WO2018163375A1 - Control device, control system, and server - Google Patents

Control device, control system, and server Download PDF

Info

Publication number
WO2018163375A1
WO2018163375A1 PCT/JP2017/009572 JP2017009572W WO2018163375A1 WO 2018163375 A1 WO2018163375 A1 WO 2018163375A1 JP 2017009572 W JP2017009572 W JP 2017009572W WO 2018163375 A1 WO2018163375 A1 WO 2018163375A1
Authority
WO
WIPO (PCT)
Prior art keywords
normal operation
control
operation range
terminals
unit
Prior art date
Application number
PCT/JP2017/009572
Other languages
French (fr)
Japanese (ja)
Inventor
中川　慎二
Original Assignee
株式会社日立製作所
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 株式会社日立製作所 filed Critical 株式会社日立製作所
Priority to PCT/JP2017/009572 priority Critical patent/WO2018163375A1/en
Publication of WO2018163375A1 publication Critical patent/WO2018163375A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G05—CONTROLLING; REGULATING
    • G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
    • G05B23/00—Testing or monitoring of control systems or parts thereof
    • G05B23/02—Electric testing or monitoring

Definitions

  • the present invention relates to a control device, a control system, and a server.
  • Patent Document 1 JP 2014-186631
  • This document describes a diagnosis processing system in which a terminal device mounted on a self-propelled machine and a server installed in a management center are connected by a wireless communication line, and the terminal device is connected to the self-propelled machine.
  • a data receiving unit that receives data from a sensor provided; and a first diagnosis unit that diagnoses an abnormality of the self-propelled machine, wherein the server diagnoses an abnormality of the self-propelled machine.
  • One of the first diagnosis unit and the second diagnosis unit performs a primary diagnosis of the abnormality of the self-propelled machine based on the data received by the data receiving unit and the primary diagnosis.
  • the diagnostic processing system is characterized in that the result of the above is transmitted to the other and the other receiving the result of the primary diagnosis performs the secondary diagnosis based on the result of the primary diagnosis. " (Refer to [Claim 1]).
  • Patent Document 2 JP-A-2016-128971.
  • the predictive diagnosis system acquires sensor data from multiple sensors installed in mechanical equipment as time-series data, and uses statistical techniques using the time-series data as learning data.
  • a state measure calculation unit that calculates a state measure that is an index indicating a state
  • an approximate expression calculation unit that calculates an approximate expression that approximates a transition of the state measure from the past to the present by a polynomial
  • an approximate expression A state measure estimating unit that estimates a state measure up to a predetermined time in the future, and a reference period setting unit that sets a period of a state measure to be referred to calculate an approximate expression.
  • the first period including the acquisition time of the latest time-series data or the second period including the acquisition time of the latest time-series data shorter than the first period is set. Solution Reference).
  • Patent Document 1 shares diagnosis processing between a terminal and a server
  • Patent Document 2 detects an abnormality of a single mechanical device. Yes, if there are multiple terminals, the abnormal range of each terminal includes an abnormal range common to each terminal and a specific abnormal range caused by individual differences, environmental differences, user characteristics differences, changes over time, etc. It is not a diagnosis method or an abnormality detection method that takes into account certain things.
  • An object of the present invention is to provide a control device or the like that can improve control reliability.
  • the present invention provides a receiving unit that receives a first normal operating range common to a plurality of terminals, a control unit that controls a machine, and an operation state of the control unit is the first operating range.
  • An abnormality detection unit that determines that the control unit is abnormal when it is not in the normal operation range.
  • FIG. 3 is an overall view of a control system in Embodiments 1 to 3. It is the figure which showed the terminal (control apparatus) and controlled object in Embodiment 1.
  • 6 is a system configuration diagram of a terminal (control device) in Embodiments 1 to 5.
  • FIG. 6 is a system configuration diagram of a server in Embodiments 1 to 5.
  • FIG. 6 is a diagram showing processing of a normal operation range learning unit common to terminals in Embodiments 1 to 4.
  • FIG. 6 is a diagram showing processing of data dividing means in the first to fourth embodiments. It is a figure which shows the example of a process result of a data division
  • FIG. 6 is a diagram showing processing of a normal operation range setting unit common to terminals in Embodiments 1 to 4.
