Wireless communication system, wireless communication method, and control station

Abstract

To provide a wireless communication system having high reliability.SOLUTION: An abnormality detection system of wireless communication in a control system is constituted of a control station and multiple moving vehicles connected to the control station via the wireless communication. Each of the multiple moving vehicles regularly measures wireless communication quality information that represents quality of location information and wireless communication and reports the measurement result to the control station. The control station groups the measurement results belonging to a first-time range on the basis of the location information out of the measurement results reported from multiple moving vehicles and learns a distribution of the wireless communication quality information by grouped locations. The control station further acquires measurement results belonging to a second time range that is different from the first time range out of the measurement results reported from multiple moving vehicles and derives an existence ratio of outliers in the measurement result belonging to the second time range by using the distribution of wireless communication quality information for each learned location range.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a wireless communication system, a wireless communication method used for controlling a moving object, and a control station constituting the wireless communication system.

Background Art

[0002] In recent years, wireless communication has been used for various purposes in various environments. As one form of use of wireless communication, there is a control system in which a central control station collects information from a plurality of moving objects such as vehicles and controls the movement of each moving object.

[0003] In the control system, if a communication interruption occurs in the wireless communication with a vehicle under the control of the central control station, not only can the central control station not sufficiently control the vehicle, but the central control station also cannot obtain the information necessary for controlling the vehicle. In addition, a communication interruption in one vehicle can also affect the control of other vehicles. Therefore, when applying wireless communication to a control system or the like, it is extremely important to ensure the reliability of wireless communication without causing a communication interruption.

[0004] On the other hand, the propagation of radio waves used for wireless communication is greatly affected by structures and the like existing in the propagation area, and due to the movement or deformation of structures and the like, the propagation of radio waves, and thus the quality of wireless communication, fluctuates, causing a communication interruption. The presence or frequency of a communication interruption varies depending on the location and also fluctuates with time. Therefore, in order to ensure the reliability of wireless communication, a technique for measuring the communication quality during communication and taking measures based on the measurement results has been proposed.

[0005] For example, Patent Document 1 discloses a wireless communication technology that enables smooth autonomous operation of a dump truck in a mine by setting a permitted driving section when the communication quality meets a predetermined quality standard, thereby ensuring the reliability of wireless communication even in mines with poor radio wave conditions. In other words, the technology disclosed in Patent Document 1 determines the connection status based on the communication quality measured in the autonomous dump truck in the mine, and allows driving in a predetermined permitted driving section when the communication quality meets a predetermined quality standard, thereby enabling smooth autonomous operation of the dump truck.

[0006] However, the technology disclosed in Patent Document 1 detects fluctuations in radio wave propagation due to movement or deformation of structures, as well as fluctuations in the quality of wireless communication due to equipment malfunctions, after the fact. Therefore, there is a problem that the target control may not be able to be executed depending on the cause of the fluctuation in communication quality. [Prior art documents] [Patent Documents]

[0007] [Patent Document 1] Japanese Patent Publication No. 2015-210595 [Overview of the project] [Problems that the invention aims to solve]

[0008] The present invention aims to solve the aforementioned problems by providing a highly reliable wireless communication system, wireless communication method, and control station constituting the wireless communication system, which can detect fluctuations in communication quality before the communication quality falls below a predetermined level. [Means for solving the problem]

[0009] The wireless communication system according to the present invention is a wireless communication system in a control system including a control station and a plurality of mobile vehicles connected to the control station via wireless communication, wherein the plurality of mobile vehicles periodically measure location information and wireless communication quality information and report the measurement results to the control station. The control station groups the measurement results from the plurality of mobile vehicles that belong to a first time range based on location information and learns the distribution of wireless communication quality information indicating the quality of the wireless communication for each grouped location range. Furthermore, the control station acquires measurement results from the plurality of mobile vehicles that belong to a second time range different from the first time range, and uses the learned distribution of wireless communication quality information for each location range to derive the ratio of outliers in the measurement results belonging to the second time range. The system has an anomaly detection unit that determines anomaly detection based on the ratio of outliers. [Effects of the Invention]

[0010] According to the present invention, it is possible to provide a wireless communication system, a wireless communication method, and a control station constituting the wireless communication system that have high reliability for wireless communication used for controlling mobile objects. [Brief explanation of the drawing]

