomaly can be estimated, there is a problem in that it is difficult to deal with the cause of the communication anomaly.

[0008] Therefore, according to an embodiment of the present invention, an estimation device capable of estimating the cause of a communication anomaly and the location where the cause of the anomaly has occurred so that the cause of the communication anomaly can be dealt with is provided.

[0009] Furthermore, according to the embodiment of the present invention, there is provided a program for causing a computer to estimate the cause of a communication anomaly and the location where the cause of the anomaly has occurred so that the cause of the communication anomaly can be dealt with.

[0010] Furthermore, according to an embodiment of the present invention, there is provided a computer-readable recording medium having recorded thereon a program for causing a computer to estimate the cause of a communication anomaly and the location where the cause of the anomaly has occurred so as to deal with the cause of the communication anomaly. [Means for solving the problem]

[0011] (Configuration 1) According to an embodiment of the present invention, the estimation means includes a cause estimation means and a position estimation means. The cause estimation means estimates an anomaly cause that is a cause of an anomaly in wireless communication when a packet transmitted by a communication device cannot be received based on a deviation degree indicating a deviation degree of observation data from learning data. The position estimation means selects a position estimation method associated with the anomaly cause estimated by the cause estimation means, and estimates a position where the anomaly cause occurs according to the selected position estimation method using QoS / channel status information indicating a communication quality when a packet is transmitted by wireless communication using a channel. Then, when the cause estimation means determines that a packet is received from an arbitrary communication device in at least one communication device during a certain period, it determines that the arbitrary communication device is not faulty, and estimates the anomaly cause when it determines that the arbitrary communication device is not faulty. The learning data is made up of first QoS / channel status information that is QoS / channel status information acquired in a normal state in which a packet is received and the packet can be decoded. The observation data is made up of second QoS / channel status information that is QoS / channel status information acquired in a state in which it is unknown whether a packet can be received.

[0012] (Configuration 2) In the configuration 1, the cause estimation means calculates a plurality of deviations indicating a degree of deviation of distribution estimation data made of third QoS / channel status information, which is QoS / channel status information acquired at a timing different from that of the first QoS / channel status information, from learning data in a normal state, by changing an acquisition timing of the third QoS / channel status information, estimates a probability density distribution of the calculated plurality of deviations, calculates an estimation deviation indicating a degree of deviation of the second QoS / channel status information from the first QoS / channel status information, and When it is determined that the received signal strength in the channel state information is an abnormal drop indicating that the estimated deviation is equal to or lower than a first threshold in the probability density distribution, or an abnormal rise indicating that the estimated deviation is equal to or higher than a second threshold that is greater than the first threshold in the probability density distribution, it is presumed that the cause of the abnormality is a drop or rise in the received signal strength, and when it is determined that the received signal strength in all base stations is neither the abnormal drop nor the abnormal rise, it is presumed that the cause of the abnormality is a drop in communication throughput.

[0013] (Configuration 3) In the second configuration, the cause estimating means further estimates that the cause of the abnormality is the installation of an obstacle when it is determined that the received signal strengths for the plurality of communication devices are abnormally decreasing or increasing.

[0014] (Configuration 4) In configuration 3, when it is determined that the received signal strength of a single communication device is abnormally increasing or decreasing, the cause estimating means further estimates that the cause of the abnormality is a change in the installation position of the communication device.

[0015] (Configuration 5) In any of configurations 2 to 4, when the cause estimation means determines that the channel occupancy rate of the communication device is abnormally low in one or more base stations or monitor communication devices, it estimates that the cause of the abnormality is an increase in traffic of the communication device, noise from installed equipment, or introduction from outside the communication device.

[0016] (Configuration 6) In configuration 5, the cause estimation means further estimates that there is no abnormality when it is determined that the channel occupancy rate of the communication device is not abnormally low in all base stations or monitor communication devices.

[0017] (Configuration 7) In the configuration 5 or 6, the cause estimating means further estimates that the cause of the abnormality is an increase in traffic of the communication device when it is determined that the communication throughput has abnormally increased in one or more base stations.

[0018] (Configuration 8) In configuration 7, when it is determined that the communication throughput is not abnormally increased in all base stations, the cause estimation means further estimates that the cause of the abnormality is noise brought in from outside the communication device or noise from installed equipment.

[0019] (Configuration 9) In configuration 8, when the cause estimation means determines that the channel occupancy rate of an unmanaged communication device has abnormally increased in one or more base stations or monitor communication devices, it further estimates that the cause of the abnormality is brought in from outside the communication device.

[0020] (Configuration 10) In configuration 8 or 9, when the cause estimation means determines that the channel occupancy rate of unmanaged communication devices is not abnormally increased in all base stations or monitored communication devices, it estimates that the cause of the abnormality is noise from installed equipment.

[0021] (Configuration 11) In any of configurations 1 to 10, when the cause estimation means estimates that the cause of the abnormality is the introduction of a communication device from outside, the location estimation means estimates the location where the cause of the abnormality occurred using the received signal strength when a packet is received in a normal state from the communication device brought in from outside, and when the cause estimation means estimates that the cause of the abnormality is noise from installed equipment, the location estimation means estimates the location where the cause of the abnormality occurred using a first abnormality degree consisting of the minimum value of the Euclidean distance between the received signal strength at the antenna end of the monitor terminal device included in the observation data and the received signal strength included in the learning data.

[0022] (Configuration 12) In configuration 11, when the cause estimation means estimates that the cause of the abnormality is an increase in traffic of the communication device, the location estimation means further detects the link with the highest second degree of abnormality, which is formed by the minimum value of the Euclidean distance of the throughput included in the observation data to the throughput included in the learning data, and estimates the location where the cause of the abnormality occurred using the locations of three monitor terminal devices with high received signal strength received from the communication device that is the transmission source of the detected link and the received signal strengths at the three monitor terminal devices.

[0023] (Configuration 13) In configuration 12, when the cause estimation means estimates that the cause of the abnormality is the installation of an obstacle, the position estimation means estimates the position where the cause of the abnormality occurred as a midpoint of any number of links detected in order of highest multiplication result of a first degree of abnormality consisting of the minimum value of the Euclidean distance of the received signal strength at the antenna end of the monitor terminal device included in the observation data to the received signal strength included in the learning data and an average value of the received signal strength included in the learning data, or a midpoint of the midpoints with the multiplication result as a weight, or a center of gravity of the midpoints with the multiplication result as a weight, when the cause estimation means estimates that the cause of the abnormality is a change in the installation position of the communication device, the initial position of the communication device that is the source of the link with the highest first degree of abnormality as the position where the cause of the abnormality occurred.

[0024] (Configuration 14) Moreover, the program according to the embodiment of the present invention is A first step in which a cause estimation means estimates a cause of an anomaly that is a cause of an anomaly in wireless communication when a packet transmitted by a communication device cannot be received, based on a deviation degree indicating a degree of deviation of observed data from learning data; a second step in which the location estimation means selects a location estimation method associated with the cause of the abnormality estimated in the first step, and estimates a location where the cause of the abnormality has occurred according to the selected location estimation method by using QoS / channel status information indicating communication quality when a packet is transmitted by wireless communication using the channel; In the first step, when it is determined that at least one communication device has received a packet from a given communication device during a certain period of time, the cause estimation means determines that the given communication device is not at fault, and when it is determined that the given communication device is not at fault, estimates a cause of the abnormality; the learning data is composed of first QoS / channel status information which is QoS / channel status information acquired in a normal state in which a packet is received and the packet is successfully decoded; The observation data is a program executed by a computer, the program comprising second QoS / channel status information, which is QoS / channel status information acquired in a state in which it is unclear whether a packet can be received.

[0025] (Configuration 15) In configuration 14, the cause estimation means, in a first step, calculates a plurality of deviations indicating a degree of deviation of distribution estimation data made of third QoS / channel status information, which is QoS / channel status information acquired at a timing different from that of the first QoS / channel status information in a normal state, from learning data, by changing an acquisition timing of the third QoS / channel status information, estimates a probability density distribution of the calculated plurality of deviations, calculates an estimation deviation indicating a degree of deviation of the second QoS / channel status information from the first QoS / channel status information, and When the received signal strength in the second QoS / channel status information at the base station is determined to be an abnormal drop indicating that the estimated deviation is equal to or lower than a first threshold in the probability density distribution, or an abnormal rise indicating that the estimated deviation is equal to or higher than a second threshold that is greater than the first threshold in the probability density distribution, it is presumed that the cause of the abnormality is a drop or rise in the received signal strength, and when the received signal strength at all base stations is determined to be neither the abnormal drop nor the abnormal rise, it is presumed that the cause of the abnormality is a drop in communication throughput.

[0026] (Configuration 16) In configuration 15, when the cause estimating means determines in the first step that the received signal strengths for the plurality of communication devices are abnormally decreasing or increasing, it estimates that the cause of the abnormality is the installation of an obstacle.

[0027] (Configuration 17) In configuration 16, when the cause estimation means determines in the first step that there is a single communication device whose received signal strength is abnormally increasing or decreasing, it estimates that the cause of the abnormality is a change in the installation location of the communication device.

[0028] (Configuration 18) In any of configurations 15 to 17, in the first step, when the cause estimation means determines that the channel occupancy rate of the communication device is abnormally decreasing in one or more base stations or monitor communication devices, the cause estimation means estimates that the cause of the abnormality is an increase in traffic of the communication device, or noise from installed equipment, or introduction from outside the communication device.

[0029] (Configuration 19) In configuration 18, the cause estimation means, in the first step, further estimates that there is no abnormality when it is determined that the channel occupancy rate of the communication device is not abnormally decreased in all base stations or monitor communication devices.

[0030] (Configuration 20) In configuration 18 or 19, when the cause estimation means determines in the first step that communication throughput has abnormally increased in one or more base stations, it estimates that the cause of the abnormality is an increase in traffic of the communication device.

[0031] (Configuration 21) In configuration 20, in the first step, when the cause estimation means further determines that the communication throughput is not abnormally increased in all base stations, it estimates that the cause of the abnormality is noise brought in from outside the communication device or noise from installed equipment.

[0032] (Configuration 22) In configuration 21, in the first step, when the cause estimation means determines that the channel occupancy rate of an unmanaged communication device has abnormally increased in one or more base stations or monitor communication devices, the cause estimation means estimates that the cause of the abnormality is brought in from outside the communication device.

[0033] (Configuration 23) In configuration 21 or 22, when the cause estimation means, in the first step, further determines that the channel occupancy rate of the non-managed communication device is not abnormally increased in all base stations or monitored communication devices, it estimates that the cause of the abnormality is noise from installed equipment.

[0034] (Configuration 24) In any of configurations 14 to 23, in the second step, when the cause estimation means estimates that the cause of the abnormality is the introduction of a communication device from outside, the location estimation means estimates the location where the cause of the abnormality occurred using the received signal strength when a packet is received in a normal state from the communication device brought in from outside, and when the cause estimation means estimates that the cause of the abnormality is noise from installed equipment, the location estimation means estimates the location where the cause of the abnormality occurred using a first abnormality degree consisting of the minimum value of the Euclidean distance between the received signal strength at the antenna end of the monitor terminal device included in the observation data and the received signal strength included in the learning data.

[0035] (Configuration 25) In configuration 24, in the second step, when the cause estimation means estimates that the cause of the abnormality is an increase in traffic of the communication device, the location estimation means detects the link with the highest second degree of abnormality, which is formed by the minimum value of the Euclidean distance of the throughput included in the observation data to the throughput included in the learning data, and estimates the location where the cause of the abnormality occurred using the locations of three monitor terminal devices with high received signal strength received from the communication device that is the source of the detected link and the received signal strengths at the three monitor terminal devices.

[0036] (Configuration 26) In configuration 25, in the second step, when the cause estimation means estimates that the cause of the abnormality is the installation of an obstacle, the position estimation means estimates as the position where the cause of the abnormality occurred a midpoint of any number of links detected in descending order of the multiplication result of the first degree of abnormality consisting of the minimum value of the Euclidean distance of the received signal strength at the antenna end of the monitor terminal device included in the observation data to the received signal strength included in the learning data and the average value of the received signal strength included in the learning data, or a midpoint of the midpoints weighted by the multiplication result, or a center of gravity of the midpoints weighted by the multiplication result, when the cause estimation means estimates that the cause of the abnormality is a change in the installation position of the communication device, the position estimation means estimates as the position where the cause of the abnormality occurred an initial position of the communication device that is the source of the link with the highest first degree of abnormality.

[0037] (Configuration 27) Furthermore, according to an embodiment of the present invention, a recording medium is a computer-readable recording medium having the program according to any one of the fourteenth to twenty-sixth configurations recorded thereon. Effect of the Invention

[0038] The cause of communication abnormalities can be addressed. [Brief description of the drawings]

[0039]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Figure 16

Figure 17

Figure 18

Figure 19

Figure 20

Figure 21

Figure 22

Figure 23

Figure 24

Figure 25

Figure 26

Figure 27

Figure 28

Figure 29

Figure 30

Figure 31

Figure 32

Figure 33

Figure 34

Figure 35

Figure 36

Figure 37

Figure 38

Figure 39

Figure 40

Figure 41

Figure 42

Figure 43

Figure 44

Figure 45

Figure 46

[0040] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described in detail with reference to the drawings. In the drawings, the same or corresponding parts are designated by the same reference characters and their description will not be repeated.

[0041] Fig. 1 is a schematic diagram of a communication system according to an embodiment of the present invention, in which solid arrows represent wired communication and dotted arrows represent wireless communication.

[0042] Referring to FIG. 1, a communication system 100 includes an estimation device 10, a server 20, a communication device 30, a base station 40, and a monitor terminal device 50.

[0043] The communication system 100 is arranged, for example, at a manufacturing site. The estimation device 10 receives QoS / channel status information indicating communication quality when a packet is transmitted by wireless communication using a channel from a base station 40 via wired communication, and also receives the QoS / channel status information from a monitor terminal device 50 via wired and wireless communication. Then, based on the QoS / channel status information, the estimation device 10 estimates the cause of the communication abnormality and the location where the cause of the communication abnormality has occurred by a method described later.

[0044] The server 20 receives the QoS / channel condition information from the base station 40 by wired communication, and stores the received QoS / channel condition information.

[0045] The communication device 30 transmits a packet to the base station 40 via wireless communication. The base station 40 receives the packet from the communication device 30 via wireless communication. Then, the base station 40 detects QoS / channel state information when the packet is received, and transmits the detected QoS / channel state information to the estimation device 10.

[0046] The monitor terminal device 50 monitors wireless communication between the communication device 30 and the base station 40 to detect QoS / channel condition information, and transmits the detected QoS / channel condition information to the estimation device 10 via wired communication and wireless communication.

[0047] In addition, although FIG. 1 illustrates one communication device 30, one base station 40, and one monitor terminal device 50, in reality, the communication system 100 includes multiple communication devices 30, multiple base stations 40, and multiple monitor terminal devices 50.

[0048] Fig. 2 is a schematic diagram of the estimation device 10 shown in Fig. 1. Referring to Fig. 2, the estimation device 10 includes a wired cable 1, an antenna 2, receiving means 3 and 4, and estimation means 5. The estimation means 5 includes a cause estimation means 51 and a position estimation means 52.

[0049] The receiving means 3 receives QoS / channel status information from the base station 40 or the monitor terminal device 50 via the wired cable 1, and outputs the received QoS / channel status information to the cause estimation means 51 and position estimation means 52 of the estimation means 5.

[0050] The receiving means 4 receives the QoS / channel condition information from the monitor terminal device 50 via the antenna 2 , and outputs the received QoS / channel condition information to the cause estimating means 51 and position estimating means 52 of the estimating means 5 .

[0051] The cause estimation means 51 receives QoS / channel status information from the receiving means 3, 4, and estimates the cause of the communication abnormality based on the received QoS / channel status information by a method described later. Then, the cause estimation means 51 outputs the estimated cause of the abnormality to the position estimation means 52.

[0052] The position estimation means 52 receives the QoS / channel status information from the receiving means 3 and 4 , and receives the cause of the communication abnormality from the cause estimation means 51 .

[0053] The location estimation means 52 holds in advance a correspondence relationship between causes of communication anomalies and location estimation methods. When the location estimation means 52 receives the cause of communication anomalies from the cause estimation means 51, it selects a location estimation method corresponding to the received cause of anomalies. Then, the location estimation means 52 uses the selected location estimation method to estimate the location where the cause of communication anomalies has occurred based on the QoS / channel status information by a method to be described later.

[0054] Fig. 3 is a diagram for explaining a cause of a communication abnormality. Referring to Fig. 3(a), when three communication devices 30-1 to 30-3 are performing wireless communication with base station 40, if an obstacle 60 such as a metal plate is placed between communication devices 30-1 to 30-3 and base station 40, the received power decreases, causing a communication abnormality due to wireless communication between communication devices 30-1 to 30-3 and base station 40.

[0055] 3B, when a communication device 30 such as a sensor or a base station 40 is relocated due to battery replacement or equipment replacement, the installation position shifts, and a drop in reception power occurs in the wireless link of the device due to the effect of multipath fading, resulting in a communication abnormality.

[0056] 3(c), when the communication device 30 managed by the manufacturing site system (communication system 100) performs software updates, traffic in the wireless system increases and the throughput of the surrounding communication devices decreases, resulting in an abnormality in communication by the surrounding communication devices.

[0057] 3(d), when a tablet terminal 70 or the like is brought in from outside, communication by the tablet terminal 70 reduces the throughput of the surrounding communication devices 30. As a result, an abnormality occurs in the wireless communication between the communication device 30 and the base station 40.

[0058] 3(e), noise is generated from a manufacturing device 80 installed at a manufacturing site, causing a decrease in the throughput of the surrounding communication devices 30. As a result, an abnormality occurs in the wireless communication between the communication device 30 and the base station 40.

[0059] In this embodiment of the present invention, the cause estimation means 51 estimates the following as causes of the communication abnormality: "installation of an obstacle", "change in installation position of the communication device", "increase in traffic on the communication device", "introduction of a communication device from outside", and "noise from equipment installed at the manufacturing site" shown in FIG. 3.

[0060] The elements of QoS / channel status information are shown in Table 1.

[0061] [Table 1]

[0062] In Table 1, "Packet successfully received" means that a packet was received and the received packet was successfully decoded.

[0063] Only the monitor terminal device detects the received signal strength at the antenna end. Also, only the base station detects the throughput in the transport layer. The throughput in the transport layer is calculated for each pair of a source communication device and a destination communication device. Also, the throughput is calculated as the sum of the multiplication results obtained by identifying each packet, acquiring the time stored in the timestamp of the packet, and multiplying the number of packets received within a specified time interval from the current time by the packet size of each packet.

[0064] The "received signal strength when a packet is normally received," "channel occupancy rate of a communication device managed by the manufacturing site system," and "channel occupancy rate of a communication device not managed by the manufacturing site system" are calculated after identifying each link after decoding the packet. Also, whether a communication device is managed by the manufacturing site system or not is determined from the MAC address of the sender.