  • FIG. 5 is a diagram illustrating processing of an abnormality detection unit in the first to third embodiments. It is the figure which showed the terminal (control apparatus) and control object in Embodiment 2.
  • FIG. It is the figure which showed the terminal (control apparatus) and control object in Embodiment 3.
  • 6 is an overall view of a control system in Embodiments 4 to 5.
  • FIG. 6 is a diagram showing terminals (control devices) and controlled objects in Embodiments 4 to 5.
  • FIG. 10 is a diagram illustrating processing of a normal operation range learning unit unique to a terminal in the fourth embodiment. It is the figure which showed the process of the data division
  • FIG. 10 is a diagram illustrating processing of a normal operation range learning unit unique to a terminal in the fourth embodiment. It is the figure which showed the process of the data division
  • FIG. 10 is a diagram illustrating processing of a normal operation range learning unit unique to a terminal in the fourth embodiment. It is the
  • FIG. 10 is a diagram illustrating processing of a normal operation range setting unit unique to a terminal in the fourth embodiment. It is the figure which showed the process of the abnormality detection means in Embodiment 4.
  • FIG. 10 is a diagram illustrating processing of a normal operation range learning unit common to terminals in the fifth embodiment.
  • FIG. 10 is a diagram illustrating processing of a normal operation range learning unit unique to a terminal in the fifth embodiment. It is the figure which showed the process of the abnormality detection means in Embodiment 5. It is the figure which showed the terminal in the modification 1. It is a figure which shows the table used for the terminal in the modification 2.
  • the terminal controls machines such as a robot, an autonomous driving vehicle, and a drone (aircraft).
  • the same numerals indicate the same parts.
  • a normal operation range learning unit that learns a first normal operation range common to the plurality of terminals;
  • An embodiment including an abnormality determination unit that determines that the operation of one terminal among the plurality of terminals is abnormal when the operation state of the one terminal is not in the first normal operation range.
  • the first normal range learning is performed on the server, and the abnormality detection is performed on each terminal.
  • the first normal range is set based on the result of machine learning based on past data.
  • the control system is a device that controls a plurality of robots.
  • FIG. 1 is a diagram showing the entire control system.
  • the plurality of terminals 1 transmit and receive information via the server 4.
  • the server 4 includes a normal operation range learning unit 5 common to terminals.
  • the normal operation range learning unit 5 common normal operation range learning unit
  • the normal operation range learning unit 5 common normal operation range learning unit
  • the details of the normal operation range learning unit 5 common to the terminals will be described later with reference to FIG.
  • the terminal 1 includes an abnormality detection means 2 and a control means 3.
  • FIG. 2 shows a robot 201 on the production line controlled by the terminal 1.
  • the control unit 3 calculates an operation amount (for example, a target angle, a target speed, a target torque, etc.) for controlling the robot 201.
  • the control means 3 controls the robot 201 (machine).
  • the detailed specification of the control means 3 for controlling the robot 201 has many known techniques and will not be described in detail here.
  • FIG. 3 is a system configuration diagram of the terminal 1.
  • the terminal 1 is provided with an input circuit 16 for processing an external signal.
  • the signal from the outside here refers to, for example, a sensor signal installed in the terminal and a normal range common to the terminal transmitted from the server.
  • These external signals are sent to the input / output port 17 through the input circuit 16 as input signals.
  • Each input information sent to the input / output port is written into the RAM 14 (Random Access Memory) through the data bus 15. Alternatively, it is stored in the storage device 11.
  • RAM 14 Random Access Memory
  • the input circuit 16 receives a normal operating range common to terminals (first normal operating range).
  • the ROM 13 Read Only Memory
  • the storage device 11 stores processing described later and is executed by the CPU 12 (Central Processing Unit).
  • the value written in the RAM 14 or the storage device 11 is used as appropriate for calculation.
  • information (value) to be sent to the outside is sent to the input / output port 17 through the data bus 15 and sent to the output circuit 18 as an output signal.
  • the output signal is output to the outside as a signal from the output circuit 18 to the outside.