[0011] [Figure 1] This is a schematic diagram of the wireless communication system configuration according to an embodiment of the present invention. [Figure 2] This is a schematic diagram of the abnormality detection unit according to an embodiment of the present invention. [Figure 3] This is a schematic diagram of the hardware configuration of the anomaly detection unit according to an embodiment of the present invention. [Figure 4A] This is an example of an information structure held in the radio station information holding unit of an embodiment of the present invention. [Figure 4B] This is an example of an information structure held in the radio station information holding unit of an embodiment of the present invention. [Figure 4C] This is an example of an information structure held in the radio station information holding unit of an embodiment of the present invention. [Figure 4D]An example of the information structure held in the wireless station information holding unit according to an embodiment of the present invention. [Figure 4E] An example of the information structure held in the wireless station information holding unit according to an embodiment of the present invention. [Figure 5] An example of the processing flow of the learning process in an embodiment of the present invention. [Figure 6A] An example of the information structure of the representative wireless communication quality according to an embodiment of the present invention. [Figure 6B] An example of the information structure of the representative wireless communication quality according to an embodiment of the present invention. [Figure 6C] An example of the information structure of the representative wireless communication quality according to an embodiment of the present invention. [Figure 7] An example of the position range according to an embodiment of the present invention. [Figure 8] An example of the processing flow of the outlier learning process in an embodiment of the present invention. [Figure 9A] The flow of processing execution by the periodic drive of the abnormality detection unit 131 according to an embodiment of the present invention. [Figure 9B] The flow of processing execution by the periodic drive of the abnormality detection unit 131 according to an embodiment of the present invention. [Figure 10] The flow of processing execution by the request drive of the abnormality detection unit 131 according to an embodiment of the present invention. [Figure 11] An example of the analysis processing performed in the abnormality cause analysis processing according to an embodiment of the present invention. [Figure 12] The flow of another example of the analysis processing performed in the abnormality cause analysis processing according to an embodiment of the present invention. [Figure 13] One of the explanatory diagrams of another example of the analysis processing performed in the abnormality cause analysis processing according to an embodiment of the present invention. [Figure 14] One of the explanatory diagrams of another example of the analysis processing performed in the abnormality cause analysis processing according to an embodiment of the present invention. [Figure 15] Yet another example of the analysis processing performed in the abnormality cause analysis processing according to an embodiment of the present invention.

Embodiments for Carrying Out the Invention

[0012] Embodiments of the present invention will be described below with reference to the drawings. In the following embodiments, multiple components may be given as examples in specific numbers, but unless otherwise specified, the number of such components may differ from those given as examples.

[0013] Furthermore, while the following embodiments may describe the processes in a specific order, the order of the processes may be changed, or they may be performed in parallel, except in cases where there is a dependency on the order, such as when the result of one process is used in another process. Moreover, even if a later process uses the result of a earlier process, each process may be executed asynchronously, and the later process may use the latest execution result of the earlier process at the time the later process is executed.

[0014] Figure 1 is a schematic diagram of the wireless communication system configuration according to an embodiment of the present invention. The wireless communication system in this embodiment comprises a mobile vehicle 100, a fixed radio station 110, a fixed network 120, a control station 130, and a mobile network 140.

[0015] The mobile vehicle 100 includes a mobile radio station 101 (101-a, 101-b) and a position sensor 102. The control station 130 includes an anomaly detection unit 131. The anomaly detection unit 131 may include a display (display unit 131D) that displays the results of anomaly detection. The mobile radio station 101 and the fixed radio station 110 are connected wirelessly by a mobile network 140, and the fixed radio station 110 and the control station 130 are connected by a fixed network 120 via wired or wireless communication. In Figure 1, three fixed radio stations 110 (110-A, 110-B, 110-C) and two mobile vehicles 100 are shown, but it goes without saying that the number is not limited to these.

[0016] Either or both of the mobile radio station 101 and the fixed radio station 110 measure the radio communication quality of the mobile network 140 and periodically notify the anomaly detection unit 131 in the control station 130 of the radio communication quality information as a result of the measurement via either or both of the fixed network 120 and the mobile network 140. The radio communication quality information is information regarding the quality of radio communication conducted between the mobile radio station 101 and the fixed radio station 110, and includes, for example, the received power, SNR (Signal-to-Noise Ratio), and information about the time the radio communication quality was measured. Furthermore, the radio communication quality information obtained by the mobile radio station 101 may include the ID of the mobile radio station 101 that performed the measurement.

[0017] In the following, when a distinction is necessary, the radio communication quality, received power, and SNR measured at the fixed radio station 110 will be referred to as "uplink radio communication quality," "uplink received power," and "uplink SNR," respectively. Furthermore, the radio communication quality, received power, and SNR measured at the mobile radio station 101 will be referred to as "downlink radio communication quality," "downlink received power," and "downlink SNR," respectively.

[0018] The fixed radio station 110 measures the radio communication quality, which is the quality of the radio communication between itself and each of the mobile radio stations 101 with which it is communicating. The radio communication quality information is not limited to specific information, but may include, for example, the ID of the mobile radio station 101 with which it is communicating. The radio communication quality information may also include location information of the mobile vehicle 100 measured using a position sensor 102, for example, a GNSS (Global Navigation Satellite System).