[0065] FIG. 4 is a diagram for explaining the QoS / channel status information detected by the monitor terminal device and the base station.

[0066] 4, communication devices A and B perform wireless communication with base station A, and communication device C performs wireless communication with base station B. In this case, communication device A transmits and receives packets of application App1 to and from base station A, communication device B transmits and receives packets of application App2 to and from base station A, and communication device C transmits and receives packets of application App3 to and from base station B.

[0067] Each of the monitor terminal devices A and B monitors the wireless communication between the communication device A and the base station A, monitors the wireless communication between the communication device B and the base station A, and monitors the wireless communication between the communication device C and the base station B.

[0068] As a result, the base stations A and B and the monitor terminal devices A and B detect the following information as QoS / channel status information: (1) Base station A - Received signal strength when packets are received from communication devices A and B ·Channel occupancy rate of communication devices A and B Throughput of application App1 to communication device A Throughput of application App2 for communication device B (2) Base station B - Received signal strength when a packet is received from communication device C -Channel occupancy rate of communication device C Throughput of application App1 to communication device C (3) Monitor terminal device A, B - Received signal strength at the antenna end - Received signal strength when base stations A and B send packets - Received signal strength when communication devices A, B, and C send packets ·Channel occupancy rate of base stations A and B ·Channel occupancy rate of communication devices A, B, and C -Channel occupancy rate of communication devices managed by the manufacturing site system -Channel occupancy of communication devices not managed by the manufacturing site system FIG. 5 is a diagram for explaining a method of calculating a received signal strength indicator (RSSI).

[0069] 5, received signal strengths RSSI_1 to RSSI_6 are received signal strengths when packets are normally received and are detected within a sampling interval SP1, while received signal strengths RSSI_7 to RSSI_10 are received signal strengths when packets are normally received and are detected within a sampling interval SP2.

[0070] The average value of received signal strengths RSSI_1 to RSSI_6 is set as the received signal strength RSSI at time t, and the average value of received signal strengths RSSI_7 to RSSI_10 is set as the received signal strength RSSI at time t+1. Sampling intervals SP1 and SP2 have the same time length, for example, 1 second.

[0071] The base station 40 calculates the received signal strength RSSI by the method described in FIG. 5, associates the calculated received signal strength RSSI with the IP address of the base station 40, and transmits the association data to the estimation device 10 and the server 20 by port mirroring.

[0072] On the other hand, the monitor terminal device 50 detects the received signal strength RSSI' consisting of the received power of the radio waves at the antenna end when the packet is received, and transmits the detected received signal strength RSSI' to the estimation device 10 via wired communication and wireless communication in association with the IP address of the monitor terminal device 50.

[0073] Fig. 6 is a conceptual diagram of a channel occupancy rate. Fig. 6 shows the channel occupancy rate detected by the monitor terminal device A when the communication devices A and B are performing wireless communication with the base station A.

[0074] Referring to FIG. 6, the monitor terminal device A monitors the time length T during which the base station A occupies the channel. LG_1 The result of dividing by the sampling interval SP1 is calculated as the channel occupancy rate of the base station A, and the time length T during which the communication device A occupies the channel is calculated. LG_2 The result of dividing by the sampling interval SP1 is calculated as the channel occupancy rate of communication device A, and the time length T during which communication device B occupies the channel is calculated. LG_3 The channel occupancy rate of communication device B is calculated by dividing the above by the sampling interval SP1.

[0075] In addition, the monitor terminal device A monitors the time length T LG_4 The result of dividing by the sampling interval SP2 is calculated as the channel occupancy rate of base station A, and the time length T during which communication device A occupies the channel is calculated. LG_5 The result of dividing by the sampling interval SP2 is calculated as the channel occupancy rate of communication device A, and the time length T during which communication device B occupies the channel is calculated. LG_6 The channel occupancy rate of communication device B is calculated by dividing the above by the sampling interval SP2.

[0076] Thus, the channel occupancy is calculated by dividing the time that the channel is occupied by the sampling interval, i.e., the channel occupancy is the percentage of time that the channel is occupied by packets transmitted by the communication device or base station.

[0077] The base station 40 calculates the channel occupancy rate by the method described in Fig. 6, associates the calculated channel occupancy rate with the IP address of the base station 40, and transmits them to the estimation device 10 and the server 20. The monitor terminal device 50 also calculates the channel occupancy rate by the method described in Fig. 6, associates the calculated channel occupancy rate with the IP address of the monitor terminal device 50, and transmits them to the estimation device 10 by wired communication and wireless communication.

[0078] [Estimation of the cause of the communication anomaly] Fig. 7 is a diagram for explaining a method of calculating the deviation. Referring to Fig. 7, the learning data consists of N (N is an integer greater than W) samples when packets are normally received. Each of the N samples consists of a QoS value (any of "1" to "5" shown in Table 1).

[0079] The distribution estimation data is obtained at a different timing from the learning data and consists of M (M is an integer greater than W) samples obtained when packets are normally received. Each of the M samples consists of a QoS value (any of "1" to "5" shown in Table 1). The magnitude relationship between N and M is arbitrary.

[0080] A curve k1 indicates the time dependency of the QoS values ​​constituting the learning data, and a curve k2 indicates the time dependency of the QoS values ​​constituting the distribution estimation data.

[0081] Each of the windows Wd_1, Wd_W+1, ···, Wd_N-W+1, Wd_1’ is a window with a width W (W is an integer of 2 or more). The width W represents the number of samples included in each of the windows Wd_1, Wd_W+1, ···, Wd_N-W+1, Wd_1’.

[0082] In the learning data, the window Wd_1 includes W samples from 1 to W, the window Wd_W+1 includes W samples from W+1 to 2W, and similarly hereinafter, the window Wd_N-W+1 includes W samples from N-W+1 to N. Also, in the distribution estimation data, the window Wd_1’ includes W samples from 1 to W.

[0083] Calculate the difference vector between the W samples in the window Wd_1 in the learning data and the W samples (represented by the curve k3) in the window Wd_1’ in the distribution estimation data, and calculate the sum n 1 of the elements of the calculated difference vector.

[0084] Next, calculate the difference vector between the W samples in the window Wd_W+1 in the learning data and the W samples (represented by the curve k3) in the window Wd_1’ in the distribution estimation data, and calculate the sum n 2 of the elements of the calculated difference vector.

[0085] Similarly hereinafter, calculate the difference vector between the W samples in the window Wd_N-W+1 in the learning data and the W samples (represented by the curve k3) in the window Wd_1’ in the distribution estimation data, and calculate the sum n N-W+1 of the elements of the calculated difference vector.

[0086] In this way, in the learning data, slide the window Wd and calculate the sums n 1 , n 2 , ···, n N-W+1 of them.

[0087] Then, the sum n 1,n 2 ,n N-W+1 The minimum value among these is taken as the deviation at timing t.

[0088] The above operation is performed for timings t+1, t+2, . . . to obtain multiple (M-W+1) deviations at timings t, t+1, t+2, . . .

[0089] Note that the window Wd_1' when calculating the deviation at timing t+1 consists of W samples from 2 to W+1, and the window Wd_1' when calculating the deviation at timing t+2 consists of W samples from 3 to W+2, and so on.

[0090] Moreover, the multiple (M-W+1) deviation degrees represent the degree of deviation (that is, the amount of deviation) of the distribution estimation data from the training data.

[0091] Furthermore, the learning data and the distribution estimation data are common in that they consist of QoS values ​​when a packet is normally received. However, the distribution estimation data is acquired at a timing different from the timing at which the learning data is acquired. Therefore, the multiple (M-W+1) deviations can be understood as deviations indicating the degree of deviation (i.e., the amount of deviation) between the QoS / channel condition information (QoS value and radio channel condition) at a first timing in a normal state and the QoS / channel condition information (QoS value and radio channel condition) at a second timing different from the first timing.

[0092] Fig. 8 is a conceptual diagram showing the probability density distribution of deviations. Referring to Fig. 8, when multiple (M-W+1) deviations are calculated, a kernel function is associated with each of the calculated multiple (M-W+1) deviations (●), and the probability density distribution of the deviations is estimated by superimposing the associated multiple (M-W+1) kernel functions (Non-Patent Document 2). As a result, the probability density distribution of the estimated deviations is represented by the curve k4.

[0093] Note that the multiple (M-W+1) deviations at timings t, t+1, t+2, and so on shown in Figure 7 are calculated for each of the "received signal strength when a packet is normally received," "received signal strength at the antenna end," "channel occupancy rate of communication devices managed by the manufacturing site system," "channel occupancy rate of communication devices not managed by the manufacturing site system," and "throughput at the transport layer" shown in Table 1, and the probability density distribution of the deviations shown in Figure 8 is estimated for each of the "received signal strength when a packet is normally received," "received signal strength at the antenna end," "channel occupancy rate of communication devices managed by the manufacturing site system," "channel occupancy rate of communication devices not managed by the manufacturing site system," and "throughput at the transport layer" shown in Table 1.

[0094] Therefore, the cause estimation means 51 calculates a plurality of deviations (M-W+1) by the above-mentioned method, and estimates the probability density distribution of the deviations based on the calculated plurality of deviations (M-W+1) for each of "received signal strength when packet is normally received", "received signal strength at antenna end", "channel occupancy rate of communication devices managed by the manufacturing site system", "channel occupancy rate of communication devices not managed by the manufacturing site system", and "throughput at transport layer" shown in Table 1, and retains the probability density distribution of the deviations in correspondence with "received signal strength when packet is normally received", "received signal strength at antenna end", "channel occupancy rate of communication devices managed by the manufacturing site system", "channel occupancy rate of communication devices not managed by the manufacturing site system", and "throughput at transport layer" shown in Table 1, respectively.

[0095] In the embodiment of the present invention, the probability density distribution of the deviation may be estimated using the histogram distribution estimation method described in Non-Patent Document 6.

[0096] FIG. 9 is a diagram showing a method for calculating the degree of deviation of observed data. When estimating the cause of a communication anomaly, observed data consisting of QoS / channel status information detected when it is unclear whether a packet has been received normally is used. Then, at a timing t at which the cause of the communication anomaly is estimated, the degree of deviation D diver is calculated by the method explained in FIG. 7 (see FIG. 9).

[0097] In this case, when estimating the cause of a communication anomaly and the location where the anomaly occurred at time t, a difference vector between W samples of observed data in a window Wd_1' (window indicated by a dotted line) whose end is set at time t and W samples of training data in windows Wd_1, Wd_W+1, . . . , Wd_N-W+1 is calculated, and the deviation degree D of the observed data is calculated as the sum of the elements of the calculated difference vector. diver_t is calculated.

[0098] In addition, when estimating the cause of a communication anomaly and the location where the anomaly occurred at timing t+1, a difference vector between W samples of the observation data in a window Wd_1' (a window indicated by a dashed line) whose ends are set at timing t+1 and W samples of the learning data in windows Wd_1, Wd_W+1, . . . , Wd_N-W+1 is calculated, and the deviation degree D of the observation data is calculated as the sum of the elements of the calculated difference vector. diver_t+1 is calculated.

[0099] Furthermore, when estimating the cause of a communication anomaly and the location where the anomaly occurred at timing t+2, a difference vector between W samples of the observation data in a window Wd_1' (a window indicated by a two-dot chain line) whose ends are set at timing t+2 and W samples of the learning data in windows Wd_1, Wd_W+1, . . . , Wd_N-W+1 is calculated, and the deviation degree D of the observation data is calculated as the sum of the elements of the calculated difference vector. diver_t+2 is calculated. The same applies below.

[0100] Fig. 10 is a diagram for explaining a method for determining an abnormal rise and fall of observed data. In Fig. 10, the vertical axis represents the probability density distribution of the deviation in a normal state, and the horizontal axis represents the deviation. Curve k4 shows the relationship between the probability density distribution of the deviation and the deviation.

[0101] Referring to FIG. 10, in the probability density distribution of the deviation (curve k4), the threshold value Ψ d ,Ψ r Set the threshold value Ψ d is the cumulative probability of being in the lower C d % (e.g., 0.1%), and the threshold Ψ r is the top C r % (e.g., 0.3%). Also, the threshold C d ,C r is set for each QoS / channel status information for which the deviation is calculated.

[0102] Then, the cause estimation means 51 estimates the deviation D of the observed data. diver is the threshold value Ψ r If the difference is more than 1, the QoS / channel status information of the observed data is determined to be abnormally rising, and the deviation degree D diver is the threshold value Ψ r If it is not, it is determined that the QoS / channel status information of the observed data is not an abnormal rise.

[0103] In addition, the cause estimation means 51 estimates the deviation D of the observed data. diver is the threshold value Ψ d If the QoS / channel status information of the observed data is abnormally descent, the deviation degree D of the observed data is diver is the threshold value Ψ d If it is not below this, it is determined that the QoS / channel status information of the observed data is not abnormally descent.

[0104] Fig. 11 is a schematic diagram of the correspondence table TBL1. Referring to Fig. 11, the correspondence table TBL1 includes QoS / channel status information in a normal state and a probability density distribution of a deviation degree. The QoS / channel status information in a normal state and the probability density distribution of a deviation degree are associated with each other.

[0105] Each of probability density distribution 1 to probability density distribution 5 is a probability density distribution estimated by the method described with reference to FIGS.

[0106] Probability density distribution 1 is associated with the received signal strength when a packet is normally received, probability density distribution 2 is associated with the received signal strength at the antenna end, probability density distribution 3 is associated with the channel occupancy rate of a communication device managed by the manufacturing site system, probability density distribution 4 is associated with the channel occupancy rate of a communication device not managed by the manufacturing site system, and probability density distribution 5 is associated with throughput in the transport layer.

[0107] Fig. 12 is a schematic diagram of the correspondence table TBL2. Referring to Fig. 12, the correspondence table TBL2 includes base stations, communication devices, monitor terminal devices, and whether or not packets are received during a certain period of time. The whether or not packets are received during a certain period of time is associated with the base stations, communication devices, and monitor terminal devices.

[0108] The base stations, communication devices and monitor terminal devices include base station IP addresses Add_AP_1 to Add_AP_I (I indicates the total number of base stations and is an integer greater than or equal to 2), communication device IP addresses Add_CM_1 to Add_CM_J (J indicates the total number of communication devices and is an integer greater than or equal to 2), and monitor terminal device IP addresses Add_MN_1 to Add_MN_M (M indicates the total number of monitor terminal devices and is an integer greater than or equal to 2).

[0109] The presence or absence of a packet being received within a certain period of time consists of yes / no. "Yes" indicates that a packet was received, and "no" indicates that a packet was not received. In the column for the presence or absence of a packet being received within a certain period of time, "yes" or "no" is stored in association with the IP addresses Add_AP_1 to Add_AP_I of the base stations, the IP addresses Add_CM_1 to Add_CM_J of the communication devices, and the IP addresses Add_MN_1 to Add_MN_M of the monitor terminal devices.

[0110] Fig. 13 is a schematic diagram of the correspondence table TBL3. Referring to Fig. 13, the correspondence table TBL3 includes QoS / channel status information in the observation data, a deviation degree, a base station / monitor terminal device that detected the QoS / channel status information, and a communication device that is a source of a packet that was the source of the detection of the QoS / channel status information. The QoS / channel status information in the observation data, the deviation degree, the base station / monitor terminal device that detected the QoS / channel status information, and the communication device that is the source of the packet that was the source of the detection of the QoS / channel status information are associated with each other.

[0111] The QoS / channel status information in the observation data includes received signal strength RSSI_1, RSSI_2,... when a packet is received normally, received signal strength RSSI'_1, RSSI'_2,... at the antenna end when the packet is received, channel occupancy rates CH_ocp_1, CH_ocp_2,... of communication devices managed by the manufacturing site system, channel occupancy rates CH'_ocp_1, CH'_ocp_2,... of communication devices not managed by the manufacturing site system, and throughput THP_1, THP_2,... in the transport layer.

[0112] The deviation is the deviation of the observed data. And the deviation is the deviation D diver_1_RSSI ,D diver_2_RSSI ,..., and the deviation D' diver_1_RSSI ,D' diver_2_RSSI , and the deviation D diver_1_CH ,D diver_2_CH ,..., and the deviation D' diver_1_CH ,D'diver_2_CH , and the deviation D diver_1_THP ,D diver_2_THP ,...and more.

[0113] Deviation degree D diver_1_RSSI ,D diver_2_RSSI , are associated with received signal strengths RSSI_1, RSSI_2, and the deviation D' diver_1_RSSI ,D' diver_2_RSSI , are associated with received signal strengths RSSI'_1, RSSI'_2, and the deviation D diver_1_CH ,D diver_2_CH , are associated with the channel occupancy rates CH_ocp_1, CH_ocp_2, and the deviation D' diver_1_CH ,D' diver_2_CH , are associated with the channel occupancy rates CH'_ocp_1, CH'_ocp_2, and the deviation D diver_1_THP ,D diver_2_THP ,··· are associated with throughputs THP_1, THP_2,···, respectively.

[0114] The base station / monitoring terminal device that detected the QoS / channel status information includes the IP address Add_AP_i (i=1 to I) of the base station and / or the IP address Add_MN_m (m=1 to M) of the monitoring terminal device.

[0115] Base station IP address Add_AP_i _RSSI and the IP address of the monitor terminal device Add_MN_m _RSSI is the deviation D diver_1_RSSI ,D diver_2_RSSI , . . . The IP address of the monitor terminal device Add_MN_m _RSSI’ is the deviation D' diver_1_RSSI ,D' diver_2_RSSI , . . . Furthermore, the IP address Add_AP_i _CH and the IP address of the monitor terminal device Add_MN_m _CH is the deviation D diver_1_CH ,D diver_2_CH, . . . Furthermore, the IP address Add_AP_i _CH’ and the IP address of the monitor terminal device Add_MN_m _CH’ is the deviation D' diver_1_CH ,D' diver_2_CH , . . . Furthermore, the IP address Add_AP_i _THP is the deviation D diver_1_THP ,D diver_2_THP ,....

[0116] In this case, the IP address of the base station Add_AP_i _RSSI ,Add_AP_i _CH ,Add_AP_i _CH’ ,Add_AP_i _THP Each of the IP addresses Add_MN_m may consist of only one base station IP address or may consist of more than one base station IP addresses. _RSSI’ ,Add_MN_m _CH ,Add_MN_m _CH’ Each of the IP addresses Add_CM_j may consist of only one IP address of a monitor terminal device, or may consist of more than one IP address of a monitor terminal device. _RSSI ,Add_CM_j _RSSI’ ,Add_CM_j _CH ,Add_CM_j _CH’ , Add_CM_s_j, and Add_CM_d_j may each consist of only the IP address of one communication device, or may consist of the IP addresses of one or more communication devices.

[0117] The communication device of the source of the packet from which the QoS / channel status information was detected includes the communication device IP addresses Add_CM_j (j=1 to J), Add_CM_s_j, and Add_d_CM_j. The IP address Add_CM_s_j indicates the IP address of the source communication device, and the IP address Add_CM_d_j indicates the IP address of the destination communication device.