  • the signal to the outside here refers to an actuator signal for causing the control target to make a desired movement, and an operation range of each terminal to be transmitted to the server 4.
  • FIG. 4 is a system configuration diagram of the server 4.
  • the server 4 is provided with an input circuit 26 for processing an external signal.
  • the signal from the outside here is operation information from each terminal.
  • An external signal is sent to the input / output port 27 as an input signal through the input circuit 26.
  • the input information sent to the input / output port is written into the RAM 24 through the data bus 25. Alternatively, it is stored in the storage device 21.
  • the input circuit 26 receives operation information (operation state) of each terminal 1 from a plurality of terminals 1 having a control unit 3 (control unit) that controls the robot 201 (machine).
  • the operation state of the terminal 1 includes the sensor output (input value) input to the control performed by the control means 3 (control unit) of each terminal 1, the operation amount (output value) output from the control, and the control parameter. Indicated by at least one of the values.
  • the control parameter value is a parameter of a function that determines an output value (operation amount) from an input value (sensor output).
  • Processing described below is written in the ROM 23 or the storage device 21 and is executed by the CPU 22. At that time, the value written in the RAM 24 or the storage device 21 is used as appropriate for calculation. Of the calculation results, information (value) to be sent to the outside is sent to the input / output port 27 through the data bus 25 and sent to the output circuit 28 as an output signal. The output signal is output to the outside as a signal from the output circuit 28 to the outside.
  • the signal to the outside here is a normal operation range common to terminals, and is sent to a plurality of terminals 1.
  • the output circuit 28 transmits the normal operation range common to the terminals (first normal operation range) to the plurality of terminals 1. Details of each process will be described below.
  • FIG. 5 is a diagram showing the entire normal operation range learning unit 5 common to terminals, and includes the following calculation means (calculation unit).
  • Data dividing means 100 Normal operation range setting means 110
  • the data dividing unit 100 divides the distribution of control parameter values and sensor output values into predetermined chunks (clusters).
  • the normal operation range setting unit 110 defines each divided cluster within a predetermined range.
  • the control parameter value and the sensor output value used here are selected to define a normal operation range common to terminals. For example, it is conceivable to use data obtained by prior verification before operating the system (software).
  • ⁇ Data division means (FIG. 6)>
  • the division information is calculated using the data of the control parameter value and the sensor output value. Specifically, it is shown in FIG.
  • the obtained division information is output by the k-means method.
  • the division information here refers to the following information.
  • FIG. 7 shows an example of the result of clustering a data group consisting of a certain two-dimensional vector with 4 clusters by the k-means method.
  • ⁇ Normal operation range setting means (FIG. 8)>
  • a data range is set for each divided data set using the above-described division information calculated by the data division unit 100, and the result is output as range information. Specifically, it is shown in FIG.
  • the minimum value of each dimension of the data belonging to each cluster is set as the lower limit of each dimension in the range corresponding to each cluster.
  • the maximum value of each dimension of data belonging to each cluster is set as the upper limit of each dimension in the range corresponding to each cluster.
  • the range information here refers to the lower limit value and the upper limit value of each dimension that define the range corresponding to each cluster (center vector).
  • the normal operation range learning unit 5 (common normal operation range learning unit) common to the terminals obtains the normal operation range (first normal operation range) from the past operation states of the plurality of terminals 1 by machine learning. Set. Note that statistical processing may be used instead of machine learning. Thereby, the reliability of the normal operation range common to the terminals is improved.
  • the center vector coordinates and range information are transmitted from the server 4 to each terminal 1.
  • FIG. 9 shows an example of the result of setting the range by the setting means based on the division result shown in FIG.
  • FIG. 10 ⁇ Abnormality detection means (FIG. 10) Detects abnormal control operation. Specifically, it is shown in FIG.
  • the area indicated by the range 1 to 4 is the normal operating range.
  • a control parameter value (vector) to be detected exists in the area corresponding to the specified center vector, it is determined to be normal, and if not, it is determined to be abnormal. For example, vector A in FIG. 10 is determined to be normal, and vector B is determined to be abnormal.