[0019] The mobile radio station 101 or the fixed radio station 110 adds time information to the measured location information and notifies the anomaly detection unit 131 in the control station 130 via the mobile network 140 and / or fixed network 120. The anomaly detection unit 131 performs anomaly detection processing of the wireless communication system based on the notified wireless communication quality information and location information.

[0020] In the above example, the position information of the mobile vehicle 100 is measured using the position sensor 102. However, instead of this, or in addition to this, the position information of the mobile radio station 101 and the mobile vehicle 100 may be measured at the fixed radio station 110 using, for example, the difference in radio wave arrival time or the direction of radio wave arrival, and based on this, the positions of the mobile radio station 101 and the mobile vehicle 100 may be measured. In this case, time information is added to the position information measured by the fixed radio station 110, and this added information is notified to the anomaly detection unit 131 in the control station 130 via the fixed network 120.

[0021] The configuration of the anomaly detection unit 131 will be described in more detail with reference to the block diagram in Figure 2. This anomaly detection unit 131 includes, as an example, an information interface processing unit 201, a learning processing unit 202, an outlier detection processing unit 203, an anomaly cause analysis processing unit 204, and a detection result interface processing unit 205. Furthermore, the anomaly detection unit 131 includes a radio station information storage unit 211, a learning result storage unit 212, a detection result storage unit 213, and an analysis result storage unit 214 as information storage units.

[0022] Figure 3 is a schematic diagram of the hardware configuration for realizing the anomaly detection unit 131. The anomaly detection unit 131 in this embodiment includes, for example, a CPU / DSP 301, a memory unit 302, a logic circuit 303, a LAN I / F 304, and a bus 305 connecting these components.

[0023] The various calculations, judgments, and input / output operations in the information interface processing unit 201, learning processing unit 202, outlier detection processing unit 203, anomaly cause analysis processing unit 204, and detection result interface processing unit 205 are executed using the resources of the CPU / DSP 301, storage unit 302, and logic circuit 303, for example, according to a computer program stored in the storage unit 302. It is desirable that each process in the anomaly detection unit 131 operates independently as a different process or thread, either individually or in groups of processes.

[0024] The radio station information storage unit 211, the learning result storage unit 212, the detection result storage unit 213, and the analysis result storage unit 214 each store their respective information in the storage unit 302. The storage unit 302 may include, for example, volatile memory such as RAM (Random Access Memory), non-volatile memory such as SSD (Solid State Drive), or magnetic storage device such as HDD (Hard Disk Drive).

[0025] The information interface processing unit 201 and the detection result interface processing unit 205, which are interface processing units of the anomaly detection unit 131, exchange information with the outside of the anomaly detection unit 131 via the LAN I / F 304. The information interface processing unit 201 receives the location information of the mobile vehicle 100 obtained by measurement by the position sensor 102 via the fixed network 120, and the wireless communication quality information of the mobile network 140 obtained by measurement by either or both of the mobile radio station 101 and the fixed radio station 110. If the wireless communication quality information does not include time information, the information interface processing unit 201 adds the time information to the received information and then has the radio station information holding unit 211 hold the data.

[0026] The radio station information storage unit 211 stores radio communication quality information and location information in association with each other based on time information. Here, for example, if the frequency of acquiring location information is lower than the frequency of acquiring radio communication quality information, interpolated location information can be associated with the radio communication quality information.

[0027] Figures 4A to 4E show an example of the data structure of the information held in the radio station information holding unit 211. Figures 4A to 4E illustrate the case where two values, received power and SNR, are used as radio communication quality indicators, but other received quality indicators may also be used.

[0028] Figure 4A shows an example of the data structure for information when measuring the wireless communication quality at the mobile radio station 101, using a wireless communication method in the mobile network 140 in which signals on the same channel are transmitted from multiple different fixed radio stations 110, such as an SFN (Single Frequency Network), without distinction required at the mobile radio station 101. In this case, as shown in Figure 4A, the radio station information storage unit records combinations of time, position coordinates, mobile radio station ID, downlink received power, and downlink SNR, using the time and mobile radio station ID as keys.

[0029] Figure 4B shows an example in which a wireless communication method is employed in the mobile network 140, such as a wireless LAN (Local Area Network) or cellular communication, where signals on different channels that need to be distinguished by the mobile radio station 101 are transmitted from multiple different fixed radio stations 110, and the wireless communication quality is measured on the mobile radio station 101 side. In this case, as shown in Figure 4B, the radio station information holding unit 211 records the time, position coordinates, wireless mobile station ID, and the combination of downlink received power and downlink SNR for each fixed radio station 110 ((A) to (C)), using the time and wireless mobile station ID as keys. If the mobile radio station 101 is unable to receive a signal from a particular fixed radio station 110, the wireless communication quality for that fixed radio station 110 is treated as a missing value.