[0118] IP address of the communication device Add_CM_j _RSSI is associated with the degree of divergence D diver_1_RSSI , D diver_2_RSSI , ···. Also, the IP address of the communication device Add_CM_j _RSSI’ is associated with the degree of divergence D' diver_1_RSSI , D' diver_2_RSSI , ···. Furthermore, the IP address of the communication device Add_CM_j _CH is associated with the degree of divergence D diver_1_CH , D diver_2_CH , ···. Furthermore, the IP address of the communication device Add_CM_j _CH’ is associated with the degree of divergence D' diver_1_CH , D' diver_2_CH , ···. Furthermore, the IP addresses of the communication devices Add_CM_s_j, Add_CM_d_j are associated with the degree of divergence D diver_1_THP , D diver_2_THP , ···.

[0119] In this case, the IP addresses Add_CM_j _RSSI , Add_CM_j _RSSI’ , Add_CM_j _CH , Add_CM_j _CH , Add_CM_s_j, Add_CM_d_j may each consist of the IP address of one communication device or may consist of the IP addresses of multiple communication devices.

[0120] By associating the QoS / channel status information with the degree of divergence, it is possible to determine whether the degree of divergence of the received signal strength RSSI, RSSI' of the communication device, the degree of divergence of the channel occupancy CH_ocp, CH'_ocp, and the degree of divergence of the throughput THP are abnormally rising (rise) or abnormally falling (descent) by the method described in FIG. 10.

[0121] The IP address Add_CM_s_j of the source communication device and the IP address Add_CM_d_j of the destination communication device are associated with the degree of divergence D diver_1_THP , D diver_2_THP, . . . are associated with each other because the throughputs THP_1, THP_2, . . . in the transport layer are calculated for a pair of a source communication device and a destination communication device.

[0122] Also, the IP address of the base station Add_AP_i _THP Only the deviation degree D diver_1_THP ,D diver_2_THP ,..., because only the base station detects the throughput in the transport layer.

[0123] In addition, the IP address of the monitor terminal device Add_MN_m _RSSI’ Only D' diver_1_RSSI ,D' diver_2_RSSI , . . . because only the monitor terminal device detects the received signal strength RSSI′ at the antenna end.

[0124] When the cause estimation means 51 receives from the receiving means 3, 4 learning data consisting of QoS / channel status information in a normal state (received signal strength RSSI, received signal strength RSSI' at the antenna end, channel occupancy rate CH_ocp of a communication device under management, channel occupancy rate CH'_ocp of a communication device not under management, and throughput THP) and distribution estimation data acquired at a timing different from the learning data and consisting of QoS / channel status information in a normal state (received signal strength RSSI, received signal strength RSSI' at the antenna end, channel occupancy rate CH_ocp of a communication device under management, channel occupancy rate CH'_ocp of a communication device not under management, and throughput THP), the cause estimation means 51 estimates a probability density distribution of deviation by the above-mentioned method based on the received learning data and the distribution estimation data for all of the received signal strength RSSI, received signal strength RSSI' at the antenna end, channel occupancy rate CH_ocp of a communication device under management, channel occupancy rate CH'_ocp of a communication device not under management, and throughput THP, and generates probability density distributions 1 to 5.

[0125] Then, the cause estimation means 51 creates a correspondence table TBL1 by corresponding the generated probability density distribution 1, probability density distribution 2, probability density distribution 3, probability density distribution 4, and probability density distribution 5 to the received signal strength RSSI, the received signal strength RSSI' at the antenna end, the channel occupancy rate CH_ocp of a communication device under management, the channel occupancy rate CH'_ocp of a communication device outside of management, and the throughput THP.

[0126] The cause estimation means 51 also detects signals S indicating that the base station, the communication device, and the monitor terminal device have received packets during a certain period of time. receive The cause estimation means 51 judges whether or not the signal S has been received from the receiving means 3, 4. Then, the cause estimation means 51 judges whether or not the signal S has been received at least once in a certain period. receive When the signal S is received from the receiving means 3, 4, receive The IP address of the source of the signal S (the IP address of the base station, the communication device, or the monitor terminal device) is set to "yes" for a certain period of time. receive When the signal S is not received from the receiving means 3, 4 for a certain period of time, receive A correspondence table TBL2 is created by setting "no" to IP addresses that have not received the command (IP addresses of the base station, the communication device, or the monitor terminal device).

[0127] Furthermore, when the cause estimation means 51 receives observation data consisting of QoS / channel status information (received signal strength RSSI, received signal strength RSSI' at the antenna end, channel occupancy rate CH_ocp of the managed communication device, channel occupancy rate CH'_ocp of the unmanaged communication device, and throughput THP) from the receiving means 3, 4, the cause estimation means 51 calculates the degree of deviation of the observation data by the above-mentioned method based on the received observation data for all of the received signal strength RSSI, received signal strength RSSI' at the antenna end, channel occupancy rate CH_ocp of the managed communication device, channel occupancy rate CH'_ocp of the unmanaged communication device, and throughput THP, and calculates the degree of deviation D diver_1_RSSI ,D diver_2_RSSI ,···;D' diver_1_RSSI ,D' diver_2_RSSI , ;D diver_1_CH ,Ddiver_2_CH ,···;D' diver_1_CH ,D' diver_2_CH , ;D diver_1_THP ,D diver_2_THP ,···get.

[0128] Then, the cause estimation means 51 detects the IP addresses Add_AP_i, Add_MN_m of the base station / monitor terminal device that detected the QoS / channel status information (received signal strength RSSI, received signal strength RSSI' at the antenna end, channel occupancy rate CH_ocp of a communication device under management, channel occupancy rate CH'_ocp of a communication device not under management, and throughput THP) from the packet including the QoS / channel status information, and detects the IP addresses Add_CM_j, Add_CM_s_j, Add_CM_d_j of the source communication device that transmitted the packet that is the basis for detecting the QoS / channel status information (received signal strength RSSI, received signal strength RSSI' at the antenna end, channel occupancy rate CH_ocp of a communication device under management, channel occupancy rate CH'_ocp of a communication device not under management, and throughput THP) from each element of the QoS / channel status information (received signal strength RSSI, received signal strength RSSI' at the antenna end, channel occupancy rate CH_ocp of a communication device under management, channel occupancy rate CH'_ocp of a communication device not under management, and throughput THP).

[0129] Then, the cause estimation means 51 calculates each element of the QoS / channel status information (received signal strength RSSI, received signal strength RSSI' at the antenna end, channel occupancy rate CH_ocp of the managed communication device, channel occupancy rate CH'_ocp of the unmanaged communication device, and throughput THP) and the deviation degree D diver_1_RSSI ,D diver_2_RSSI ,···;D' diver_1_RSSI ,D' diver_2_RSSI , ;D diver_1_CH ,D diver_2_CH ,···;D' diver_1_CH ,D' diver_2_CH , ;D diver_1_THP ,D diver_2_THP,···, the IP addresses Add_AP_i, Add_MN_m, and the IP addresses Add_CM_j, Add_CM_s_j, Add_CM_d_j are associated with each other to create a correspondence table TBL3.

[0130] After creating correspondence tables TBL1, TBL2, and TBL3, the cause estimation means 51 refers to correspondence table TBL2 to determine whether all base stations, communication devices, and monitor terminal devices within the communication system 100 have not received packets from any communication device during a certain period of time.

[0131] Then, when the cause estimation means 51 determines that all base stations, communication devices, and monitor terminal devices within the communication system 100 have not received packets from any communication device for a certain period of time, it estimates that the given communication device has failed.

[0132] On the other hand, when it is determined that at least one of the base stations, communication devices, and monitor terminal devices in the communication system 100 has received a packet from any communication device during a certain period, the cause estimation means 51 refers to the correspondence table TBL3 and calculates the IP address Add_AP_i of each base station. _RSSI The deviation D associated with diver_1_RSSI ,D diver_2_RSSI ,..., and then the probability density distribution 1 is detected by referring to the correspondence table TBL1.

[0133] Then, the cause estimation means 51 judges whether or not there is an abnormal descent in the received signal strength RSSI in one or more base stations. In this case, the cause estimation means 51 estimates the deviation D diver_1_RSSI ,D diver_2_RSSI ,..., one or more deviations D diver_j_RSSI is the threshold value Ψ of the probability density distribution 1 d If it is determined that the RSSI is below 0, it is determined that there is an abnormal descent in the received signal strength (RSSI) of one or more base stations, and the deviation degree D diver_1_RSSI ,D diver_2_RSSI ,... are all thresholds Ψ of probability density distribution 1 dIf it is determined that the received signal strength RSSI at one or more base stations is greater than 1, it is determined that there is no abnormal descent in the received signal strength RSSI at one or more base stations.

[0134] When determining that there is an abnormal descent in the received signal strength RSSI of one or more base stations, the cause estimation means 51 determines whether there is an abnormal descent in the received signal strength RSSI of multiple communication devices. In this case, the cause estimation means 51 uses the IP addresses Add_CM_j of the multiple communication devices in the correspondence table TBL3. _RSSI The deviation D associated with diver_1_RSSI ,D diver_2_RSSI ,..., the deviations of some of them are below the threshold value Ψ of the probability density distribution 1. d If it is determined that the received signal strength RSSI for the plurality of communication devices is abnormally decreased, the IP address Add_CM_j of one of the communication devices in the correspondence table TBL3 is determined to be equal to or less than the predetermined value. _RSSI Only the deviation corresponding to the threshold value Ψ of the probability density distribution 1 d or the IP addresses Add_CM_j of multiple communication devices in the correspondence table TBL3 _RSSI The deviation D associated with diver_1_RSSI ,D diver_2_RSSI ,..., the deviations of some of them are below the threshold value Ψ of the probability density distribution 1. d If it is determined that the received signal strengths RSSI for the plurality of communication devices are greater than 1, then it is determined that there is no abnormal descent in the received signal strengths RSSI for the plurality of communication devices.

[0135] When it is determined that there is an abnormal descent in the received signal strength RSSI for a plurality of communication devices, the cause estimation means 51 estimates that the cause of the communication abnormality is the "installation of an obstacle."

[0136] On the other hand, when the cause estimation means 51 determines that there is no abnormal descent in the received signal strength RSSI for the multiple communication devices, it estimates that the cause of the communication abnormality is "change in the installation location of the communication device".

[0137] In addition, when it is determined that there is no abnormal descent in the received signal strength RSSI in one or more base stations, the cause estimation means 51 determines whether there is an abnormal descent in the channel occupancy rate CH_ocp of the communication device managed by the communication system 100 in one or more base stations / monitor terminal devices. In this case, the cause estimation means 51 determines whether there is an abnormal descent in the channel occupancy rate CH_ocp of the communication device managed by the communication system 100 in one or more base stations / monitor terminal devices. _CH / Add_MN_m _CH The deviation D corresponding to diver_1_CH ,D diver_2_CH ,... is the threshold value Ψ of probability density distribution 3 in correspondence table TBL1. d If it is determined that the channel occupancy rate CH_ocp of the communication device managed by the communication system 100 is abnormally decreased (descent) in one or more base stations / monitor terminal devices, the deviation degree D diver_1_CH ,D diver_2_CH ,... is the threshold value Ψ of probability density distribution 3 d If it is determined that the channel occupancy rate CH_ocp of the communication device managed by the communication system 100 is greater than 1, it is determined that there is no abnormal descent in the channel occupancy rate CH_ocp of the communication device managed by the communication system 100 in one or more base stations / monitor terminal devices.

[0138] The cause estimation means 51 determines that there is no abnormal cause of communication when it determines that there is no abnormal descent in the channel occupancy rate CH_ocp of the communication devices managed by the communication system 100 in one or more base stations / monitor terminal devices.

[0139] On the other hand, when the cause estimation means 51 determines that there is an abnormal descent in the channel occupancy rate CH_ocp of the communication device managed by the communication system 100 in one or more base stations / monitor terminal devices, it determines whether there is an abnormal rise in the throughput THP in one or more base stations. In this case, the cause estimation means 51 determines whether there is an abnormal rise in the throughput THP in one or more base stations. _THP The deviation D associated with diver_1_THP ,D diver_2_THP , . . . One or more of the deviations is equal to or exceeds the threshold value Ψ of the probability density distribution 5 in the correspondence table TBL1. rIf this is the case, it is determined that there is an abnormal rise in throughput THP in one or more base stations, and the deviation D diver_1_THP ,D diver_2_THP ,... are all thresholds Ψ of probability density distribution 5 r If the THP is smaller than the threshold, it is determined that there is no abnormal rise in throughput THP in one or more base stations.

[0140] When it is determined that there is an abnormal rise in throughput THP in one or more base stations, the cause estimation means 51 estimates that the cause of the communication abnormality is an increase in traffic of the communication device.

[0141] On the other hand, when it is determined that there is no abnormal rise in the throughput THP in one or more base stations, the cause estimation means 51 determines whether there is an abnormal rise in the channel occupancy rate CH'_ocp of a communication device not managed by the communication system 100 in one or more base stations / monitor terminal devices. In this case, the cause estimation means 51 determines whether there is an abnormal rise in the channel occupancy rate CH'_ocp of a communication device not managed by the communication system 100 in one or more base stations / monitor terminal devices. _CH’ / Add_MN_m _CH’ The deviation D' corresponding to diver_1_CH ,D' diver_2_CH At least one of , ... r If it is determined that the channel occupancy rate CH'_ocp of a communication device not managed by the communication system 100 is abnormally increased in one or more base stations / monitor terminal devices, the deviation degree D' diver_1_CH ,D' diver_2_CH ,..., the deviation D' of all the r If it is determined that the channel occupancy rate CH'_ocp of a communication device not managed by the communication system 100 in one or more base stations / monitor terminal devices is not abnormally increased.

[0142] When the cause estimation means 51 determines that there is an abnormal rise in the channel occupancy rate CH'_ocp of a communication device not managed by the communication system 100 in one or more base stations / monitor terminal devices, it estimates that the cause of the communication abnormality is something that was brought in from outside the communication device.

[0143] On the other hand, when the cause estimation means 51 determines that there is no abnormal rise in the channel occupancy rate CH'_ocp of one or more base stations / monitor terminal devices of a communication device not managed by the communication system 100, it estimates that the cause of the communication abnormality is noise from equipment installed at the manufacturing site.

[0144] When the cause estimation means 51 estimates the cause of the communication abnormality by the above-mentioned method, the cause estimation means 51 outputs the estimated cause of the communication abnormality to the position estimation means 52 .

[0145] The cause estimating means 51 estimates the cause of the communication anomaly by the above-mentioned method at each of timings t, t+1, t+2, . . . at which the cause of the communication anomaly is estimated.

[0146] [Estimation of the location of the cause of the abnormality] Fig. 14 is a diagram for explaining a method of calculating the degree of abnormality (Rarity). Referring to Fig. 14, the position estimation means 52 calculates a difference vector between W samples in a window Wd_1' of the observation data and W samples in a window Wd_1 of the learning data at a timing t by the method explained in Fig. 9, and calculates a sum n' of the absolute values ​​of the elements of the calculated difference vector. 1 Calculate.

[0147] Thereafter, the position estimation means 52 sequentially slides the window to windows Wd_2, . . . , Wd_N-W+1 in the learning data, and similarly calculates the sum of absolute values ​​of the elements of the difference vector n′ 2 ,n' N-W+1 Calculate.

[0148] Then, the position estimation means 52 calculates the sum n' 1 ,n' 2 ,n'N-W+1 The minimum value of these is taken as the rarity at time t. In other words, the rarity is the minimum Euclidean distance of the observed data to the learning data. 1 ,n' 2 ,n' N-W+1 If there are multiple minimum values ​​among them, any one of the multiple minimum values ​​is determined as the degree of abnormality (Rarity) at timing t.

[0149] Similarly, the position estimation means 52 calculates the rarity at the timings t+1, t+2, . . . (1) Location estimation method 1 A method for estimating the location of the cause of the abnormality in a case where the cause of the abnormality in communication is the installation of an obstacle will be described.

[0150] 15 and 16 are diagrams for explaining a method of estimating the location where the cause of a communication anomaly occurs when the cause is the installation of an obstacle.

[0151] 15, for example, a plurality of monitor terminal devices 50 are arranged in a grid pattern. A communication device 30 performs wireless communication with a base station 40.

[0152] In such a situation, when an obstacle 60 is placed on the link between the communication device 30 and the base station 40, the received signal strength RSSI of the observation data is lower than the received signal strength RSSI of the learning data. The shorter the link length of the link is, the closer the communication device 30 and the base station 40 are to the obstacle 60.

[0153] 16, the position estimation means 52 calculates the rarity of the received signal strength RSSI at time t by the above-mentioned method.

[0154] Furthermore, the position estimation means 52 calculates the average value of the received signal strength RSSI of the learning data at time t.

[0155] FIG. 17 is a conceptual diagram of the rarity of the received signal strength indicator RSSI and the normalized value of the average value of the received signal strength indicator RSSI.

[0156] 17(a), the position estimation means 52 calculates the abnormality degree (Rarity) of the received signal strength RSSI at time t for all of the links between the monitor terminal device 50_1 and the base station 40_1, between the monitor terminal device 50_1 and the base station 40_2, between the monitor terminal device 50_1 and the communication device 30_1, ..., between the base station 40_1 and the communication device 30_2, and between the base station 40_1 and the communication device 30_1. Then, the position estimation means 52 acquires a normalized value by normalizing the abnormality degree (Rarity) by setting the minimum value of the calculated multiple abnormality degrees (Rarity) of the received signal strength RSSI to "0" and the maximum value of the calculated multiple abnormality degrees (Rarity) of the received signal strength RSSI to "1."

[0157] 17(b), the position estimation means 52 calculates an average value of the received signal strength RSSI of each link in the learning data for all of the links between the monitor terminal device 50_1 and the base station 40_1, between the monitor terminal device 50_1 and the base station 40_2, between the monitor terminal device 50_1 and the communication device 30_1, ..., between the base station 40_1 and the communication device 30_2, and between the base station 40_1 and the communication device 30_1. Then, the position estimation means 52 acquires a normalized value by normalizing the average value of the received signal strength RSSI by setting the minimum value of the calculated average values ​​of the received signal strength RSSI to "0" and the maximum value of the calculated average values ​​of the received signal strength RSSI to "1."

[0158] Then, the position estimation means 52 multiplies the normalized value of the degree of abnormality (Rarity) of the received signal strength RSSI by the normalized value of the average value of the received signal strength RSSI of the learning data to obtain a multiplication result.

[0159] The position estimation means 52 calculates the multiplication result of the normalized value of the rarity of the received signal strength RSSI and the normalized value of the average value of the received signal strength RSSI for all of the multiple monitor terminal devices 50 shown in FIG.

[0160] Referring again to FIG. 15, the position estimation means 52 then selects N 2 (N 2 is a natural number.) links (for example, three links between the communication device 30 and the monitor terminal devices 50A, 50B, and 50C). The position estimation means 52 then detects a midpoint MD1(x 1 ,y 1 ), the midpoint MD2(x 2 ,y 2 ) and the midpoint MD3(x 3 ,y 3 In this case, since the position estimation means 52 holds the positions of the communication device 30 and all the monitor terminal devices 50 in advance, it can calculate the intermediate points MD1, MD2, and MD3.