  • the abnormality detection unit 2 indicates that the control unit 3 is abnormal when the operation state of the control unit 3 (control unit) is not in the normal operation range common to the terminals (first normal operation range). Judge that there is. Thereby, the abnormality detection which considered the normal operation
  • the parameter to be detected may be an operation amount corresponding to the output of the control means 3. Also, an internal parameter calculated inside the control means 3 may be used. Moreover, the sensor output installed in the control object used with the control means 3 may be sufficient.
  • FIG. 10 shows a two-dimensional case, but it can be expanded to N dimensions (N: natural number).
  • the normality / abnormality of the parameters related to the control of each terminal is determined using the normal range common to the terminals that respectively control the plurality of robots in the production line. Reliability is improved. That is, the reliability of control can be improved.
  • a normal operation range learning unit that learns a first normal operation range common to the plurality of terminals;
  • An embodiment including an abnormality determination unit that determines that the operation of one terminal among the plurality of terminals is abnormal when the operation state of the one terminal is not in the first normal operation range.
  • control system is a device that controls a plurality of autonomous vehicles.
  • FIG. 1 is a diagram showing the entire control system, but since it is the same as that of the first embodiment, it will not be described in detail.
  • FIG. 11 shows the terminal 1 and the automatic driving vehicle 202 controlled by the terminal 1.
  • the control means 3 calculates an operation amount (for example, a target speed, a target rotation speed, etc.) for controlling the autonomous driving vehicle 202.
  • the detailed specifications of the control means 3 for controlling the autonomous vehicle 202 are not described here because there are many known techniques.
  • FIG. 3 is a system configuration diagram of the terminal 1, but since it is the same as that of the first embodiment, it will not be described in detail.
  • FIG. 4 is a system configuration diagram of the server 4, which is the same as that of the first embodiment and will not be described in detail. Details of each process will be described below.
  • FIG. 5 is a diagram showing the entire normal operation range learning unit 5 common to terminals, but since it is the same as that of the first embodiment, it will not be described in detail.
  • ⁇ Data division means (FIG. 6)>
  • the division information is calculated using the data of the control parameter value and the sensor output value. Specifically, although it is shown in FIG. 6, it is the same as that of the first embodiment, and therefore will not be described in detail.
  • ⁇ Abnormality detection means (FIG. 10) Detects abnormal control operation. Specifically, since it is the same as the first embodiment shown in FIG. 10, it will not be described in detail.
  • the normality / abnormality of the parameters related to the control of each terminal is determined using the normal range common to the terminals that respectively control the autonomous driving vehicle. improves.
  • a normal operation range learning unit that learns a first normal operation range common to the plurality of terminals;
  • An embodiment including an abnormality determination unit that determines that the operation of one terminal among the plurality of terminals is abnormal when the operation state of the one terminal is not in the first normal operation range.
  • control system is a device that controls a plurality of drones.
  • FIG. 1 is a diagram showing the entire control system, but since it is the same as that of the first embodiment, it will not be described in detail.
  • FIG. 12 shows the terminal 1 and the drone 203 controlled by the terminal 1.
  • An operation amount for example, a target rotation speed of each rotor
  • the detailed specification of the control means 3 for controlling the drone 203 is not described in detail here because there are many known techniques.
  • FIG. 3 is a system configuration diagram of the terminal 1, but since it is the same as that of the first embodiment, it will not be described in detail.
  • FIG. 4 is a system configuration diagram of the server 4, which is the same as that of the first embodiment and will not be described in detail. Details of each process will be described below.
  • FIG. 5 is a diagram showing the entire normal operation range learning unit 5 common to terminals, but since it is the same as that of the first embodiment, it will not be described in detail.
  • ⁇ Data division means (FIG. 6)>
  • the division information is calculated using the data of the control parameter value and the sensor output value. Specifically, although it is shown in FIG. 6, it is the same as that of the first embodiment, and therefore will not be described in detail.
  • ⁇ Abnormality detection means (FIG. 10) Detects abnormal control operation. Specifically, since it is the same as the first embodiment shown in FIG. 10, it will not be described in detail.
  • the normality / abnormality of the parameters related to the control of each terminal is determined using the normal range common to the terminals that respectively control the drone, so that the reliability of the entire system is improved.