[0030] Figure 4C shows an example of measuring the wireless communication quality at the fixed radio station 110. In this case, as shown in Figure 4C, the radio station information storage unit 211 records the time, position coordinates, radio mobile station ID, and the combination of uplink received power and uplink SNR for each fixed radio station 110 that was measured, using the time and radio mobile station ID as keys. If it is not possible to receive a signal from the mobile radio station 101 at a particular fixed radio station 110, the wireless communication quality for that fixed radio station 110 is treated as a missing value.

[0031] Figure 4D shows an example of the data structure for information when a wireless communication method is adopted in the mobile network 140, such as an SFN (Single Frequency Network), in which signals on the same channel that do not need to be distinguished on the mobile radio station 101 side are transmitted from multiple different fixed radio stations 110, and wireless communication quality is measured at both the mobile radio station 101 and the fixed radio stations 110. In this case, as shown in Figure 4D, the radio station information holding unit 211 records a combination of time, position coordinates, radio mobile station ID, downlink received power, downlink SNR, uplink received power for each fixed radio station 110, and uplink SNR, using time and radio mobile station ID as keys. The handling of missing values ​​when signal reception is impossible is the same as in the case of Figure 4C.

[0032] Figure 4E shows an example of the data structure for information when a wireless communication method is adopted in the mobile network 140, such as a wireless LAN (Local Area Network) or cellular communication, in which signals on different channels that need to be distinguished on the mobile radio station 101 side are transmitted from multiple different fixed radio stations 110, and wireless communication quality is measured at both the mobile radio station 101 and the fixed radio stations 110. In this case, as shown in Figure 4E, the radio station information holding unit 211 records the time, position coordinates, radio mobile station ID, and the combination of downlink received power and downlink SNR for each fixed radio station 110, and uplink received power and uplink SNR for each fixed radio station 110, using the time and radio mobile station ID as keys. The handling of missing values ​​when signal reception is impossible is the same as in the cases of Figures 4B and 4C.

[0033] The learning processing unit 202 retrieves information held in the radio station information holding unit 211 at fixed time and location intervals, learns the distribution of representative radio communication quality for each, and stores the learned results in the learning result holding unit 212.

[0034] Referring to the flowchart in Figure 5, an example of the learning process procedure of the learning processing unit 202 in this embodiment will be explained. In the learning processing unit 202, first, in step S510, the time range to be learned is set. Here, the time range to be learned refers to, for example, a 24-hour period on a specific day.

[0035] Next, loop processing (grouping) is performed for each location range from steps S511 to S517. In step S512, data corresponding to the time range set in step S510 and the location range of the current loop processing is extracted from the radio station information holding unit 211. In the following step S513, the amount of data extracted is checked, and if the amount of data is less than a specified value, for example less than 100 points (No), the loop processing corresponding to the currently processed location range is terminated and the process returns to step S511 to perform processing for the next location range. If the amount of data is equal to or greater than the specified value, the process proceeds to step S514.

[0036] Next, in step S514, a representative radio communication quality value is created. Here, the representative radio communication quality value is the value of one selected downlink radio communication quality when there are multiple downlink radio communication quality values ​​for a single key consisting of time and radio mobile station ID, as shown in Figures 4B and 4E. Also, it is the value of one selected uplink radio communication quality when there are multiple uplink radio communication quality values ​​for a single key consisting of time and radio mobile station ID, as shown in Figures 4C, 4D, and 4E. As for the method of selecting the representative radio communication quality, for example, it is possible to select the radio communication quality value with the largest value that does not have any missing values ​​among the multiple radio communication quality values, such as received power or SNR.

[0037] If the wireless communication quality information stored in the radio station information storage unit 211 is in the format shown in Figure 4A or Figure 4B, the representative wireless communication quality value will be an array consisting of two elements: downlink representative received power and downlink representative SNR, as shown in Figure 6A. If the wireless communication quality information stored in the radio station information storage unit 211 is in the format shown in Figure 4C, the representative wireless communication quality value will be an array consisting of two elements: uplink representative received power and uplink representative SNR, as shown in Figure 6B. If the wireless communication quality information stored in the radio station information storage unit 211 is in the format shown in Figure 4D or Figure 4E, the representative wireless communication quality value will be an array consisting of four elements: downlink representative received power, downlink representative SNR, uplink representative received power, and uplink representative SNR, as shown in Figure 6C.