[0161] The position estimation means 52 estimates the intermediate point MD1(x 1 ,y 1 ),MD2(x 2 ,y 2 ),MD3(x 3 ,y 3 ) is calculated, the midpoint MD1(x 1 ,y 1 ),MD2(x 2 ,y 2 ),MD3(x 3 ,y 3 ) The normalized value of the abnormality degree (Rarity) of the received signal strength RSSI and the normalized value of the average value of the received signal strength RSSI are multiplied by ML 1 ,ML 2 ,ML 3 The center of gravity G(x G ,y G) as the location where the cause of the abnormality (installation of an obstacle) occurred. In this case, the position estimation means 52 estimates the abnormality degree R 1 ~R 3 Multiply each result by ML 1 ,ML 2 ,ML 3 Change to the center of gravity G(x G ,y G ) is calculated.

[0162] In addition, the position estimation means 52 2 When the value is "2", for example, two links between the communication device 30 and the monitor terminal devices 50A and 50B are detected, and the normalized value of the abnormality degree (Rarity) of the received signal strength RSSI at the midpoints MD1 and MD2 is multiplied by the normalized value of the average value of the received signal strength RSSI, ML 1 ,ML 2 The position estimation means 52 estimates the midpoint of the weighted N as the occurrence position of the cause of the abnormality (installation of an obstacle). 2 If N is "1", for example, one link between the communication device 30 and the monitor terminal device 50A is detected, and the midpoint MD1 is estimated as the location where the cause of the abnormality (installation of an obstacle) occurs. 2 The estimated position (x E ,y E ) is R in the following formula (6). m Multiply the result by ML 1 ,ML 2 It is calculated by substituting.

[0163] In this way, the position estimation means 52 ranks the detected N stations in descending order of the product of the normalized value of the abnormality degree (Rarity) of the received signal strength RSSI and the normalized value of the average value of the received signal strength RSSI. 2 The midpoint of the links, or the midpoint of the midpoints weighted by the multiplication result, or the center of gravity G of the midpoints weighted by the multiplication result is estimated as the location where the cause of the abnormality (installation of an obstacle) has occurred.

[0164] In the above, it has been explained that the location of the cause of the anomaly is estimated using the rarity of the received signal strength RSSI. However, instead of the rarity of the received signal strength RSSI, the mean square error between the observed data and the learning data may be used, or the anomaly calculated by an anomaly detection method such as the k-nearest neighbor method (k-NN method) and the Singular Spectrum Transform may be used (Non-Patent Documents 3 and 4). (2) Location estimation method 2 A method for estimating the location of the cause of a communication abnormality when the cause of the abnormality is a change in the installation location of a communication device will be described.

[0165] FIG. 18 is a diagram for explaining a method of estimating the location where the cause of a communication anomaly occurs when the cause of the anomaly is a change in the installation location of a communication device.

[0166] 18, for example, a plurality of monitor terminal devices 50 are arranged in a grid pattern. The installation positions of the communication devices 30 are shifted by several tens of centimeters. As a result, the received signal strength RSSI is reduced in the link Link_1 between the communication devices 30 and the base station 40 due to the influence of multipath fading.

[0167] Therefore, the location estimation means 52 calculates the abnormality degree (Rarity) of the received signal strength RSSI for a plurality of links, and estimates the location where the cause of the abnormality has occurred from the link Link_1 which has the highest abnormality degree (Rarity) of the received signal strength RSSI.

[0168] More specifically, the location estimation means 52 detects the link Link_1 having the highest rarity from among the multiple links, and estimates the initial location of the communication device 30 that is the source of the detected link Link_1 as the location where the cause of the abnormality occurred. Note that the location estimation means 52 holds the initial location of the communication device 30 in advance, and therefore can estimate the location where the cause of the abnormality occurred.

[0169] The reason why the location where the cause of the anomaly occurs is estimated from link Link_1, which has the highest rarity of the received signal strength RSSI, is that link Link_1 is affected by multipath fading and therefore experiences the greatest decrease in the received signal strength RSSI rarity, and therefore experiences the greatest decrease in the received signal strength RSSI rarity.

[0170] When there are a plurality of links Link_1 having the highest rarity of the received signal strength RSSI, any one of the plurality of links is detected as the link Link_1. (3) Location estimation method 3 A method for estimating the location of the cause of a communication anomaly when the cause of the anomaly is an increase in traffic will be described.

[0171] FIG. 19 is a diagram for explaining a method of estimating the location where the cause of a communication anomaly occurs when the cause of the anomaly is an increase in traffic.

[0172] 19, for example, a plurality of monitor terminal devices 50 are arranged in a grid pattern. Then, traffic increases in wireless communication between the communication device 30 and the base station 40.

[0173] As traffic increases, the rarity of the throughput THP of the observed data increases.

[0174] Therefore, the position estimation means 52 calculates the abnormality degree (Rarity) of the throughput THP for a plurality of links, and detects the link Link_2 having the highest abnormality degree (Rarity) of the throughput THP from the plurality of links. When there are a plurality of links having the highest abnormality degree (Rarity) of the throughput THP, any one of the plurality of links is detected as the link Link_2.

[0175] Then, the position estimation means 52 calculates the position of the communication device 30 based on the received signal strength RSSI when the base station 40 successfully receives a packet from the communication device 30 that is the source of link Link_2, and estimates the calculated position of the communication device 30 as the position where the cause of the abnormality occurred. (3-1) Method 1 for estimating the position of a communication device using received signal strength 20 is a diagram for explaining a communication device position estimation method 1 using received signal strength. FIG. 21 is a diagram showing received signal strengths RSSI at a plurality of monitor terminal devices 50.

[0176] 20, the wireless communication space is divided into a grid (for example, at 5 m intervals), and each intersection is set as a candidate point for the position of the communication device 30. In addition, the monitor terminal devices 50 are arranged in a grid pattern.

[0177] Then, using a radio wave propagation model, the received signal strength RSSI_EST when the monitor terminal device 50 receives a packet from the communication device 30 when the communication device 30 is placed at each candidate point is estimated.

[0178] Furthermore, each monitor terminal device 50 detects the received signal strength RSSI_Detect when the packet is actually received.

[0179] 21, received signal strengths RSSI_1_1, RSSI_1_2,..., RSSI_1_M are received signal strengths RSSI in M ​​monitor terminal devices MN_1 to MN_M estimated using a radio wave propagation model when communication device 30 is placed at candidate point 1, received signal strengths RSSI_2_1, RSSI_2_2,..., RSSI_2_M are received signal strengths RSSI in M ​​monitor terminal devices MN_1 to MN_M estimated using a radio wave propagation model when communication device 30 is placed at candidate point 2, and similarly, received signal strengths RSSI_S_1, RSSI_S_2,..., RSSI_S_M are received signal strengths RSSI in M ​​monitor terminal devices MN_1 to MN_M estimated using a radio wave propagation model when communication device 30 is placed at candidate point S. Here, S is an integer of 2 or more.

[0180] The received signal strength RSSI_EST is composed of received signal strengths RSSI_1_1, RSSI_1_2,...,RSSI_1_M; RSSI_2_1, RSSI_2_2,...,RSSI_2_M;...; RSSI_S_1, RSSI_S_2,...,RSSI_S_M.

[0181] Moreover, the received signal strengths RSSI_1, RSSI_2, ..., RSSI_M are the received signal strengths when the M monitor terminal devices MN_1 to MN_M actually receive packets when the communication device 30 is placed at the actual position PS_CM.

[0182] The received signal strength RSSI_Detect is made up of received signal strengths RSSI_1, RSSI_2, . . . , RSSI_M.

[0183] The position estimation means 52 calculates S Euclidean errors between the received signal strength RSSI_Detect (=RSSI_1, RSSI_2,..., RSSI_M) and the received signal strength RSSI_EST (=RSSI_1_1, RSSI_1_2,..., RSSI_1_M; RSSI_2_1, RSSI_2_2,..., RSSI_2_M;...; RSSI_S_1, RSSI_S_2,..., RSSI_S_M), and estimates the candidate point at which the smallest Euclidean error is obtained among the S calculated Euclidean errors as the position of the communication device 30. Note that when there are multiple smallest Euclidean errors, the position estimation means 52 estimates the candidate point at which any one of the multiple smallest Euclidean errors is obtained as the position of the communication device 30.

[0184] The position estimation means 52 estimates each of the received signal strengths RSSI_1_1, RSSI_1_2, ..., RSSI_1_M; RSSI_2_1, RSSI_2_2, ..., RSSI_2_M; ...; RSSI_S_1, RSSI_S_2, ..., RSSI_S_M using the ITU-R recommended indoor radio wave propagation loss model (non-patent document 5) expressed by the following equation.

[0185] [Number]

[0186] In Equation (1A), L total is the radio wave propagation loss, f is the radio wave frequency, N is the distance attenuation constant, d is the distance between each candidate point and each monitoring terminal device, and L f is the signal attenuation amount due to the obstacle wall, and c is the number of obstacle walls. Also, in Equation (1B), P t is the transmission power of the radio wave.

[0187] The position estimation means 52 preliminarily holds the radio wave frequency f, the distance attenuation constant N, the number of obstacle walls c, the signal attenuation amount L f due to the obstacle wall, and the position of each intersection of the grid shown in FIG. 20. As a result, the position estimation means 52 calculates the distance d based on the position of each intersection of the grid. Also, the position estimation means 52 receives the transmission power P t of the radio wave in the communication device 30 from the base station 40 that performs wireless communication with the communication device 30. Therefore, the position estimation means 52 substitutes the radio wave frequency f, the distance attenuation constant N, the distance d, the signal attenuation amount L f due to the obstacle wall, and the number of obstacle walls c into Equation (1A) to calculate the radio wave propagation loss L total . Then, the position estimation means 52 substitutes the radio wave propagation loss L total and the transmission power P t of the radio wave into Equation (1B) to calculate the received signal strength RSSI.

[0188] Therefore, the position estimation means 52 calculates the received signal strengths RSSI_1_1, RSSI_1_2, ···, RSSI_1_M; RSSI_2_1, RSSI_2_2, ···, RSSI_2_M; ···; RSSI_S_1, RSSI_S_2, ···, RSSI_S_M while changing the position of the candidate point of the communication device 30, that is, while changing the distance d. (3-2) Position Estimation Method 2 of Communication Device Using Received Signal Strength FIGS. 22 to 24 are diagrams for explaining the position estimation method 2 of the communication device using the received signal strength.

[0189] 22, the wireless communication space is divided into a grid with 5 m intervals, and each intersection is set as a candidate point for the position of the communication device 30. The monitor terminal devices 50 are arranged in a grid pattern. The position of the communication device 30 to be estimated is set as position PS_pfb.

[0190] 23 and 24, the vertical axis represents the received signal strength RSSI, and the horizontal axis represents the number of the monitor terminal device 50. Furthermore, curve k5 represents the received signal strength RSSI sequence when the 20 monitor terminal devices 50 actually received a packet, curve k6 represents the received signal strength RSSI sequence estimated by equation (1) at the positions of the 20 monitor terminal devices 50 when the communication device 30 is placed at the candidate point CP, and curve k7 represents the received signal strength RSSI sequence estimated by equation (1) at the positions of the 20 monitor terminal devices 50 when the communication device 30 is placed at the position PS_pfb.

[0191] Referring to FIG. 23, the received signal strength RSSI sequence (curve k6) estimated at the positions of the 20 monitor terminal devices 50 when the communication device 30 is placed at candidate point CP, and the received signal strength RSSI sequence (curve k7) estimated at the positions of the 20 monitor terminal devices 50 when the communication device 30 is placed at position PS_pfb deviate from the received signal strength RSSI sequence (curve k5) actually detected at the positions of the 20 monitor terminal devices 50.

[0192] As a result, when the absolute values ​​of the RSSI sequence are used, the accuracy of estimating the position of the communication device 30 may be degraded due to the discrepancy between the radio wave propagation model and the actual environment.

[0193] Therefore, in the method 2 for estimating the position of a communication device using received signal strength, the position estimation means 52 calculates the average value of the received signal strength RSSI sequence, and calculates the error vector between the absolute value of the received signal strength RSSI sequence and the average value of the received signal strength RSSI sequence for the received signal strength RSSI sequence shown by curve k5 and the received signal strength RSSI sequence shown by curve k6.

[0194] As a result, the position estimation means 52 calculates one error vector calculated based on the received signal strength RSSI sequence shown by curve k5, and S difference vectors (i.e., difference vectors equal to the number of candidate points of the communication device 30) calculated based on the received signal strength RSSI sequence shown by curve k6.

[0195] Then, the position estimation means 52 calculates S errors between one error vector and each of the S difference vectors, and estimates the candidate point at which the smallest error of the S calculated errors is obtained as the position of the communication device 30. Note that when there are multiple smallest errors, the position estimation means 52 estimates the candidate point at which any one of the multiple smallest errors is obtained as the position of the communication device 30. (3-3) Method 3 for estimating the position of a communication device using received signal strength FIG. 25 is a diagram for explaining a third method for estimating a position of a communication device using received signal strength.

[0196] With reference to FIG. 25, three monitor terminal devices 50A, 50B, and 50C arranged in the vicinity of the link with the highest abnormality degree (Rarity) of the throughput THP are selected. Then, the received signal strengths acquired by the three monitor terminal devices 50A, 50B, and 50C are set as received signal strengths RSSI_1, RSSI_2, and RSSI_3. The position of the monitor terminal device 50A is defined as (x 1 ,y 1 ), and the position of the monitor terminal device 50B is (x 2 ,y 2 ), and the position of the monitor terminal device 50C is (x 3 ,y 3 ), and the position of the communication device 30 to be estimated is (x, y).

[0197] D 1 is the radius of a circle centered on the position of the monitor terminal device 50A, and D 2 is the radius of a circle centered on the position of the monitor terminal device 50B, and D 3is the radius of a circle centered on the position of the monitor terminal device 50C.

[0198] As a result, the position (x, y) of the communication device 30 is estimated based on the position (x, y) of the monitor terminal device 50A. 1 ,y 1 ) to D 1 and the position of the monitor terminal device 50B (x 2 ,y 2 ) to D 2 and the position of the monitor terminal device 50C (x 3 ,y 3 ) to D 3 is located at a distance of

[0199] Therefore, the following equation holds:

[0200]

number

[0201] By expanding equations (2A), (2B), and (2C), respectively, we obtain the following equations.

[0202]

number

[0203] By subtracting equation (3B) from equation (3A) and subtracting equation (3C) from equation (3B), we obtain the following equation.

[0204]

number

[0205] The position estimation means 52 substitutes the received signal strength RSSI_1 into the equation (1B) to calculate the distance d according to the equations (1A) and (1B), and the calculated distance d is represented as D 1 In addition, the position estimation means 52 substitutes the received signal strength RSSI_2 into the equation (1B) to calculate the distance d according to the equations (1A) and (1B), and the calculated distance d is represented as D 2Furthermore, the position estimation means 52 substitutes the received signal strength RSSI_3 into the equation (1B) to calculate the distance d according to the equations (1A) and (1B), and the calculated distance d is represented as D 3 Let us assume that.

[0206] Then, the position estimation means 52 calculates x 1 ,x 2 ,x 3 ,y 1 ,y 2 ,y 3 ,D 1 ,D 2 ,D 3 into equation (4), and the two simultaneous equations expressed by equations (4A) and (4B) are solved to calculate x and y. The calculated (x, y) is then estimated as the position of the communication device 30.

[0207] As described above, in position estimation method 3, position estimation means 52 estimates the position of communication device 30 by any one of position estimation method 1 of a communication device using received signal strength, position estimation method 2 of a communication device using received signal strength, and position estimation method 3 of a communication device using received signal strength, and estimates the estimated position of communication device 30 as the position where the cause of the communication abnormality (increase in traffic) has occurred. (4) Location estimation method 4 A method for estimating the location where the cause of the communication abnormality occurs when the cause of the abnormality is brought in from outside the communication device will be described.

[0208] FIG. 26 is a diagram for explaining a method for estimating the occurrence location of the cause of a communication abnormality when the cause of the abnormality is brought in from outside the communication device.

[0209] 26, a plurality of monitor terminal devices 50 are arranged in a grid pattern, and a communication device 90 is brought in from outside.

[0210] The plurality of monitor terminal devices 50 detects the received signal strength RSSI when a packet is normally received from the communication device 90 brought in from outside.

[0211] The position estimation means 52 estimates the position of the communication device 90 using the above-mentioned position estimation method 3 (any of the position estimation methods 1 to 3 of the communication device using the received signal strength) based on the received signal strength RSSI detected by the multiple monitor terminal devices 50, and estimates the estimated position of the communication device 90 as the position where the cause of the abnormality occurred. (5) Location estimation method 5 A method for estimating the location of the cause of the abnormality in a case where the cause of the abnormality in communication is noise from equipment installed at the manufacturing site will be described.

[0212] FIG. 27 is a diagram for explaining a method for estimating the location where the cause of a communication anomaly occurs when the cause is noise from an installed device at the manufacturing site.

[0213] 27, a plurality of monitor terminal devices 50 are arranged in a grid pattern. When a manufacturing device 80 generates noise, the received signal strength RSSI' at the antenna terminal increases.

[0214] Therefore, the position estimation means 52 calculates the rarity of the received signal strength RSSI' observed at the antenna end by the monitor terminal device 50 for all nine monitor terminal devices 50 by the above-mentioned method, and selects N monitor terminal devices 50 in descending order of rarity based on the nine rarity of the received signal strength RSSI' thus calculated. 3 (N 3 is a natural number.) The positions of monitor terminal devices are obtained.

[0215] In FIG. 27, the position estimation means 52 estimates the positions (x 1 ,y 1 ),(x 2 ,y 2 ),(x 3 ,y 3 Then, the position estimation means 52 obtains the abnormality degree R abn_1 , the abnormality degree R of the received signal strength at the monitor terminal device 50B abn_2, and the abnormality degree R of the received signal strength at the monitor terminal device 50C. abn_3 and the position (x 1 ,y 1 ),(x 2 ,y 2 ),(x 3 ,y 3 ) and the anomaly degree R abn_1 ,R abn_2 ,R abn_3 The center of gravity G(x G ,y G ) and calculate the calculated center of gravity G(x G ,y G ) is estimated to be the location where the cause of the abnormality occurred.

[0216] In this case, the position estimation means 52 calculates x G ,y G Calculate.

[0217]

number

[0218] In formula (5), R 1 is the abnormality R abn_1 R 2 is the abnormality R abn_2 R 3 is the abnormality R abn_3 It consists of:

[0219] In addition, for example, when the positions of the two monitor terminal devices 50A and 50B are acquired, the degree of abnormality R abn_1 and abnormality R abn_2 The midpoint between 50A and 50B weighted by is estimated as the location where the cause of the abnormality occurred. Furthermore, for example, when the position of one monitor terminal device 50A is acquired, the position of 50A is estimated as the location where the cause of the abnormality occurred.