  • a normal operation range learning unit that learns a first normal operation range common to the plurality of terminals;
  • An embodiment including an abnormality determination unit that determines that the operation of one terminal among the plurality of terminals is abnormal when the operation state of the one terminal is not in the first normal operation range.
  • the present embodiment also includes normal operation range learning means for learning a second normal operation range unique to the terminal due to individual differences, environment, user characteristics, changes with time, and the like.
  • the second normal operation range is learned based on the parameter value relating to the control of each terminal or the sensor information provided in each terminal.
  • the second normal range learning is performed at each terminal.
  • FIG. 13 is a diagram showing the entire control system.
  • the plurality of terminals 1 transmit and receive information via the server 4.
  • the server 4 includes a normal operation range learning unit 5 common to terminals.
  • the terminal 1 includes a normal operation range learning unit 6, an abnormality detection unit 7, and a control unit 3 unique to the terminal.
  • the normal operation range learning unit 6 (specific normal operation range learning unit) unique to the terminal learns the normal operation range (second normal operation range) unique to the control means 3 (control unit). Specifically, the terminal-specific normal operation range learning unit 6 learns a normal operation range (second normal operation range) caused by at least one of individual differences, environment, user characteristics, and changes with time. Details of the normal operation range learning unit 6 unique to the terminal will be described later with reference to FIG.
  • FIG. 14 shows a robot 201 on the production line controlled by the terminal 1.
  • An operation amount for example, a target angle, a target speed, a target torque, etc.
  • the detailed specification of the control means 3 for controlling the robot 201 has many known techniques and will not be described in detail here.
  • FIG. 3 is a system configuration diagram of the terminal 1, but since it is the same as that of the first embodiment, it will not be described in detail.
  • FIG. 4 is a system configuration diagram of the server 4, which is the same as that of the first embodiment and will not be described in detail. Details of each process will be described below.
  • FIG. 5 is a diagram showing the entire normal operation range learning unit 5 common to terminals, but since it is the same as that of the first embodiment, it will not be described in detail.
  • ⁇ Data division means (FIG. 6)>
  • the division information is calculated using the data of the control parameter value and the sensor output value. Specifically, although it is shown in FIG. 6, it is the same as that of the first embodiment, and therefore will not be described in detail.
  • ⁇ Normal operation range setting means (FIG. 8)>
  • a data range is set for each divided data set using the above-described division information calculated by the data division unit 100, and the result is output as range information.
  • FIG. 8 it is the same as that of the first embodiment, and thus will not be described in detail.
  • FIG. 15 is a diagram showing the entire normal operation range learning unit 6 unique to the terminal, and includes the following calculation units.
  • Data dividing means 102 Normal operation range setting means 112
  • the data dividing unit 102 divides the distribution of control parameter values and sensor output values into predetermined chunks (clusters).
  • the normal operation range setting unit 112 defines each divided cluster within a predetermined range.
  • the control parameter value and the sensor output value used here are selected to define a normal operation range unique to the terminal. For example, it is conceivable to use performance data when the system (software) is operating.
  • the obtained division information is output by the k-means method.
  • the division information here refers to the following information.
  • Cluster number to which data divided by k-means method belongs ⁇ Average value of data belonging to each cluster (center vector) The details of the k-means method are described in many literatures and books, and therefore will not be described in detail here.
  • the minimum value of each dimension of the data belonging to each cluster is set as the lower limit of each dimension in the range corresponding to each cluster.
  • the maximum value of each dimension of data belonging to each cluster is set as the upper limit of each dimension in the range corresponding to each cluster.
  • the range information here refers to the lower limit value and the upper limit value of each dimension that define the range corresponding to each cluster (center vector).
  • the normal operation range learning unit 6 (inherent normal operation range learning unit) unique to the terminal obtains the normal operation range (second normal operation range) from the past operation state of each terminal 1 by machine learning. Set. Note that statistical processing may be used instead of machine learning.
  • the coordinates and range information of the center vector are sent from the normal operation range setting means 112 to the abnormality detection means 7.
  • FIG. 18 ⁇ Abnormality detection means
  • the area indicated by the ranges 1a to 4a is the normal operating range common to terminals.