[0038] In the following step S515, the characteristics of the distribution of representative wireless communication quality values ​​are learned. One method for learning the distribution is to assume a Gaussian distribution with dimensions equal to the number of elements of the representative wireless communication quality values, and calculate the mean and standard deviation for each dimension of the Gaussian distribution such that, for example, 95% of the target data falls within 3σ. Another method for learning the distribution is to construct a One Class SVM classifier such that, for example, 95% of the data falls within the normal range.

[0039] Next, in step S516, the learning results are stored in the learning result storage unit 212 as a classifier. The above process is repeated, and in step S517, if there are still position ranges that have not been processed, the process returns to step S511, and if processing for all position ranges has been completed, the learning step S202 is terminated.

[0040] Figure 7 shows an example of a position range in this embodiment. In this embodiment, the overall area range 710 is defined to include the travel trajectory 700 of the moving vehicle 100, and the overall area range 710 is divided into multiple sub-areas 720 at equal intervals, for example. Each sub-area 720 serves as a processing unit for the learning process in the learning processing unit 202.

[0041] The outlier detection processing unit 203 retrieves wireless communication quality information and location information stored in the wireless station information storage unit 211 at fixed time and location intervals, classifies the retrieved information using a classifier stored in the learning result storage unit 212, and stores the classification results in the detection result storage unit 213.

[0042] Referring to the flowchart in Figure 8, an example of the processing procedure in the outlier detection processing unit 203 in this embodiment will be explained. In the outlier detection processing unit 203, first in step S810, the time range to be targeted for outlier detection processing and the time range of the classifier used for outlier detection processing are set. Here, the time range to be targeted for outlier detection processing refers to, for example, a specific time range within a specific day (e.g., 1 hour). The time range of the classifier used for outlier detection processing refers to, for example, the time range of the day before the time range to be targeted for outlier detection processing (e.g., 24 hours of the previous day).

[0043] Next, loop processing is performed for each position range from step S811 to step S818. In step S812, the classifiers for the target position range and classifier time, which will be used for outlier detection processing, are extracted from the learning result storage unit 212.

[0044] In the following step S813, it is checked whether the classifier in question has been extracted (whether it exists or not). In step S812, if the corresponding classifier does not exist in the learning result storage unit 212 and extraction has failed (No), the loop processing for the currently processed position range is terminated, and the process returns to step S811 to perform processing for the next position range. If the classifier in question has been extracted, the process proceeds to step S814.

[0045] In step S814, information corresponding to the time range to be targeted for outlier detection processing set in step S810, and the position range of the current loop processing, is extracted from the radio station information holding unit 211. Then, in step S815, the amount of data of the extracted information is checked. If the amount of data is less than a specified value, for example less than 50 points, the loop processing corresponding to the currently processed position range is terminated and the process returns to step S811 to perform processing for the next position range. If the amount of data is equal to or greater than the specified value, the process proceeds to step S816.

[0046] In step S816, the data extracted in step S814 is classified using the classifier extracted in step S812. In the following step S817, the classification results and statistical information are recorded in the detection result holding unit 213. The classification results and statistical information include, as an example, the following: • Number of data points that were classified • Number of data points within the normal range of the classification result • Number of data points that were identified as outliers in the classification process. • Average wireless communication quality value of all data • Average wireless communication quality value of data within the normal range • Average wireless communication quality value of outlier data

[0047] The above steps are repeated, and if there are any position ranges in step S818 for which the above steps have not been performed, the process returns to step S811 and the same process is repeated. On the other hand, if processing for all position ranges has been completed, the outlier detection process is terminated. The anomaly cause analysis processing unit 204 analyzes the anomaly when an anomaly occurs in wireless communication based on the information held in the detection result holding unit 213, and records the results in the analysis result holding unit 214.

[0048] The detection result interface processing unit 205 outputs the analysis results recorded in the analysis result holding unit 214 to, for example, an external monitoring terminal via the fixed network 120. The detection result interface processing unit 205 also performs various processes in the learning processing unit 202, outlier detection processing unit 203, and anomaly factor analysis processing unit 204 as needed, based on analysis request instructions requested from, for example, an external monitoring terminal via the fixed network 120, and then outputs the analysis results recorded in the analysis result holding unit 214 to, for example, an external monitoring terminal via the fixed network 120.

[0049] Referring to the flowcharts in Figures 9A and 9B, the procedure for processing by the timed drive of the anomaly detection unit 131 will be explained. As for the timed drive processing, for example, as shown in step S910 in Figure 9A, processing is started once a day when the radio station information holding unit 211 has stored the radio communication quality for the previous day. In step S911, the learning processing unit 202 is started with the learning target time range set to the previous day, and the result is recorded in the learning result holding unit 212.