[0220] In addition, N 3 The estimated position (x E ,y E ) is calculated by the following formula:

[0221]

number

[0222] In formula (6), R m In 3 N observed at the antenna ends of the monitor terminal devices 50 3 Received signal strengths RSSI'_1 to RSSI'_N 3 N 3 The abnormality (rarity) of each is assigned.

[0223] Fig. 28 is a schematic diagram of correspondence table TBL4. Referring to Fig. 28, correspondence table TBL4 includes causes of communication abnormalities and methods of estimating a position. The causes of communication abnormalities and the methods of estimating a position are associated with each other.

[0224] Referring to FIG. 28, position estimation method 1 corresponds to the installation of an obstacle, position estimation method 2 corresponds to a change in the installation position of the communication device, position estimation method 3 corresponds to an increase in traffic of the communication device, position estimation method 4 corresponds to the bringing in of the communication device from outside, and position estimation method 5 corresponds to noise from equipment installed at the manufacturing site.

[0225] The position estimation means 52 holds a correspondence table TBL4, and when it receives a cause of a communication anomaly from the cause estimation means 51, it refers to the correspondence table TBL4, selects a position estimation method corresponding to the cause of the communication anomaly received from the cause estimation means 51, and estimates the occurrence position of the cause of the anomaly using the selected position estimation method.

[0226] Fig. 29 is a flowchart for explaining the operation of the estimation device 10 shown in Fig. 2. With reference to Fig. 29, when the operation of the estimation device 10 is started, the cause estimation means 51 reads the learning data (step S1) and reads the distribution estimation data (step S2).

[0227] Then, the cause estimation means 51 calculates the degree of deviation of the distribution estimation data based on the learning data and the distribution estimation data (step S3). That is, the cause estimation means 51 calculates the degree of deviation in a normal state.

[0228] Thereafter, the cause estimation means 51 estimates the probability density distribution of the calculated deviation degree (step S4).

[0229] Then, the cause estimation means 51 reads the observed data (step S5), and calculates the degree of discrepancy between the observed data and the learned data (step S6).

[0230] Then, the cause estimation means 51 judges whether the calculated deviation degree is abnormally increased or decreased (step S7).

[0231] Thereafter, the cause estimating means 51 estimates the cause of the communication abnormality based on the abnormal increase and the abnormal decrease (step S8).

[0232] Subsequently, the position estimation means 52 judges whether or not a "failure" or "no abnormality" has been estimated in step S8 (step S9).

[0233] When it is determined in step S9 that the communication is not estimated to be "fault" or "no abnormality", the position estimation means 52 reads the cause of the communication abnormality from the cause estimation means 51 (step S10). Then, the position estimation means 52 selects a position estimation method corresponding to the read cause of the communication abnormality (step S11).

[0234] Then, the location estimation means 52 estimates the location where the cause of the communication abnormality occurs by using the selected location estimation method (step S12).

[0235] Then, when it is estimated in step S9 that the communication system 100 is “fault” or “no abnormality”, or after step S12, the estimation means 5 judges whether or not the termination condition is satisfied (step S13). In this case, if the communication system 100 is operating, the estimation means 5 judges that the termination condition is not satisfied, and if the communication system 100 is stopped, the estimation means 5 judges that the termination condition is satisfied.

[0236] When it is determined in step S13 that the end condition is not satisfied, the series of operations proceeds to step S5, and steps S5 to S13 are repeatedly executed until it is determined in step S13 that the end condition is satisfied.

[0237] Then, if it is determined in step S13 that the termination condition is satisfied, the operation of the estimation device 10 ends.

[0238] FIG. 30 is a flowchart for explaining the detailed operation of step S3 shown in FIG.

[0239] 30, after step S2 in FIG. 29, the cause estimation means 51 sets q=1 (step S31) and sets T=t (step S32). q is made up of "1" to "5" which are arguments of each element of the QoS / channel status information shown in Table 1. Furthermore, timing T indicates the timing of executing the flowchart shown in FIG. 29.

[0240] After step S32, the cause estimation means 51 selects the q-th element QoS_est_q of the QoS / channel condition information from the distribution estimation data using a window of width W (step S33).

[0241] Then, the cause estimation means 51 sets v=1 (step S34). v is an argument of the windows Wd_1, Wd_W+1, . . . , Wd_N-W+1 shown in FIG.

[0242] After step S34, the cause estimation means 51 selects the q-th element QoS_lan_q of the QoS / channel state information from the learning data using a window Wd_v of width W (step S35).

[0243] Then, the cause estimation means 51 calculates a difference vector between the W elements QoS_est_q and the W elements QoS_lan_q (step S36), and calculates the sum n of the elements of the calculated difference vector. q_v_T is calculated (step S37).

[0244] Thereafter, the cause estimating means 51 judges whether or not v=V (step S38). When it is judged in step S38 that v=V is not true, the cause estimating means 51 sets v=v+1 (step S39). Thereafter, the series of operations proceeds to step S35, and steps S35 to S39 are repeatedly executed until it is judged in step S38 that v=V is true.

[0245] Then, in step S38, when it is determined that v=V, the cause estimation means 51 calculates the sum n q_1_T ,n q_2_T ,n q_V_T The smallest value of n q_min_T The deviation D of the qth element QoS_est_q diver_q_T (Step S40).

[0246] Thereafter, the cause estimation means 51 judges whether or not T=Tmax (step S41), where Tmax represents the timing at which the last sample of the M samples of the distribution estimation data is acquired.

[0247] When it is determined in step S41 that T is not equal to Tmax, the cause estimation means 51 sets T to t+1 (step S42). After that, the series of operations proceeds to step S33, and steps S33 to S42 are repeatedly executed until it is determined in step S41 that T is equal to Tmax.

[0248] Then, in step S41, when it is determined that T=Tmax, the cause estimation means 51 calculates the deviation degree D diver_q_1 ~D diver_q_Tmax is retained (step S43).

[0249] Thereafter, the cause estimating means 51 judges whether or not q=Q (=5) is true (step S44).

[0250] When it is determined in step S44 that q is not equal to Q (=5), the cause estimation means 51 sets q to q+1 (step S45). After that, the series of operations proceeds to step S32, and steps S32 to S45 are repeatedly executed until it is determined in step S44 that q is equal to Q (=5).

[0251] Then, in step S44, when it is determined that q=Q (=5), the cause estimation means 51 calculates the deviation degree D diver_1_1 ~D diver_1_Tmax ;D diver_2_1 ~D diver_2_Tmax ;···;D diver_Q_1 ~D diver_Q_Tmax (Step S46). After that, the series of operations proceeds to Step S4 in FIG.

[0252] In the flowchart shown in Fig. 30, when steps S35 to S37 are executed once, one sum of the elements of the difference vector shown in Fig. 7 is calculated. Then, when steps S35 to S37 are executed V times (i.e., when the learning data window is slid from window Wd_1 to window Wd_N-W+1 in Fig. 7), V sums n q_1_T ,n q_2_T ,n q_V_T is obtained.

[0253] In the flowchart shown in FIG. 30, each time steps S33 to S40 are executed, the deviation degree D of the q-th element QoS_est_q at each timing T is calculated. diver_q_TWhen it is acquired and steps S33 to S40 are executed Tmax times, Tmax divergence degrees D diver_q_1 ~D diver_q_Tmax are acquired.

[0254] Furthermore, in the flowchart shown in FIG. 30, every time steps S32 to S43 are executed once, for one element of the QoS / channel status information shown in Table 1, Tmax divergence degrees D diver_q_1 ~D diver_q_Tmax are acquired. When steps S32 to S43 are executed Q(=5) times, for each of the Q(=5) elements of the QoS / channel status information, Tmax divergence degrees D diver_q_1 ~D diver_q_Tmax are acquired.

[0255] FIG. 31 is a flowchart for explaining the detailed operation of step S4 shown in FIG. 29.

[0256] Referring to FIG. 31, after step S3 in FIG. 29, the cause estimation means 51 sets q = 1 (step S51). Then, the cause estimation means 51 plots Tmax divergence degrees D diver_q_1 ~D diver_q_Tmax on the horizontal axis (step S52).

[0257] After that, the cause estimation means 51 associates a kernel function with each of the Tmax divergence degrees D diver_q_1 ~D diver_q_Tmax (step S53).

[0258] Subsequently, the cause estimation means 51 superimposes the kernel functions to estimate the probability density distribution PDD_q of the divergence degree (step S54).

[0259] Then, the cause estimation means 51 determines whether q = Q (step S55). When it is determined in step S55 that q ≠ Q, the cause estimation means 51 sets q = q + 1 (step S56). Thereafter, the series of operations proceeds to step S52, and steps S52 to S56 are repeatedly executed until it is determined in step S55 that q = Q.

[0260] And when it is determined in step S55 that q = Q, the series of operations proceeds to step S5 in FIG. 29.

[0261] In the flowchart shown in FIG. 31, every time steps S52 to S54 are executed once, the probability density distribution PDD of the divergence degree is estimated for one element of the QoS / channel status information shown in Table 1. When steps S52 to S54 are executed Q (= 5) times, the probability density distribution PDD of the divergence degree is estimated for all elements of the QoS / channel status information shown in Table 1. That is, all of the probability density distributions 1 to 5 of the correspondence table TBL1 shown in FIG. 11 are estimated.

[0262] FIG. 32 is a flowchart for explaining the detailed operation of step S6 shown in FIG. 29.

[0263] Referring to FIG. 32, after step S5 in FIG. 29, the cause estimation means 51 sets q = 1 (step S61).

[0264] Then, the cause estimation means 51 selects the q-th element QoS_obs_q of the QoS / channel status information from the observation data using a window of width W (step S62).

[0265] Then, the cause estimation means 51 sets v = 1 (step S63). Thereafter, the cause estimation means 51 selects the q-th element QoS_lan_q of the QoS / channel status information from the learning data using a window of width W (step S64).

[0266] Then, the cause estimation means 51 calculates a difference vector between the W elements QoS_obs_q and the W elements QoS_lan_q (step S65), and calculates the sum n' of the elements of the calculated difference vector. q_v is calculated (step S66).

[0267] Thereafter, the cause estimating means 51 judges whether or not v=V (step S67). When it is judged in step S67 that v=V is not true, the cause estimating means 51 sets v=v+1 (step S68). Thereafter, the series of operations proceeds to step S64, and steps S64 to S68 are repeatedly executed until it is judged in step S67 that v=V is true.

[0268] Then, in step S67, when it is determined that v=V, the cause estimation means 51 calculates the sum n' q_1 ,n' q_2 ,...,n' q_V The minimum value n' of q_min The deviation D' of the qth element QoS_obs_q diver_q (step S69).

[0269] Thereafter, the cause estimating means 51 judges whether or not q=Q (=5) is true (step S70).

[0270] When it is determined in step S70 that q is not equal to Q (=5), the cause estimation means 51 sets q to q+1 (step S71). After that, the series of operations proceeds to step S62, and steps S62 to S71 are repeatedly executed until it is determined in step S70 that q is equal to Q (=5).

[0271] Then, in step S70, when it is determined that q=Q (=5), the cause estimation means 51 calculates the deviation degree D' diver_1 ~D' diver_Q (Step S72). After that, the series of operations proceeds to Step S7 in FIG.

[0272] In the flowchart shown in Fig. 32, when steps S64 to S66 are executed once, one sum of the elements of the difference vector shown in Fig. 14 is calculated. Then, when steps S64 to S66 are executed V times (i.e., when the learning data window is slid from window Wd_1 to window Wd_N-W+1 in Fig. 14), V sums n' q_1 ,n' q_2 ,...,n' q_V is obtained.

[0273] In the flowchart shown in FIG. 32, each time steps S62 to S69 are executed, the deviation degree D′ of the q-th element QoS_est_q is calculated. diver_q is obtained, and when steps S62 to S69 are executed Q times, Q deviation degrees D' are obtained. diver_1 ~D diver_Q is obtained.

[0274] FIG. 33 is a flowchart for explaining the detailed operation of step S7 shown in FIG.

[0275] With reference to FIG. 33, after step S6 in FIG. 29, cause estimating means 51 sets q=1 (step S81).

[0276] Then, the cause estimation means 51 selects the probability density distribution PDD_q of the deviation degree corresponding to the q-th element QoS_lan_q of the QoS / channel state information of the learning data (step S82).

[0277] Then, the cause estimation means 51 calculates the deviation D' of the q-th element QoS_obs_q of the observation data. diver_q is the threshold value Ψ in the probability density distribution PDD_q of the deviation degree r It is determined whether or not it is equal to or greater than this (step S83).

[0278] In step S83, the deviation D' diver_q is the threshold value Ψ rIf it is determined that the value is equal to or greater than this, the cause estimation means 51 determines that the q-th element QoS_obs_q of the observation data is abnormally rising (step S84).

[0279] On the other hand, in step S83, the deviation D' diver_q is the threshold value Ψ r If it is determined that the deviation is not greater than the threshold, the cause estimation means 51 estimates the deviation D' diver_q is the threshold value Ψ in the probability density distribution PDD_q of the deviation degree d It is determined whether or not it is equal to or less (step S85).

[0280] In step S85, the deviation D' diver_q is the threshold value Ψ d If it is determined that the q-th element QoS_obs_q of the observation data is equal to or less than the threshold value, the cause estimation means 51 determines that the q-th element QoS_obs_q of the observation data is abnormally descent (step S86).

[0281] On the other hand, in step S85, the deviation D' diver_q is the threshold value Ψ d If it is determined that the result is not equal to or greater than the threshold value, the cause estimation means 51 determines that the q-th element QoS_obs_q of the observation data is not abnormal (step S87).

[0282] Then, after any one of step S84, step S86, and step S87, the cause estimating means 51 judges whether or not q=Q (step S88).

[0283] When it is determined in step S88 that q is not equal to Q, the cause estimation means 51 sets q to q+1 (step S89). After that, the series of operations proceeds to step S82, and steps S82 to S89 are repeatedly executed until it is determined in step S88 that q is equal to Q.

[0284] Then, if it is determined in step S88 that q=Q, the series of operations proceeds to step S8 in FIG.

[0285] In the flowchart shown in FIG. 33, when steps S82 to S87 are executed for the first time, the probability density distribution PDD_q (=1) (probability density distribution 1 shown in FIG. 11) of the deviation corresponding to the q (=1)-th element QoS_lan_q ("received signal strength when a packet is normally received" shown in Table 1) of the QoS / channel state information of the learning data is selected, and the deviation D' of the q (=1)-th element QoS_obs_q of the observation data is diver_q (=1) is the threshold value Ψ of the probability density distribution PDD_q(=1) r Is it greater than or equal to the threshold value Ψ d By determining whether or not the q(=1)th element QoS_obs_q of the observation data is an abnormal rise or descent, it is determined whether the q(=1)th element QoS_obs_q of the observation data is an abnormal rise or descent.

[0286] Furthermore, when steps S82 to S87 are executed for the second time, the probability density distribution PDD_q(=2) (probability density distribution 2 shown in FIG. 11) of the deviation corresponding to the q(=2)-th element QoS_lan_q ("received signal strength at the antenna end" shown in Table 1) of the QoS / channel state information of the learning data is selected, and the deviation D' of the q(=2)-th element QoS_obs_q of the observation data is selected. diver_q (=2) is the threshold value Ψ of the probability density distribution PDD_q(=2) r Is it greater than or equal to the threshold value Ψ d By determining whether or not the q(=2)th element QoS_obs_q of the observation data is an abnormal rise or an abnormal descent, it is determined whether the q(=2)th element QoS_obs_q of the observation data is an abnormal rise or an abnormal descent.

[0287] Furthermore, when steps S82 to S87 are executed for the third time, the probability density distribution PDD_q (=3) (probability density distribution 3 shown in FIG. 11) of the deviation corresponding to the q (=3)-th element QoS_lan_q ("channel occupancy rate of the communication device managed by the system at the manufacturing site" shown in Table 1) of the QoS / channel status information of the learning data is selected, and the deviation D' of the q (=3)-th element QoS_obs_q of the observation data is diver_q (=3) is the threshold Ψ of the probability density distribution PDD_q(=3)r Is it greater than or equal to the threshold value Ψ d By determining whether or not the q(=3)th element QoS_obs_q of the observation data is an abnormal rise or an abnormal descent, it is determined whether the q(=3)th element QoS_obs_q of the observation data is an abnormal rise or an abnormal descent.

[0288] Furthermore, when steps S82 to S87 are executed for the fourth time, the probability density distribution PDD_q (=4) (probability density distribution 4 shown in FIG. 11) of the deviation corresponding to the q (=4)-th element QoS_lan_q ("channel occupancy rate of communication devices not managed by the manufacturing site system" shown in Table 1) of the QoS / channel status information of the learning data is selected, and the deviation D' of the q (=4)-th element QoS_obs_q of the observation data is selected. diver_q (=4) is the threshold Ψ of the probability density distribution PDD_q(=4) r Is it greater than or equal to the threshold value Ψ d By determining whether or not the q(=4)th element QoS_obs_q of the observation data is an abnormal rise or descent, it is determined whether the q(=4)th element QoS_obs_q of the observation data is an abnormal rise or descent.

[0289] Furthermore, when steps S82 to S87 are executed for the fifth time, the probability density distribution PDD_q (=5) (probability density distribution 5 shown in FIG. 11) of the deviation corresponding to the q (=5)-th element QoS_lan_q ("Throughput in the transport layer" shown in Table 1) of the QoS / channel state information of the learning data is selected, and the deviation D' of the q (=5)-th element QoS_obs_q of the observation data is diver_q (=5) is the threshold Ψ of the probability density distribution PDD_q(=5) r Is it greater than or equal to the threshold value Ψ d By determining whether or not the q(=5)th element QoS_obs_q of the observation data is an abnormal rise or an abnormal descent, it is determined whether the q(=5)th element QoS_obs_q of the observation data is an abnormal rise or an abnormal descent.

[0290] FIG. 34 is a flowchart for explaining the detailed operation of step S8 shown in FIG.

[0291] 34, after step S7 in FIG. 29, the cause estimation means 51 judges whether or not all communication devices in the communication system 100 have received packets from any communication device during a certain period of time (step S91). In this case, the cause estimation means 51 judges whether or not all communication devices in the communication system 100 have received packets from any communication device during a certain period of time by referring to the correspondence table TBL2 shown in FIG. 12. Then, when "no" is stored in the "whether or not packets have been received during a certain period of time" column corresponding to all of the IP addresses Add_AP_1, Add_AP_2,...,Add_AP_I; Add_CM_1, Add_CM_2,...,Add_CM_J; Add_MN_1, Add_MN_2,...,Add_MN_M in the correspondence table TBL2, the cause estimation means 51 judges that all communication devices in the communication system 100 have not received packets from any communication device during a certain period of time. On the other hand, when "yes" is stored in the "Whether packets have been received in a certain period of time" column corresponding to at least one of the IP addresses Add_AP_1, Add_AP_2, ···, Add_AP_I; Add_CM_1, Add_CM_2, ···, Add_CM_J; Add_MN_1, Add_MN_2, ···, Add_MN_M in correspondence table TBL2, the cause estimation means 51 determines that packets have been received from any communication device in at least one communication device of communication system 100 in a certain period of time.

[0292] When it is determined in step S91 that no packets have been received from any communication device in all communication devices of the communication system 100 for a certain period of time, the cause estimation means 51 estimates that the given communication device has a failure (step S92).

[0293] On the other hand, when it is determined in step S91 that at least one communication device in the communication system 100 has received a packet from any communication device during a certain period of time, the cause estimation means 51 determines whether the received signal strength RSSI in one or more base stations is abnormally descent or abnormally rises (step S93).

[0294] In this case, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13 and enters the IP address Add_AP_i of one or more base stations in the column of "Base station / monitor terminal device that detected QoS / channel status information" corresponding to one of the received signal strengths RSSI_1, RSSI_2, . . . _RSSI is stored, and the IP address of one or more base stations Add_AP_i _RSSI The deviation D' corresponding to diver (D in the correspondence table TBL3 diver_1_RSSI ,D diver_2_RSSI , ...) is abnormally descent or abnormally rises, it is determined that the received signal strength RSSI is abnormally descent or abnormally rises in one or more base stations.

[0295] On the other hand, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13 and enters the IP address Add_AP_i of one or more base stations in the column of "Base station / monitor terminal device that detected QoS / channel status information" corresponding to one of the received signal strengths RSSI_1, RSSI_2, . . . _RSSI is stored, and the IP address of one or more base stations Add_AP_i _RSSI The deviation D' corresponding to diver (D in the correspondence table TBL3 diver_1_RSSI ,D diver_2_RSSI,...) is neither abnormally descent nor abnormally rise, it is determined that the state is not one in which "the received signal strength RSSI is abnormally descent or abnormally rise at one or more base stations" (the received signal strength RSSI is neither abnormally descent nor abnormally rise at any base station).

[0296] 13, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13 and enters the IP address Add_AP_i of one or more base stations in the column of "Base station / monitor terminal device that detected QoS / channel status information" corresponding to one of the received signal strengths RSSI_1, RSSI_2, . . . _RSSI When the IP address of one or more base stations Add_AP_i is not stored, _RSSI The deviation D' corresponding to diver (D in the correspondence table TBL3 diver_1_RSSI ,D diver_2_RSSI If the received signal strength RSSI (any of (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 6

[0297] In step S93, when it is determined that the received signal strength RSSI in one or more base stations is abnormally descent or abnormally rises, the cause estimation means 51 determines whether the received signal strength RSSI for multiple communication devices is abnormally descent or abnormally rises (step S94).

[0298] In this case, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13, and in step S93, selects the IP addresses Add_AP_i of one or more base stations whose received signal strength RSSI is determined to be abnormally descending (descent) or abnormally rising (rise). _RSSIIn response to the above, multiple IP addresses of multiple communication devices are stored in the "Communication device that transmitted the packet that was the source of the detection of QoS / channel status information" field, and the deviation D' diver (D in the correspondence table TBL3 diver_1_RSSI ,D diver_2_RSSI When two or more of the communication devices (any of the communication devices 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 6

[0299] On the other hand, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13 and in step S93, selects the IP address Add_AP_i of one or more base stations whose received signal strength RSSI is determined to be abnormally descending (descent) or abnormally rising (rise). _RSSI In response to the above, multiple IP addresses of multiple communication devices are stored in the "Communication device that transmitted the packet that was the source of the detection of QoS / channel status information" field, and the deviation D' diver (D in the correspondence table TBL3 diver_1_RSSI ,D diver_2_RSSI ,...) is abnormally descending or abnormally rising for only one communication device, the cause estimation means 51 determines that the state is not one in which "the received signal strength RSSI for multiple communication devices is abnormally descending or abnormally rising" (=there is only one communication device in which the received signal strength RSSI is abnormally descending or abnormally rising). Also, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13 and in step S93, calculates the IP addresses Add_AP_i of one or more base stations determined to be abnormally descending or abnormally rising for the received signal strength RSSI. _RSSI In response to the above, when multiple IP addresses of multiple communication devices are not stored in the "Communication device that sent the packet that was the source of the QoS / channel status information detection" field, or when the deviation D'diver (D in the correspondence table TBL3 diver_1_RSSI ,D diver_2_RSSI When there is only one communication device in which the received signal strength RSSI for multiple communication devices is abnormally descent or abnormally rises, it is determined that the state is not one in which "the received signal strength RSSI for multiple communication devices is abnormally descent or abnormally rises."

[0300] When it is determined in step S94 that the received signal strengths RSSI for the multiple communication devices are abnormally descending or rising, the cause estimation means 51 estimates that the cause of the communication abnormality is the installation of an obstacle (step S95).

[0301] On the other hand, when it is determined in step S94 that the received signal strength RSSI for multiple communication devices is not in an abnormal descent or rise state, the cause estimation means 51 estimates that the cause of the communication abnormality is a change in the installation location of the communication device (step S96).

[0302] On the other hand, when it is determined in step S93 that the state is not one in which "the received signal strength RSSI is abnormally descent or abnormally rising in one or more base stations," the cause estimation means 51 determines whether the channel occupancy rate of the communication devices managed by the communication system 100 is abnormally descent in one or more base stations / monitor terminal devices (step S97).

[0303] In this case, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13, and in step S93, the IP addresses Add_AP_i of one or more base stations used in the judgment are _RSSI (=Add_AP_i _CH ) along with the IP address of one or more monitor terminals Add_MN_m _CHis stored in the "Base station / Monitor terminal device that detected QoS / Channel status information" column in correspondence with the channel occupancy rate CH_ocp, and the deviation D' diver (D in the correspondence table TBL3 diver_1_CH ,D diver_2_CH , . . . ) is abnormally descent, the channel occupancy rate of the communication device managed by the communication system 100 is determined to be abnormally descent.

[0304] On the other hand, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13 and in step S93, the IP addresses Add_AP_i of one or more base stations used in the judgment are _RSSI (=Add_AP_i _CH ) along with the IP address of the monitor terminal Add_MN_m _CH is stored in the "Base station / Monitor terminal device that detected QoS / Channel status information" column in correspondence with the channel occupancy rate CH_ocp, and the deviation D' diver (D in the correspondence table TBL3 diver_1_CH ,D diver_2_CH , . . .) is not abnormally descent, the cause estimation means 51 determines that the channel occupancy rate of the communication device managed by the communication system 100 is not abnormally descent. Also, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13 and in step S93, determines that the IP addresses Add_AP_i _RSSI (=Add_AP_i _CH ) and the IP address of the monitor terminal device Add_MN_m _CH is not stored in the "Base station / Monitor terminal device that detected QoS / Channel status information" column in correspondence with the channel occupancy rate CH_ocp, or the deviation D' diver (D in the correspondence table TBL3 diver_1_CH ,D diver_2_CH, . . . ) is not abnormally descent, it is determined that the channel occupancy rates of the communication devices managed by the communication system 100 are not abnormally descent.

[0305] When it is determined in step S97 that the channel occupancy rate of the communication device is not abnormally descent in all of the one or more base stations / monitor terminal devices, the cause estimation means 51 estimates that there is no abnormality (step S98).

[0306] On the other hand, when it is determined in step S97 that the channel occupancy rate of the communication device managed by the communication system 100 is abnormally descent in one or more base stations / monitor terminal devices, the cause estimation means 51 determines whether the throughput THP is abnormally rising in one or more base stations (step S99).

[0307] In this case, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13, and in step S97, the IP addresses Add_AP_i of one or more base stations used in the judgment are _CH Same IP address as Add_AP_i _CH’ is stored in the "Base station / monitor terminal device that detected QoS / channel status information" column, and the IP address Add_AP_i of one or more base stations is _CH’ The deviation D' corresponding to diver (D in the correspondence table TBL3 diver_1_THP ,D diver_2_THP , . . . ) is abnormally rising, it is determined that the throughput THP is abnormally rising in one or more base stations.

[0308] On the other hand, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13 and in step S97, the IP addresses Add_AP_i of one or more base stations used in the judgment are _CH Same IP address as Add_AP_i _CH’is stored in the "Base station / monitor terminal device that detected QoS / channel status information" column, and the IP address Add_AP_i of one or more base stations is _CH’ The deviation D' corresponding to diver (D in the correspondence table TBL3 diver_1_THP ,D diver_2_THP , ...) is not an abnormal rise, it is determined that the state is not one in which "the throughput THP is abnormally rising at one or more base stations" (= the throughput THP is not abnormally rising at all base stations).

[0309] 13, in step S97, the cause estimation means 51 determines the IP addresses Add_AP_i of one or more base stations used in the determination. _CH Same IP address as Add_AP_i _CH’ When the IP address of one or more base stations Add_AP_i is not stored in the "Base station / monitor terminal device that detected QoS / channel status information" column, _CH’ The deviation D' corresponding to diver (D in the correspondence table TBL3 diver_1_THP ,D diver_2_THP , . . . ) is not abnormally rising, it is determined that the state is not one in which "the throughput THP is abnormally rising in one or more base stations."

[0310] In step S99, when it is determined that the throughput THP is abnormally rising in one or more base stations, the cause estimation means 51 estimates that the cause of the communication abnormality is an increase in traffic of the communication device (step S100). On the other hand, when it is determined in step S99 that the throughput THP is not abnormally rising in all base stations, the cause estimation means 51 determines whether or not the channel occupancy rate of a communication device not managed by the communication system 100 is abnormally rising in one or more base stations / monitor terminal devices (step S101).

[0311] In this case, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13, and in step S99, the IP addresses Add_AP_i of one or more base stations used in the judgment are _THP Same IP address as Add_AP_i _CH’ Along with one or more monitor terminal IP addresses Add_MN_m _CH’ is stored in the "Base station / monitor terminal device that detected QoS / channel status information" column, and the IP address Add_AP_i of one or more base stations / monitor terminal devices is _CH’ / IP Address Add_MN_m _CH’ The deviation D' corresponding to diver (D' in the correspondence table TBL3 diver_1_CH ,D' diver_2_CH , . . . ) is abnormally rising, one or more base stations / monitor terminal devices determine that the channel occupancy rate of a communication device not managed by the communication system 100 is abnormally rising.

[0312] On the other hand, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13 and in step S99, the IP addresses Add_AP_i of one or more base stations used in the judgment are _THP Same IP address as Add_AP_i _CH’ Along with one or more monitor terminal IP addresses Add_MN_m _CH’ is stored in the "Base station / monitor terminal device that detected QoS / channel status information" column, and the IP address Add_AP_i of one or more base stations / monitor terminal devices is _CH’ / IP Address Add_MN_m _CH’ The deviation D' corresponding to diver (D' in the correspondence table TBL3 diver_1_CH ,D' diver_2_CH , ...) is not abnormally rising, it is determined that the state is not one in which "in one or more base stations / monitor terminal devices, the channel occupancy rates of communication devices not managed by the communication system 100 are abnormally rising" (=in all base stations / monitor terminal devices, the channel occupancy rates of communication devices not managed by the communication system 100 are not abnormally rising).

[0313] Further, the cause estimation means 51 refers to the correspondence table TBL3 shown in FIG. 13, and in step S99, for the IP address Add_AP_i of one or more base stations used in the determination _THP the same IP address Add_AP_i _CH’ and the IP address Add_MN_m of one or more monitor terminal devices _CH’ when they are not stored in the column of "Base station / monitor terminal device that detected QoS / channel status information", or for the IP addresses Add_AP_i of one or more base stations / monitor terminal devices _CH’ / IP address Add_MN_m _CH’ if the degree of deviation D' diver (any one of D' in the correspondence table TBL3 diver_1_CH , D' diver_2_CH , ···) are not all abnormally rising (rise), it is determined that the state where "in one or more base stations / monitor terminal devices, the channel occupancy rate of communication devices not managed by the communication system 100 is abnormally rising (rise)" does not exist.

[0314] In step S101, when it is determined that in one or more base stations / monitor terminal devices, the channel occupancy rate of communication devices not managed by the communication system 100 is abnormally rising (rise), the cause estimation means 51 estimates that the cause of the communication abnormality is brought in from outside the communication device (step S102).

[0315] On the other hand, in step S101, when it is determined that in all base stations / monitor terminal devices, the channel occupancy rate of communication devices not managed by the communication system 100 is not abnormally rising (rise), the cause estimation means 51 estimates that the cause of the communication abnormality is noise from the installed devices at the manufacturing site (step S103).

[0316] And after any one of step S92, step S95, step S96, step S98, step S100, step S102, and step S103, the series of operations proceeds to step S9 in FIG. 29.

[0317] 34, in step S91, determining that at least one communication device in the communication system 100 has received a packet from an arbitrary communication device during a certain period of time corresponds to determining that the arbitrary communication device managed by the communication system 100 is not malfunctioning. Here, when it is determined that all communication devices are not malfunctioning, the cause estimation means 51 determines that all communication devices in the communication system 100 are operating normally.

[0318] Then, after confirming that all communication devices in the communication system 100 are operating normally, the cause estimating means 51 executes steps S92 to S103 to estimate the cause of the communication abnormality.

[0319] In the flowchart shown in Fig. 34, steps S94 to S96 are steps for estimating the cause of the communication abnormality caused by a decrease in received signal strength RSSI, and steps S97 to S103 are steps for estimating the cause of the communication abnormality caused by a decrease in throughput THP.

[0320] Therefore, when the cause estimation means 51 determines in step S93 that the received signal strength RSSI is descent or rise in one or more base stations, it estimates that the cause of the communication abnormality is due to a decrease or increase in the received signal strength RSSI, and when the cause estimation means 51 determines in step S93 that the received signal strength RSSI is neither descent nor rise in all of the one or more base stations, it estimates that the cause of the communication abnormality is due to a decrease in throughput.

[0321] When an obstacle is installed, the received signal strength RSSI of packets transmitted and received in the link that sandwiches the obstacle decreases or increases. Also, when the installation location of a communication device is changed, the received signal strength RSSI of packets received by the base station for the communication device whose installation location has been changed decreases or increases due to the influence of multipath fading.

[0322] On the other hand, an increase in traffic of a communication device managed by the communication system 100 or noise from equipment installed at a manufacturing site does not cause the received signal strength RSSI of packets received by surrounding communication devices to decrease or increase.

[0323] Therefore, when it is determined in step S93 that the received signal strength RSSI is descent or rise in one or more base stations, it is presumed that the cause of the communication abnormality is a decrease or increase in the received signal strength RSSI, and when it is determined in step S93 that the received signal strength RSSI is neither descent nor rise in all base stations, it is presumed that the cause of the communication abnormality is a decrease in throughput.

[0324] In addition, when an obstacle is installed, the received signal strength RSSI decreases or increases in multiple links, whereas when the installation position of a communication device is changed, the received signal strength RSSI decreases or increases in only a single link.

[0325] Therefore, in the flowchart shown in FIG. 34, when it is determined in step S94 that the received signal strength RSSI for multiple communication devices is abnormally descent or abnormally rises, it is presumed that the cause of the communication abnormality is the installation of an obstacle, and when it is determined that there is only a single communication device with an abnormally descent or abnormally rises in the received signal strength RSSI, it is presumed that the cause of the communication abnormality is a change in the installation location of the communication device.

[0326] Furthermore, an increase in traffic of a communication device managed by the communication system 100 reduces the channel occupancy rate of other communication devices managed by the communication system 100. Also, the introduction of a communication device not managed by the communication system 100 reduces the channel occupancy rate of the communication devices managed by the communication system 100. Furthermore, noise from equipment installed at the manufacturing site reduces the channel occupancy rate of the communication devices managed by the communication system 100. When no communication abnormality occurs, the channel occupancy rate does not change.

[0327] Therefore, in the flowchart shown in FIG. 34, when it is determined in step S97 in one or more base stations / monitor terminal devices that the channel occupancy rate of the communication devices managed by the communication system 100 is abnormally descent, it is presumed that the cause of the communication abnormality is an increase in traffic of the communication device or an external factor brought into the communication device, and when it is determined in step S97 in all base stations / monitor terminal devices that the channel occupancy rate of the communication devices managed by the communication system 100 is not abnormally descent, it is presumed that there is no abnormality.

[0328] Furthermore, when the traffic of a communication device managed by the communication system 100 increases, the throughput of the corresponding base station increases. On the other hand, the throughput of the base station does not increase when the traffic of a communication device not managed by the communication system 100 increases or when there is noise from equipment installed at a manufacturing site.

[0329] Therefore, in the flowchart shown in FIG. 34, when it is determined in step S99 that the throughput is abnormally rising in one or more base stations, it is presumed that the cause of the communication abnormality is an increase in traffic of the communication device, and when it is determined in step S99 that the throughput is not abnormally rising in all base stations, it is presumed that the cause of the communication abnormality is noise brought in from outside the communication device or noise from equipment installed at the manufacturing site.

[0330] Furthermore, when a communication device not managed by the communication system 100 is brought in and performs communication, the channel occupancy rate of the communication device not managed by the communication system 100 observed by the base station / monitor terminal device increases. On the other hand, since noise from the equipment installed at the manufacturing site cannot be decoded, the channel occupancy rate of the communication device not managed by the communication system 100 observed by the base station / monitor terminal device does not increase.

[0331] Therefore, in the flowchart shown in FIG. 34, when it is determined in step S101 in one or more base stations / monitor terminal devices that the channel occupancy rate of a communication device not managed by the communication system 100 is abnormally rising, it is presumed that the cause of the communication abnormality is an external communication device brought in, and when it is determined in step S101 in all base stations / monitor terminal devices that the channel occupancy rate of a communication device not managed by the communication system 100 is not abnormally rising, it is presumed that the cause of the communication abnormality is noise from equipment installed at the manufacturing site.

[0332] FIG. 35 is a flowchart for explaining the detailed operation of step S12 shown in FIG.

[0333] 35, after step S11 in FIG. 29, the position estimation means 52 determines whether or not the cause of the communication abnormality is the installation of an obstacle (step S121).

[0334] In step S121, when it is determined that the cause of the communication abnormality is the installation of an obstacle, the position estimation means 52 refers to the correspondence table TBL4, selects the position estimation method 1 corresponding to “installation of an obstacle” (step S122), and estimates the occurrence position of the cause of the communication abnormality using the selected position estimation method 1 (step S123).

[0335] On the other hand, when it is determined in step S121 that the cause of the communication abnormality is not the installation of an obstacle, the position estimation means 52 determines whether or not the cause of the communication abnormality is a change in the installation position of the communication device (step S124).

[0336] In step S124, when it is determined that the cause of the communication abnormality is a change in the installation location of the communication device, the location estimation means 52 refers to the correspondence table TBL4, selects the location estimation method 2 corresponding to the “change in the installation location of the communication device” (step S125), and estimates the location where the cause of the communication abnormality occurred using the selected location estimation method 2 (step S126).