[0050] As a time-driven process, as shown in step S920 of Figure 9B, processing can also be started once every hour when the radio station information holding unit 211 has stored one hour's worth of wireless communication quality information. Then, in step S921, using the previous day's learning results created in step S911, and with the wireless communication quality information for the previous hour as the classification target, the outlier detection processing unit 203 performs outlier detection processing, and the results are recorded in the detection result holding unit 213. Subsequently, in step S922, the anomaly cause analysis processing unit 204 performs cause analysis processing, and the results are recorded in the analysis result holding unit 214. Then, in step S923, the analysis results are output externally through the detection result interface processing unit 205.

[0051] Referring to the flowchart in Figure 10, the procedure for detection processing by the anomaly detection unit 131 when an analysis request (request-driven) is received will be explained. For example, when the detection result interface processing unit 205 receives an analysis request instruction, the process shown in Figure 10 is started.

[0052] When an analysis request initiates the process shown in Figure 10, the first step, S1010, checks whether the analysis result corresponding to the analysis request instruction exists in the analysis result holding unit 214. If the analysis result exists, the process proceeds to step S1016; otherwise, it proceeds to step S1011.

[0053] In step S1011, it is checked whether a detection result corresponding to the analysis request instruction exists in the detection result storage unit 213. If a detection result exists, the process proceeds to step S1015; otherwise, it proceeds to step S1012. In step S1012, it is checked whether a learning result corresponding to the analysis request instruction exists in the learning result storage unit 212. If a learning result exists, the process proceeds to step S1014; otherwise, it proceeds to step S1013.

[0054] In step S1013, the learning processing unit 202 executes a learning process within the learning target time range corresponding to the analysis request instruction received by the detection result interface processing unit 205, and records the results in the learning result storage unit 212. In step S1014, the outlier detection processing unit 203 executes an outlier detection process within the classifier and classification target time range corresponding to the analysis request instruction received by the detection result interface processing unit 205, and records the results in the detection result storage unit 213. In the following step S1015, the anomaly cause analysis processing corresponding to the analysis request instruction received by the detection result interface processing unit 205 is executed in the anomaly cause analysis processing unit 204, and the results are recorded in the analysis result storage unit 214. In step S1016, the analysis results corresponding to the analysis request instruction received by the detection result interface processing unit 205 are output externally from the information recorded in the analysis result storage unit 214.

[0055] Referring to Figure 11, an example of the analysis process in the anomaly cause analysis processing unit 204 will be explained. In the analysis process in the example in Figure 11, the travel trajectory 700 of the moving vehicle 100 and the entire area range 710 are displayed on the display 131D of the anomaly detection unit 131, and the time range information 730 displays the learning target time range of the classifier used for the classification process and the target time range of the data that was classified.

[0056] Furthermore, for each location range, the outlier ratio is calculated by dividing the number of data points classified as outliers by the outlier detection processing unit 203 by the total number of data points. Location ranges where the outlier ratio exceeds a threshold (e.g., 0.5 or higher) are highlighted, for example, by code 740. This area of ​​code 740 is recognized as an area where an anomaly has occurred. If the number of areas where anomalies have occurred exceeds a certain level, it can be recognized that an anomaly has occurred in the entire wireless communication system.

[0057] Referring to the flowchart in Figure 12, an example of another procedure for the analysis process performed by the anomaly cause analysis processing unit 204 will be explained. In this analysis process in Figure 12, first, in step S1210, the learning time range of the classifier and the time range of the wireless communication quality information to be classified are set. Next, in step S1211, the statistical values ​​corresponding to the time range set in step S1210 are read from the detection result holding unit 213.

[0058] In the following step S1212, similar to the analysis process in Figure 11, the ratio of outliers is calculated for each location range, and location ranges with an outlier ratio of 0.5 or more are classified as outlier location ranges, and location ranges with an outlier ratio of less than 0.5 are classified as normal location ranges. Then, in step S1213, for each location range, the distance from all fixed radio stations (ground radio stations) 110 is calculated, as illustrated in Figure 13. Here, the location information of the ground radio stations 110 used in the calculation may be location information measured using, for example, GNSS, or it may be a value that has been stored as a constant in advance.

[0059] Next, in step S1214, the classification results of the normal position range and outlier position range obtained in step S1212 based on the ratio of outliers are used as training data, and the distance from each overhead radio station is used as a parameter to train a binary classifier. Furthermore, using the trained binary classifier, the entire area range 710 is classified into a normal position range and an outlier position range by the binary classifier, as illustrated in Figure 14. The binary classifier used here can be a simple classifier such as a decision tree, or a classifier such as Adaboost which is a combination of decision trees.