[0337] On the other hand, when it is determined in step S124 that the cause of the communication abnormality is not a change in the installation location of the communication device, the location estimation means 52 determines whether or not the cause of the communication abnormality is an increase in traffic of the communication device (step S127).

[0338] In step S127, when it is determined that the cause of the communication abnormality is an increase in traffic, the location estimation means 52 refers to the correspondence table TBL4, selects the location estimation method 3 corresponding to “increase in traffic” (step S128), and estimates the location where the cause of the communication abnormality has occurred using the selected location estimation method 3 (step S129).

[0339] On the other hand, when it is determined in step S127 that the cause of the communication abnormality is not an increase in traffic, the position estimation means 52 determines whether or not the cause of the communication abnormality is due to an external device being brought in (step S130).

[0340] In step S130, when it is determined that the cause of the communication abnormality is brought in from outside the communication device, the location estimation means 52 refers to the correspondence table TBL4, selects the location estimation method 4 corresponding to "brought in from outside the communication device" (step S131), and estimates the location where the cause of the communication abnormality occurred using the selected location estimation method 4 (step S132).

[0341] On the other hand, when it is determined in step S130 that the cause of the communication abnormality is not brought in from outside the communication device, the location estimation means 52 determines that the cause of the communication abnormality is noise from equipment installed at the manufacturing site, refers to correspondence table TBL4, selects location estimation method 5 corresponding to "noise from equipment installed at the manufacturing site" (step S133), and estimates the occurrence location of the cause of the communication abnormality using the selected location estimation method 5 (step S134).

[0342] Then, after any one of step S123, step S126, step S129, step S132, and step S134, the series of operations proceeds to step S13 in FIG.

[0343] FIG. 36 is a flowchart for explaining the detailed operation of step S123 shown in FIG.

[0344] 36, after step S122 in FIG. 35, the position estimation means 52 calculates the rarity of the received signal strength RSSI of the observation data at time t (step S123-1).

[0345] Then, the position estimation means 52 calculates the average value of the received signal strength RSSI of the learning data at time t (step S123-2).

[0346] Thereafter, the position estimation means 52 normalizes the rarity of the received signal strength RSSI of the observation data (step S123-3).

[0347] Next, the position estimation means 52 normalizes the average value of the received signal strength RSSI of the learning data (step S123-4).

[0348] Then, the position estimation means 52 multiplies the normalized value of the degree of abnormality (Rarity) by the normalized value of the average value of the received signal strength RSSI, and obtains the multiplication results for all of the multiple monitor terminal devices (step S123-5).

[0349] Then, the position estimation means 52 selects N 2 links are detected (step S123-6), and the detected N 2 For each link, the midpoint of the link, or the midpoint of the midpoint weighted by the multiplication result of the normalized value of the degree of anomaly (Rarity) and the normalized value of the average value of the received signal strength RSSI, or the center of gravity of the midpoint weighted by the multiplication result of the normalized value of the degree of anomaly (Rarity) and the normalized value of the average value of the received signal strength RSSI is estimated as the occurrence position of the cause of the anomaly (installation of an obstacle) (step S123-7). After that, the series of operations proceeds to step S13 in FIG. 29.

[0350] FIG. 37 is a flowchart for explaining the detailed operation of step S126 shown in FIG.

[0351] 37, after step S125 in FIG. 35, the position estimation means 52 calculates the degree of abnormality (Rarity) of the received signal strength RSSI of the observation data for a plurality of links (step S126-1).

[0352] Then, the position estimation means 52 detects the link having the highest rarity of the received signal strength RSSI from among the multiple links (step S126-2).

[0353] Then, the position estimation means 52 estimates that the initial position of the communication device 30 of the link transmission source detected in step S126-2 is the occurrence position of the cause of the abnormality (change in the installation position of the communication device) (step S126-3). After that, the series of operations proceeds to step S13 in FIG.

[0354] FIG. 38 is a flowchart for explaining the detailed operation of step S129 shown in FIG.

[0355] 38, after step S128 in FIG. 35, the position estimation means 52 calculates the degree of abnormality (Rarity) of the throughput THP of the observation data for a plurality of links (step S129-1).

[0356] Then, the position estimation means 52 detects the link having the highest rarity of the throughput THP from among the multiple links (step S129-2).

[0357] Thereafter, the position estimation means 52 calculates the position PS_CM of the source communication device 30 based on the received signal strength RSSI when a packet is normally received from the source communication device 30 of the link detected in step S129-2 (step S129-3).

[0358] Then, the position estimation means 52 estimates that the calculated position PS_CM of the source communication device 30 is the position where the cause of the abnormality (increase in traffic) occurs (step S129-4). After that, the series of operations proceeds to step S13 in FIG.

[0359] FIG. 39 is a flowchart for explaining the detailed operation of step S129-3 shown in FIG.

[0360] Referring to FIG. 39, after step S129-2 in FIG. 38, the position estimation means 52 detects M received signal strengths RSSI_1 to RSSI_M when M monitor terminal devices MN_1 to MN_M actually receive packets from the communication device 30 (step S201).

[0361] Then, the position estimation means 52 sets s=1 (step S202) and sets m=1 (step S203).

[0362] Then, when the communication device 30 is placed at a candidate point CD_s consisting of an intersection of a grid pattern, the position estimation means 52 estimates the received signal strength RSSI_s_m when the monitor terminal device MN_m receives a packet from the communication device 30 using a radio wave propagation model (equation (1)) (step S204).

[0363] Then, the position estimation means 52 judges whether m=M or not (step S205). When it is judged in step S205 that m=M is ​​not true, the position estimation means 52 sets m=m+1 (step S206). After that, the series of operations proceeds to step S204, and steps S204 to S206 are repeatedly executed until it is judged in step S205 that m=M is ​​true.

[0364] If it is determined in step S205 that m=M, the position estimation means 52 calculates the error ERR_s between the received signal strengths RSSI_s_1 to RSSI_s_M and the received signal strengths RSSI_1 to RSSI_M (step S207).

[0365] Thereafter, the position estimation means 52 judges whether or not s=S (step S208). When it is judged in step S208 that s=S is not true, the position estimation means 52 sets s=s+1 (step S209). Thereafter, the series of operations proceeds to step S203, and steps S203 to S209 are repeatedly executed until it is judged in step S208 that s=S is true.

[0366] Then, when it is determined in step S208 that s=S, the position estimation means 52 estimates the candidate point CD_min at which the smallest error ERR_min is obtained among the S errors ERR_1 to ERR_S as the position PS_CM of the communication device 30 (step S210). After that, the series of operations proceeds to step S129-4 in FIG.

[0367] FIG. 40 is another flowchart for explaining the detailed operation of step S129-3 shown in FIG.

[0368] Referring to FIG. 40, after step S129-2 in FIG. 38, the position estimation means 52 detects M received signal strengths RSSI_1 to RSSI_M when M monitor terminal devices MN_1 to MN_M actually receive packets from the communication device 30 (step S211).

[0369] Then, the position estimation means 52 calculates an average value RSSI_ave of the M received signal strengths RSSI_1 to RSSI_M (step S212).

[0370] Thereafter, the position estimation means 52 calculates an error vector VC between the M received signal strengths RSSI_1 to RSSI_M and the average value RSSI_ave. ERR is calculated (step S213).

[0371] Then, the position estimation means 52 sets s = 1 (step S214), and when the communication device 30 is placed at a candidate point CD_s consisting of an intersection of the grid pattern, estimates the received signal strengths RSSI_s_1 to RSSI_s_M when the monitor terminal devices MN_1 to MN_M receive packets from the communication device 30 using a radio wave propagation model (equation (1)) (step S215).

[0372] Thereafter, the position estimation means 52 calculates an error vector VC between the received signal strengths RSSI_s_1 to RSSI_s_M and the average value RSSI_ave. ERR_s is calculated (step S216).

[0373] Then, the position estimation means 52 judges whether or not s=S (step S217). When it is judged in step S217 that s=S is not true, the position estimation means 52 sets s=s+1 (step S218). After that, the series of operations proceeds to step S215, and steps S215 to S218 are repeatedly executed until it is judged in step S217 that s=S is true.

[0374] Then, in step S217, when it is determined that s=S, the position estimation means 52 calculates the error vector V ERR and S error vectors VC ERR_1 ~VC ERR_S Then, S errors ERR_1 to ERR_S between the input and output are calculated (step S219).

[0375] Then, the position estimation means 52 estimates the candidate point CD_min at which the smallest error ERR_min is obtained among the S errors ERR_1 to ERR_S as the position PS_CM of the communication device 30 (step S220). After that, the series of operations proceeds to step S129-4 in FIG.

[0376] FIG. 41 is yet another flowchart for explaining the detailed operation of step S129-3 shown in FIG.

[0377] 41, after step S129-2 in FIG. 38, the position estimation means 52 selects three monitor terminal devices MN_1 to MN_3 arranged in the vicinity of the link having the highest abnormality degree (Rarity) of the throughput THP (step S221).

[0378] Then, the position estimation means 52 sets the received signal strengths acquired by the three monitor terminal devices MN_1 to MN_3 as RSSI_1 to RSSI_3 (step S222).

[0379] Thereafter, the position estimation means 52 estimates the positions of the monitor terminal devices MN_1 to MN_3 as (x 1 ,y 1 ),x 2 ,y 2 ),x 3 ,y 3 ) (step S223), and the position of the communication device 30 to be estimated is set to (x, y) (step S224).

[0380] Then, the position estimation means 52 sets m=1 (step S225) and calculates the position (x m ,y m ) and the distance D between position (x, y) m is estimated using the received signal strength RSSI_m and the radio wave propagation model (Equation (1)) (step S226).

[0381] Then, the position estimation means 52 judges whether or not m=M (step S227). When it is judged in step S227 that m=M is ​​not true, the position estimation means 52 sets m=m+1 (step S228). After that, the series of operations proceeds to step S226, and steps S226 to S228 are repeatedly executed until it is judged in step S227 that m=M is ​​true.

[0382] Then, in step S227, when it is determined that m=M, the position estimation means 52 calculates the position (x 1 ,y 1 ),x 2 ,y 2 ),x 3 ,y 3 ) and distance D 1 ,D 2 ,D 3 into equation (2) and solve the simultaneous equations to calculate the position (x, y) (step S229).

[0383] Then, the position counting means 52 estimates the calculated position (x, y) as the position PS_CM of the communication device 30 (step S230). After that, the series of operations proceeds to step S129-4 in FIG.

[0384] The flowcharts shown in Figs. 39 to 41 are common in that the position PS_CM of a communication device is estimated based on the received signal strength RSSI at the monitor terminal device when a packet is received from the communication device, the position of which is to be estimated.

[0385] The flowcharts shown in Figures 39 and 40 have in common the point that the received signal strength RSSI when a packet is received from the communication device whose position is to be estimated is estimated using a radio wave propagation model (Equation (1)), and the candidate point of the communication device when the error between the estimated received signal strength RSSI and the received signal strength RSSI when the packet is actually received is minimized is estimated to be the position PS_CM of the communication device.

[0386] In addition, in the flowchart shown in Figure 41, the distance between the communication device, the location of which is to be estimated, and the monitor terminal device is estimated using a radio wave propagation model (equation (1)) based on the received signal strength RSSI when the monitor terminal device actually receives a packet from the communication device, the location of which is to be estimated, and the location PS_CM of the communication device is estimated by three-point positioning using the estimated distance.

[0387] Therefore, according to the flowcharts shown in FIGS. 39 to 41, step S129-3 in FIG. 38 is to estimate the received signal strength RSSI when a packet is received from the communication device whose position is to be estimated or the distance D between the communication device whose position is to be estimated and the monitor terminal device. m is estimated using the radio wave propagation model (Equation (1)), and the estimated received signal strength RSSI or distance D m This is done by estimating the position PS_CM of the communication device using

[0388] FIG. 42 is a flowchart for explaining the detailed operation of step S132 shown in FIG.

[0389] Referring to Figure 42, after step S131 in Figure 35, the position estimation means 52 acquires M received signal intensities RSSI_1 to RSSI_M detected by the M monitor terminal devices MN_1 to MN_M arranged in a grid pattern when the M monitor terminal devices MN_1 to MN_M successfully receive a packet from a communication device CM_EXT brought in from outside (step S132-1).

[0390] Then, the position estimation means 52 estimates the position of the communication device CM_EXT estimated according to the flowchart shown in Fig. 38 (including any of the flowcharts shown in Fig. 39 to Fig. 41) based on the M received signal strengths RSSI_1 to RSSI_M as the occurrence position of the cause of the abnormality (brought in from outside the communication device) (step S132-2). After that, the series of operations proceeds to step S13 in Fig. 29.

[0391] FIG. 43 is a flowchart for explaining the detailed operation of step S134 shown in FIG.

[0392] 43, after step S133 in FIG. 35, the position estimation means 52 calculates a plurality of rarities of the plurality of received signal strengths RSSI' observed at the antenna ends by the plurality of monitor terminal devices 50 (step S134-1).

[0393] Then, the position estimation means 52 estimates N 3 N of monitor terminal devices 50 3 Position of each item (x 1 ,y 1 )~(x N3 ,y N3 ) is obtained (step S134-2).

[0394] Thereafter, the position estimation means 52 3 Position of each item (x 1 ,y 1 )~(x N3 ,y N3 ) and N 3 Individual abnormality R abn_1 ~R abn_N3 Based on this, the abnormality R abn N with weights 3 Position of each item (x 1 ,y 1 )~(x N3 ,y N3 ) or the abnormality R abn N with weights 3 The midpoint of N positions, or 3 Indicates the position of (x G ,y G ) is calculated using equation (6) (step S134-3).

[0395] Then, the position estimation means 52 calculates the calculated center of gravity (x G ,y G ) is estimated to be the location of the cause of the abnormality (noise from equipment installed at the manufacturing site) (step S134-4). After that, the series of operations proceeds to step S13 in FIG.

[0396] As described above, the estimation device 10 estimates the cause of the communication anomaly and the location where the cause of the anomaly occurs, and therefore it is possible to deal with the cause of the communication anomaly. The location estimation means 52 estimates the location where the cause of the anomaly occurs using a location estimation method associated with the cause of the communication anomaly estimated by the cause estimation means 51, and therefore it is possible to accurately estimate the location where the cause of the anomaly occurs. As a result, it is possible to deal with the cause of the communication anomaly more quickly and accurately.

[0397] In the embodiment of the present invention, the operation of the estimation device 10 may be realized by software. In this case, the estimation device 10 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), and a RAM (Random Access Memory). The ROM stores a program Prog_A including each step of the flowchart shown in FIG. 29 (including the flowcharts shown in FIG. 30 to FIG. 34, the flowchart shown in FIG. 35 (including the flowchart shown in FIG. 36, the flowchart shown in FIG. 37, and the flowchart shown in FIG. 38 (including any of the flowcharts shown in FIG. 39 to FIG. 41)), the flowchart shown in FIG. 42, and the flowchart shown in FIG. 43).

[0398] The CPU reads out the program Prog_A from the ROM and executes the read out program Prog_A to estimate the cause of the communication anomaly and the location where the anomaly has occurred. The RAM temporarily stores the deviation, the probability density distribution of the deviation, the determined abnormal rise and fall, the cause of the communication anomaly, etc.

[0399] Furthermore, the program Prog_A may be distributed in a form recorded on a recording medium such as a CD or a DVD. When the recording medium on which the program Prog_A is recorded is attached to a computer, the computer reads out the program Prog_A from the recording medium, executes the program, and estimates the cause of the communication abnormality and the location where the abnormality has occurred.

[0400] Therefore, the recording medium on which the program Prog_A is recorded is a computer-readable recording medium.

[0401] [simulation] A simulation was performed on the estimation device 10 for estimating the cause of a communication anomaly and the location where the cause of the anomaly has occurred.

[0402] In the simulation, in order to verify whether each estimation method was functioning correctly, we set the assumption that the cause estimation results of the location estimation method were always correct, and evaluated the performance of each estimation method independently.

[0403] The simulation specifications are shown in Table 2.

[0404] [Table 2]

[0405] FIG. 44 is a diagram showing the arrangement of the communication device 30, the base station 40, and the monitor terminal device 50 in the vehicle body inspection factory used in the simulation.

[0406] The size of the factory is 30m x 40m, and we assume that image inspection is performed on the body frames moving on the line. The following equipment is installed in the factory. Vehicle inspection camera Network camera ·Sensor for measuring chain elongation in lines These connect to a base station via IEEE802.11g wireless LAN using the same wireless channel, and from there communicate with the application server via a wired connection.

[0407] The body inspection camera transmits images of the body frame that has arrived at the inspection position, and the network camera transmits video footage of the inside of the factory. The elongation measurement sensor senses the elongation of the line chain and transmits the data. The monitor terminal devices 50 are arranged in a grid pattern at 10m intervals in the factory, receive frames transmitted by the communication device 30 and base station 40, and transmit QoS information and the like to the cause estimation / position estimation control server (server 20). It is assumed that the QoS information in the base station 40 can be acquired ideally. The QoS information acquired in the base station 40 and the monitor terminal device 50 is as follows: - RSSI (received signal strength) of successfully received packets Received signal strength RSSI' at the antenna end (obtained only by the monitor terminal device 50) -Channel occupancy rate of each terminal managed within the manufacturing site (hereinafter referred to as "managed terminal") (calculated from the time length of normally received frames) -Channel occupancy rate of unmanaged devices (hereinafter referred to as unmanaged devices) in the manufacturing site Throughput in the transport layer (obtained only by base station 40) The packet of the wireless application is identified from the sending and receiving IP addresses and port numbers in the information elements of the packet acquired at the base station, and the throughput at the base station 40 is calculated from the timestamp when the packet was observed and the packet size.

[0408] The vehicle body frames move along the line at a speed of 0.5 m / sec, and when they arrive at the position of the vehicle body inspection camera, they are inspected for 50 seconds. During this time, the line stops, and after the inspection is completed, they resume movement at a speed of 0.5 m / sec. Assuming that the arrival intervals of the vehicle body frames vary due to factors such as the factory process and manual work, we modeled it using a truncated normal distribution with a lower limit of 0 seconds, an upper limit of 50 seconds, a mean of 40 seconds, and a standard deviation of 2.26 seconds.

[0409] In this simulation, the following events were set as causes of abnormalities: (1) Installation of obstacles: A new metal plate is installed (Figure 44:1) (2) Changing the location of the communication device: The sensor position shifts by several tens of centimeters (Figure 44:2) (3) Increase in traffic on communication devices: Additional traffic temporarily occurs on network cameras (Figure 44:3) (4) Bringing in communication devices from outside: An unmanaged device is brought into the factory (Figure 44:4) (5) Noise from equipment installed at the manufacturing site: Radio noise is emitted from equipment installed in the factory (Figure 44:5) In this simulation, data was obtained for 2500 seconds of normal conditions where no abnormalities occurred, with the first 2000 seconds being learning data 1 and the last 500 seconds being learning data 2. In addition, data for 3300 seconds that was different from the learning data was obtained as observed data, and cause and location estimation was performed on this. Here, the observed data was set so that an abnormality occurred at 2500 seconds.