[0060] Next, in step S1215, the installation location of each ground radio station 110 is classified as either an outlier location range or a normal location range. If, as a result, a ground radio station 110 is classified as being in an outlier location range, it is possible to recognize that some kind of communication problem (such as a communication interruption) is occurring at that ground radio station.

[0061] In the subsequent step S1216, the feature importance during binary classifier training is evaluated. For ground radio stations 110 classified as being within the normal position range in step S1215, their importance is considered to be 0. For ground radio stations 110 classified as being within the outlier position range in step S1215, it becomes possible to recognize that the higher the feature importance, the higher the likelihood of an anomaly occurring.

[0062] Figure 15 shows another example of the analysis process performed by the anomaly cause analysis processing unit 204. The horizontal axis of the graph shows the change over time (time), and the vertical axis shows the feature importance obtained every hour according to the flow shown in Figure 12. By checking the time variation of these feature importance values, it is possible to estimate the time range in which anomalies occurred for each ground radio station 110.

[0063] It should be noted that the present invention is not limited to the embodiments described above, but includes various modifications and equivalent configurations within the spirit of the attached claims. For example, the embodiments described above are described in detail for the purpose of clearly illustrating the present invention, and the present invention is not necessarily limited to having all of the described configurations. Furthermore, a part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Furthermore, a configuration of another embodiment may be added to the configuration of one embodiment. Furthermore, a part of the configuration of each embodiment may be added, deleted, or replaced with other configurations.

[0064] Furthermore, some or all of these may be implemented in hardware, such as integrated circuits. Also, each of the aforementioned configurations, functions, processing units, and processing means may be implemented in software by a processor interpreting and executing computer programs that realize their respective functions. If multiple identical functions exist in the configuration, the hardware or software that realizes each function may be implemented separately, or a single implemented hardware or software may be used in time-multiplexing to perform multiple processes. Even if a function is a single function in the configuration, distributed processing may be performed using multiple pieces of hardware or software with the same function.

[0065] Information such as programs, tables, and files that implement each function can be stored in memory, hard disks, SSDs (Solid State Drives), or other storage media such as IC cards, SD cards, and DVDs.

[0066] Furthermore, the control lines and information lines shown are those deemed necessary for explanation purposes and do not necessarily represent all control lines and information lines required for implementation. In reality, it can be assumed that almost all components are interconnected. [Explanation of symbols]

[0067] 100...Mobile vehicle, 101...Mobile radio station, 102...Position sensor, 110...Fixed radio station, 120...Fixed network, 130...Control station, 131...Anomaly detection unit, 140...Mobile network, 201...Information interface processing unit, 202...Learning processing unit, 203...Outlier detection processing unit, 204...Anomaly cause analysis processing unit, 205...Detection result interface processing unit, 211...Radio station information storage unit, 212...Learning result storage unit, 213...Detection result storage unit, 214...Analysis result storage unit, 301...CPU / DSP, 302...Memory unit, 303...Logic circuit, 304...LAN interface, 305...Bus.

Claims

1. A wireless communication system in a control system including a control station and a plurality of mobile vehicles connected to the control station via wireless communication, the plurality of moving vehicles or fixed wireless stations connected to the plurality of moving vehicles by wireless communication periodically measure location information and wireless communication quality information indicating quality of the wireless communication, and report the measurement results to the control station; the control station defines travel trajectories of the plurality of moving vehicles and an area including the travel trajectories based on the position information reported from the plurality of moving vehicles, and divides the area into a plurality of small divided areas; among the measurement results reported from the plurality of moving vehicles, a first time range is set as a learning object, and measurement results belonging to the first time range are grouped based on the position information so as to correspond to the plurality of small divided areas; and learns a distribution of the wireless communication quality information for each of the grouped small divided areas; acquiring measurement results belonging to a second time range different from the first time range among the measurement results reported from the plurality of moving vehicles by subjecting the measurement results to a detection process, and deriving an existence ratio of outliers in the measurement results belonging to the second time range for each small divided area using the learned distribution of the wireless communication quality information for each small divided area; the control station has an anomaly detection device that detects an occurrence of an anomaly in wireless communication in each of the plurality of small divided areas in the second time range based on the existence ratio of the outliers; The abnormality detection device is a wireless communication system having a display unit that displays the area, each of the multiple small divided areas, and the traveling trajectory of the moving vehicle, and highlights a small divided area among the multiple small divided areas in which an abnormality is occurring in wireless communication.

2. 2. The wireless communication system according to claim 1, wherein the control station determines that an abnormality has occurred in the wireless communication system when there is a small divided area in which the proportion of the outliers is equal to or greater than a threshold value.

3. 2. The wireless communication system according to claim 1, The wireless communication system further comprises an anomaly detection device that classifies a sub-divided area as an outlier sub-divided area when the proportion of outliers present in each sub-divided area is equal to or greater than a threshold, and classifies the sub-divided area as a normal range when the proportion of outliers present in each sub-divided area is less than the threshold, and executes a learning process for a binary classifier using training data, and when the position of the moving vehicle is classified using the binary classifier, determines that an anomaly has been detected for the moving vehicle when the position of the moving vehicle is included in the outlier sub-divided area.

4. 1. A wireless communication method between a control station and a plurality of moving vehicles connected to the control station via wireless communication, comprising: the plurality of moving vehicles or fixed wireless stations connected to the plurality of moving vehicles by wireless communication periodically measure location information and wireless communication quality information indicating quality of the wireless communication and report the measurement results to the control station; defining travel trajectories of the plurality of moving vehicles and an area including the travel trajectories based on the position information reported from the plurality of moving vehicles, and dividing the area into a plurality of small divided areas; among the measurement results reported from the plurality of moving vehicles, a first time range is set as a learning object, and measurement results belonging to the first time range are grouped based on the position information so as to correspond to the plurality of small divided areas; and learning a distribution of the wireless communication quality information for each of the grouped small divided areas; acquiring measurement results belonging to a second time range different from the first time range among the measurement results reported from the plurality of moving vehicles by subjecting the measurement results to a detection process, and deriving an existence ratio of outliers in the measurement results belonging to the second time range for each small divided area using the learned distribution of the wireless communication quality information for each small divided area; detecting an occurrence of an abnormality in wireless communication in each of the plurality of small divided areas in the second time range based on the existence ratio of the outliers; The area, each of the plurality of small divided areas, and the travel path of the moving vehicle are displayed, and a small divided area in which an abnormality occurs in wireless communication is highlighted among the plurality of small divided areas.

23. A wireless communication method comprising:

5. 5. The wireless communication method according to claim 4, further comprising: determining that an abnormality has occurred in the wireless communication when there is a small divided area in which the proportion of the outliers is equal to or greater than a threshold value.

6. 5. The wireless communication method according to claim 4, A wireless communication method, comprising: classifying a sub-division area as an outlier sub-division area when the proportion of outliers present in each sub-division area is equal to or greater than a threshold; classifying the sub-division area as a normal range when the proportion of outliers present in each sub-division area is less than the threshold; executing a learning process for a binary classifier using training data; and determining that an abnormality has been detected for the moving vehicle when the position of the moving vehicle is included in the outlier sub-division area when the position of the moving vehicle is classified using the binary classifier.

7. A control station that is connected to a plurality of moving vehicles via wireless communication to configure a control system, The control station includes an abnormality detection device that detects an abnormality in the wireless communication, The abnormality detection device includes: periodically receiving position information and measurement results of wireless communication quality information indicating quality of the wireless communication from the plurality of moving vehicles; defining travel trajectories of the plurality of moving vehicles and an area including the travel trajectories based on the position information reported from the plurality of moving vehicles, and dividing the area into a plurality of small divided areas; among the measurement results reported from the plurality of moving vehicles, a first time range is set as a learning object, and measurement results belonging to the first time range are grouped based on the position information so as to correspond to the plurality of small divided areas; and learning a distribution of the wireless communication quality information for each of the grouped small divided areas; acquiring measurement results belonging to a second time range different from the first time range among the measurement results reported from the plurality of moving vehicles by subjecting the measurement results to a detection process, and deriving an existence ratio of outliers in the measurement results belonging to the second time range for each small divided area using the learned distribution of the wireless communication quality information for each small divided area; the anomaly detection device detects an occurrence of an anomaly in wireless communication in each of the plurality of small divided areas in the second time range based on the existence ratio of the outliers; the abnormality detection device has a display unit that displays the area, each of the plurality of small divided areas, and a travel trajectory of the moving vehicle, and highlights a small divided area among the plurality of small divided areas in which an abnormality occurs in wireless communication. Control station.

8. 8. The control station according to claim 7, wherein the anomaly detection device determines that an anomaly has occurred in the wireless communication system when there is a small divided area in which the existence ratio of the outliers is equal to or greater than a threshold value.

9. 8. The control station according to claim 7, The anomaly detection device classifies a sub-divided area as an outlier sub-divided area when the proportion of outliers in each sub-divided area is equal to or greater than a threshold, and performs a learning process for a binary classifier using data in which the sub-divided area is classified as a normal range when the proportion of outliers is less than the threshold, and when classifying the position of the moving vehicle using the binary classifier, determines that an anomaly has been detected for the moving vehicle if the position of the moving vehicle is included in the outlier sub-divided area.

Images