[0410] [Evaluation Results] Five assumed patterns of communication anomalies were generated, and the causes of communication anomalies and the locations where the causes of the anomalies were estimated using the methods for estimating the causes of the anomalies and the methods for estimating the locations where the causes of the anomalies were occurring described above are shown in the following figures.

[0411] [Evaluation results of methods for estimating the cause of abnormalities] FIG. 45 is a diagram showing the evaluation results of the cause estimation method when each abnormality cause occurs. In FIG. 45, the horizontal axis represents time, and the vertical axis represents the scenario number of the abnormality cause. The solid lines in the diagram represent the state of wireless communication at that time (no abnormality, or abnormality causes 1 to 5). Abnormality cause 1 is "installation of an obstacle," abnormality cause 2 is "change in the installation position of the communication device," abnormality cause 3 is "increase in traffic on the communication device," abnormality cause 4 is "introduction of a communication device from outside," and abnormality cause 5 is "noise from equipment installed at the manufacturing site." Also, the circles plot the state of wireless communication estimated by the cause estimation method at that time.

[0412] Referring to Fig. 45, for anomaly cause 2 and anomaly cause 3, it is clear that the cause of the anomaly can be correctly estimated almost all the time after the anomaly occurs (see Fig. 45(b) and (c)). For anomaly cause 1, the cause can be correctly estimated after the anomaly occurs, but it is determined that some kind of anomaly has occurred even at times when no anomaly has occurred. When there is little deviation between the fluctuation patterns of the input series of learning data 1 and learning data 2, the spread of the probability density distribution of the deviation estimated by kernel density estimation becomes small, and even a fluctuation in the input series pattern that is not actually abnormal is determined to be a rise or descent, and it is considered that an anomaly has been estimated to have occurred.

[0413] For anomaly causes 4 and 5, it is correctly estimated that no anomaly has occurred at times when no anomaly has occurred, but it is erroneously estimated that no anomaly has occurred even after an anomaly has occurred. This is thought to be because the deviation of the input sequence pattern in the observation data is small compared to the rise / descent judgment threshold (Ψ), and so no anomaly was judged. Increasing Ψ makes false alarms more likely to occur, while decreasing Ψ makes it more likely that misdetection of an anomaly will occur. In other words, Ψ is a parameter that controls the trade-off between false alarms and detection and oversight in cause estimation, and it is preferable to set it to an appropriate value.

[0414] [Evaluation results of methods for estimating the location of anomaly causes] Fig. 46 is a diagram showing the evaluation results of the position estimation method when each anomaly cause occurred. In Fig. 46, the x marks indicate the occurrence position of the anomaly cause, and the circles indicate the occurrence position of the anomaly cause estimated by the position estimation method. The plot of each estimated position starts from the time when the anomaly occurred.

[0415] 46, it can be seen that the location of the anomaly cause can be estimated with an accuracy of about 5 m for all anomaly causes. In other words, it is possible to accurately estimate the location of the anomaly cause by considering the QoS / channel status information that is affected by any anomaly cause and switching the location estimation method.

[0416] In the above, the estimation device 10 has been described as estimating the cause of a communication anomaly and the location where the anomaly cause occurs indoors. However, the embodiment of the present invention is not limited to this. The estimation device 10 may also estimate the cause of a communication anomaly and the location where the anomaly cause occurs outdoors by the above-mentioned method.

[0417] According to the above-described embodiment, the estimation device according to the embodiment of the present invention includes: a cause estimation means for estimating a cause of an anomaly that is a cause of an anomaly in wireless communication when a packet transmitted by a communication device cannot be received, based on a deviation degree indicating a degree of deviation of observed data from learned data; a location estimation means for selecting a location estimation method associated with the cause of the abnormality estimated by the cause estimation means, and estimating a location where the cause of the abnormality has occurred according to the selected location estimation method by using QoS / channel status information indicating a communication quality when a packet is transmitted by wireless communication using the channel, the learning data is composed of first QoS / channel status information which is QoS / channel status information acquired in a normal state in which a packet is received and the packet is successfully decoded; The observation data may be made up of second QoS / channel status information, which is QoS / channel status information acquired in a state in which it is unclear whether a packet can be received.

[0418] This is because the estimation device has such a configuration and is able to estimate the cause of the communication anomaly and the location where the cause of the anomaly has occurred, and can take measures to eliminate the cause of the communication anomaly based on the estimation result.

[0419] Further, according to the above-described embodiment, the program according to the embodiment of the present invention is a first step in which a cause estimation means estimates a cause of an anomaly that is a cause of an anomaly in wireless communication when a packet transmitted by a communication device cannot be received, based on a deviation degree indicating a degree of deviation of observed data from learning data; a second step in which the location estimation means selects a location estimation method associated with the cause of the abnormality estimated in the first step, and estimates a location where the cause of the abnormality has occurred in accordance with the selected location estimation method by using QoS / channel status information indicating a communication quality when a packet is transmitted by wireless communication using a channel, the learning data is composed of first QoS / channel status information which is QoS / channel status information acquired in a normal state in which a packet is received and the packet is successfully decoded; The observation data may be made up of second QoS / channel status information, which is QoS / channel status information acquired in a state in which it is unclear whether a packet can be received.

[0420] This is because, by having the program cause the computer to execute the first and second steps, the cause of the communication anomaly and the location where the cause of the anomaly occurs are estimated, and measures can be taken to resolve the cause of the communication anomaly based on the estimation results.

[0421] In an embodiment of the present invention, the QoS / channel condition information constituting the learning data constitutes the "first QoS / channel condition information", the QoS / channel condition information constituting the observation data constitutes the "second QoS / channel condition information", and the QoS / channel condition information constituting the distribution estimation data constitutes the "third QoS / channel condition information".

[0422] Then, estimating the probability density distribution of deviations (see curve k4 in Fig. 10) from the learning data and the distribution estimation data corresponds to "calculating a plurality of deviations indicating the degree of deviation of the third QoS / channel state information from the first QoS / channel state information, and estimating the probability density distribution based on the calculated plurality of deviations." Here, calculating one deviation by the method shown in Fig. 7 for a window of width W that is slid sequentially at timings t, t+1, t+2, ... in the observation data as shown in Fig. 9 constitutes "calculating a plurality of deviations."

[0423] In addition, in the embodiment of the present invention, the threshold Ψ r constitutes the "first threshold" and the threshold Ψ d constitutes the "second threshold value."

[0424] Furthermore, in this embodiment of the present invention, the deviation of the observation data constitutes an "estimation deviation."

[0425] The embodiments disclosed herein should be considered to be illustrative and not restrictive in all respects. The scope of the present invention is defined by the claims, not by the description of the embodiments described above, and is intended to include all modifications within the meaning and scope of the claims. [Industrial Applicability]

[0426] The present invention is applicable to an estimation device, a program, and a computer-readable recording medium having the program recorded thereon. [Explanation of symbols]

[0427] 1 wired cable, 2 antenna, 3, 4 receiving means, 5 estimation means, 10 estimation device, 20 server, 30, 90 communication device, 40 base station, 50 monitor terminal device, 51 cause estimation means, 52 position estimation means, 60 obstacle, 70 tablet terminal, 80 manufacturing device, 100 communication system.

Claims

1. a cause estimation means for estimating a cause of an anomaly that is a cause of an anomaly in wireless communication when a packet transmitted by a communication device cannot be received, based on a deviation degree indicating a degree of deviation of observed data from learned data; a location estimation means for estimating a location where the cause of the abnormality has occurred by using QoS / channel status information indicating a communication quality when a packet is transmitted by the wireless communication using a channel, in accordance with the location estimation method selected in correspondence with the cause of the abnormality estimated by the cause estimation means, the cause estimation means, when determining that a packet has been received from a given communication device in at least one communication device during a certain period of time, determines that the given communication device is not at fault, and when determining that the given communication device is not at fault, estimates the cause of the abnormality; the learning data is made up of first QoS / channel status information, which is the QoS / channel status information acquired in a normal state in which the packet is received and the packet is successfully decoded; The observation data is composed of second QoS / channel condition information, which is the QoS / channel condition information acquired in a state in which it is unclear whether the packet can be received.

2. The cause estimation means calculates a plurality of deviations indicating a degree of deviation from the learning data of distribution estimation data made of third QoS / channel status information, which is the QoS / channel status information acquired at a timing different from that of the first QoS / channel status information in the normal state, by changing an acquisition timing of the third QoS / channel status information, estimates a probability density distribution of the calculated plurality of deviations, calculates an estimated deviation indicating a degree of deviation from the first QoS / channel status information of the second QoS / channel status information in one or more base stations, 2. The estimation device according to claim 1, wherein when it is determined that the received signal strength in the base station status information is an abnormal drop indicating that the deviation for estimation is equal to or less than a first threshold in the probability density distribution, or an abnormal rise indicating that the deviation for estimation is equal to or greater than a second threshold greater than the first threshold in the probability density distribution, it is estimated that the cause of the abnormality is due to a drop or rise in the received signal strength, and when it is determined that the received signal strength in all base stations is neither the abnormal drop nor the abnormal rise, it is estimated that the cause of the abnormality is due to a drop in communication throughput.

3. 3. The estimation device according to claim 2, wherein the cause estimation means further estimates that the cause of the abnormality is installation of an obstacle when it is determined that the received signal strengths for a plurality of communication devices are abnormally decreased or abnormally increased.

4. 4. The estimation device according to claim 3, wherein the cause estimation means, when determining that the received signal strength of a single communication device is abnormally decreased or abnormally increased, estimates that the cause of the abnormality is a change in an installation location of the communication device.

5. 5. The estimation device according to claim 2, wherein the cause estimation means, when determining that the channel occupancy rate of the communication device is abnormally decreasing in one or more base stations or monitor terminal devices, estimates that the cause of the abnormality is an increase in traffic of the communication device, noise from installed equipment, or introduction from outside the communication device.

6. 6. The estimation device according to claim 5, wherein said cause estimation means further estimates that there is no abnormality when it is determined that the channel occupancy rate of said communication device is not abnormally low in all base stations or monitor terminal devices.

7. 7. The estimation device according to claim 5, wherein the cause estimation means further estimates that the cause of the abnormality is an increase in traffic of the communication device when it is determined that the communication throughput has abnormally increased in one or more base stations.

8. 8. The estimation device according to claim 7, wherein the cause estimation means, when determining that the communication throughput is not abnormally increased in all base stations, estimates that the cause of the abnormality is noise brought in from outside the communication device or noise from installed equipment.

9. 9. The estimation device according to claim 8, wherein the cause estimation means further estimates that the cause of the abnormality is brought in from outside the communication device when it is determined in the one or more base stations or monitor terminal devices that the channel occupancy rate of a communication device not under management is the abnormal increase.

10. 10. The estimation device according to claim 9, wherein the cause estimation means further estimates that the cause of the abnormality is noise from installed equipment when it is determined in the monitor terminal device that the channel occupancy rate of the communication device not under management is not the abnormal increase.

11. 9. The estimation device according to claim 5 or claim 8, wherein, when the cause estimation means estimates that the cause of the abnormality is the introduction of a communication device from the outside, the location estimation means estimates a location where the cause of the abnormality has occurred using a received signal strength in the second QoS / channel status information at one or more base stations when the packet is received in the normal state from the communication device brought in from the outside, and, when the cause estimation means estimates that the cause of the abnormality is noise from the installed equipment, the location estimation means estimates a location where the cause of the abnormality has occurred using a first anomaly degree consisting of a minimum value of a Euclidean distance between the received signal strength at the antenna end of a monitor terminal device included in the observation data and the received signal strength included in the learning data.

12. 12. The estimation device according to claim 11, wherein the location estimation means, when the cause estimation means estimates that the cause of the abnormality is an increase in traffic of the communication device, further detects a link with the highest second degree of abnormality, which is formed by a minimum value of a Euclidean distance between the communication throughput included in the observation data and the communication throughput included in the learning data, and estimates a location where the cause of the abnormality has occurred using the locations of three monitor terminal devices with high received signal strengths received from the communication device that is a transmission source of the detected link and the received signal strengths at the three monitor terminal devices.

13. 13. The estimation device according to claim 12, further comprising: when the cause estimation means estimates that the cause of the abnormality is the installation of an obstacle, the position estimation means estimates a position where the cause of the abnormality has occurred, as a midpoint of any number of links detected in descending order of a multiplication result of a first degree of abnormality consisting of a minimum value of a Euclidean distance of the received signal strength at the antenna end of the monitor terminal device included in the observation data to the received signal strength included in the learning data and an average value of the received signal strength included in the learning data, or a midpoint of the midpoints weighted by the multiplication result, or a center of gravity of the midpoints weighted by the multiplication result, when the cause estimation means estimates that the cause of the abnormality is a change in installation position of the communication device, the initial position of the communication device that is a source of the link with the highest first degree of abnormality, as the position where the cause of the abnormality has occurred.

14. a first step in which a cause estimation means estimates a cause of an anomaly that is a cause of an anomaly in wireless communication when a packet transmitted by a communication device cannot be received, based on a deviation degree indicating a degree of deviation of observed data from learning data; a second step in which a location estimation means estimates a location where the cause of the abnormality has occurred by using QoS / channel status information indicating a communication quality when a packet is transmitted by wireless communication using a channel, in accordance with the location estimation method selected in the first step and associated with the cause of the abnormality; the cause estimation means, when determining that a packet has been received from a given communication device in at least one communication device during a certain period of time in the first step, determines that the given communication device is not at fault, and when determining that the given communication device is not at fault, estimates the cause of the abnormality; the learning data is made up of first QoS / channel status information, which is the QoS / channel status information acquired in a normal state in which the packet is received and the packet is successfully decoded; The observation data is composed of second QoS / channel status information, which is the QoS / channel status information acquired in a state in which it is unclear whether the packet can be received.

15. The cause estimation means, in the first step, calculates a plurality of deviations indicating a degree of deviation of distribution estimation data made of third QoS / channel status information, which is the QoS / channel status information acquired at a timing different from that of the first QoS / channel status information in the normal state, from the learning data, by changing an acquisition timing of the third QoS / channel status information, estimates a probability density distribution of the calculated plurality of deviations, calculates an estimated deviation indicating a degree of deviation of the second QoS / channel status information from the first QoS / channel status information, and 15. A program for causing a computer to execute the program of claim 14, wherein when it is determined that the received signal strength in the status information is an abnormal drop indicating that the estimation deviation is equal to or less than a first threshold in the probability density distribution, or an abnormal rise indicating that the estimation deviation is equal to or greater than a second threshold greater than the first threshold in the probability density distribution, it is presumed that the cause of the abnormality is due to a drop or rise in the received signal strength, and when it is determined that the received signal strength in all base stations is neither the abnormal drop nor the abnormal rise, it is presumed that the cause of the abnormality is due to a drop in communication throughput.

16. 16. The program for causing a computer to execute the program according to claim 15, wherein, in the first step, the cause estimation means further estimates that the cause of the abnormality is the installation of an obstacle when it is determined that the received signal strengths for a plurality of communication devices are abnormally decreasing or abnormally increasing.

17. 17. The program for causing a computer to execute the program according to claim 16, wherein, in the first step, when it is determined that there is a single communication device in which the received signal strength is abnormally decreased or abnormally increased, the cause estimation means estimates that the cause of the abnormality is a change in installation location of the communication device.

18. 18. A program for causing a computer to execute the program of claim 15, wherein in the first step, when it is determined in one or more base stations or monitor terminal devices that the channel occupancy rate of the communication device is abnormally decreasing, the cause estimation means estimates that the cause of the abnormality is an increase in traffic of the communication device, or noise from installed equipment, or introduction from outside the communication device.

19. 20. The program for causing a computer to execute the program of claim 18, wherein the cause estimation means, in the first step, further estimates that there is no abnormality when it is determined that the channel occupancy rate of the communication device is not abnormally decreasing in all base stations or monitor terminal devices.

20. 20. The program for causing a computer to execute the program according to claim 18 or 19, wherein in the first step, when it is determined that the communication throughput in one or more base stations has abnormally increased, the cause estimation means estimates that the cause of the abnormality is an increase in traffic of the communication device.

21. 21. The program to be executed by a computer according to claim 20, wherein in the first step, when it is determined that the communication throughput is not abnormally increased in all base stations, the cause estimation means estimates that the cause of the abnormality is noise brought in from outside the communication device or noise from installed equipment.

22. 22. A program for causing a computer to execute the program described in claim 21, wherein in the first step, the cause estimation means further estimates that the cause of the abnormality is brought in from outside the communication device when it is determined in the one or more base stations or monitor terminal devices that the channel occupancy rate of a communication device not under management is the abnormal increase.

23. 23. The program for causing a computer to execute the program according to claim 22, wherein in the first step, the cause estimation means further estimates that the cause of the abnormality is noise from installed equipment when it is determined that the channel occupancy rate of the communication device not under management is not the abnormal increase in all of the base stations or monitor terminal devices.

24. 22. The program for causing a computer to execute the program according to claim 18 or claim 21, wherein, in the second step, when the cause estimation means estimates that the cause of the abnormality is the bringing-in of a communication device from the outside, the location estimation means estimates a location where the cause of the abnormality has occurred using a received signal strength in the second QoS / channel status information at one or more base stations when the packet is received in the normal state from the communication device brought in from the outside, and when the cause estimation means estimates that the cause of the abnormality is noise from the installed equipment, the location estimation means estimates a location where the cause of the abnormality has occurred using a first anomaly degree consisting of a minimum value of Euclidean distance between the received signal strength at the antenna end of a monitor terminal device included in the observation data and the received signal strength included in the learning data.

25. 25. The program for causing a computer to execute the program according to claim 24, wherein in the second step, when the cause estimation means estimates that the cause of the abnormality is an increase in traffic of the communication device, the location estimation means further detects a link with the highest second degree of abnormality, which is formed by a minimum value of the Euclidean distance of the communication throughput included in the observation data to the communication throughput included in the learning data, and estimates a location where the cause of the abnormality has occurred using the locations of three monitor terminal devices with high received signal strengths received from the communication device that is the transmission source of the detected link and the received signal strengths at the three monitor terminal devices.

26. 26. The program for causing a computer to execute the program according to claim 25, wherein in the second step, when the cause estimation means estimates that the cause of the abnormality is the installation of an obstacle, the position estimation means further estimates as the position where the cause of the abnormality has occurred a midpoint of any number of links detected in descending order of a multiplication result of a first degree of abnormality consisting of a minimum value of a Euclidean distance of the received signal strength at the antenna end of the monitor terminal device included in the observation data to the received signal strength included in the learning data and an average value of the received signal strength included in the learning data, or a midpoint of the midpoints weighted by the multiplication result, or a center of gravity of the midpoints weighted by the multiplication result, when the cause estimation means estimates that the cause of the abnormality is a change in installation position of the communication device, the initial position of the communication device that is the source of the link with the highest first degree of abnormality as the position where the cause of the abnormality has occurred.

27. A computer-readable recording medium having the program according to any one of claims 14 to 26 recorded thereon.

Images