  • control parameter value (vector) to be detected exists within the area corresponding to the specified center vector, it is determined to be normal (the value of the abnormal flag a is 0), and if it does not exist, It is determined that there is an abnormality (the value of the abnormality flag a is 1).
  • the area indicated by the ranges 1b to 4b is the normal operating range unique to the terminal.
  • control parameter value (vector) to be detected exists within the area corresponding to the specified center vector, it is determined as normal (the value of the abnormality flag b is 0), and if it does not exist, It is determined that there is an abnormality (the value of the abnormality flag b is 1).
  • the abnormality detection unit 7 is configured such that the operation state of the control unit 3 (control unit) is a normal operation range common to the terminals (first normal operation range) or a normal operation range unique to the terminals (second In the normal operation range), it is determined that the control means 3 is abnormal. Thereby, the abnormality detection which considered the normal operation range common to a terminal and the normal operation range specific to a terminal can be performed.
  • the parameter to be detected may be an operation amount corresponding to the output of the control means 3. Also, an internal parameter calculated inside the control means 3 may be used. Moreover, the sensor output installed in the control object used with the control means 3 may be sufficient.
  • FIG. 18 shows a two-dimensional case, but it can be expanded to N dimensions (N: natural number).
  • a normal operation range learning unit that learns a first normal operation range common to the plurality of terminals;
  • An embodiment including an abnormality determination unit that determines that the operation of one terminal among the plurality of terminals is abnormal when the operation state of the one terminal is not in the first normal operation range.
  • a normal operation range learning means for learning a second normal operation range unique to the terminal due to individual differences, environment, user characteristics, changes with time, and the like.
  • the second normal operation range is learned based on the parameter value relating to the control of each terminal or the sensor information provided in each terminal.
  • the second normal range learning is performed at each terminal.
  • the first normal range and the second normal operating range are determined.
  • the normal range is set on a rule basis.
  • the first normal range (first normal operation range) and the second normal range (at least one of the second normal operation ranges may be set on a rule basis. Anomaly detection can be performed based on empirically obtained rules.
  • FIG. 13 is a diagram showing the entire control system, but since it is the same as that of the fourth embodiment, it will not be described in detail.
  • FIG. 14 shows the robot 201 on the production line controlled by the terminal 1, but since it is the same as that of the fourth embodiment, it will not be described in detail.
  • the detailed specification of the control means 3 for controlling the robot 201 has many known techniques and will not be described in detail here.
  • FIG. 3 is a system configuration diagram of the terminal 1, but since it is the same as that of the first embodiment, it will not be described in detail.
  • FIG. 4 is a system configuration diagram of the server 4, which is the same as that of the first embodiment and will not be described in detail. Details of each process will be described below.
  • FIG. 19 is a diagram showing a normal operation range learning unit 5 common to terminals, and includes the following calculation units.
  • control parameter 1 The minimum value of control parameter 1 is K1a_L and the maximum value is K1a_H
  • control parameter 2 is K2a_L and the maximum value is K2a_H
  • control parameter n Kna_L and the maximum value is Kna_H
  • control parameter values and sensor output values used here are selected to define the normal operating range common to terminals. For example, it is conceivable to use data obtained by prior verification before operating the system (software).
  • FIG. 20 is a diagram showing the normal operation range learning unit 6 unique to the terminal, and includes the following calculation units.
  • control parameter 1 The minimum value of control parameter 1 is K1b_L and the maximum value is K1b_H
  • control parameter 2 is K2b_L and the maximum value is K2b_H
  • control parameter n The minimum value of the control parameter n is Knb_L and the maximum value is Knb_H
  • control parameter values and sensor output values used here are selected to define the normal operating range specific to the terminal. For example, it is conceivable to use performance data when the system (software) is operating.
  • FIG. 21 ⁇ Abnormality detection means (FIG. 21)> Detects abnormal control operation. Specifically, it is shown in FIG.
  • abnormality detection is performed considering that there is a unique abnormality range, both control performance and reliability of a control system having a plurality of terminals are improved.
  • the abnormality detection is performed on a rule basis in the present embodiment, the explanation at the time of abnormality detection in which the abnormality detection method is explicitly given is also improved.
  • Modification 1 In the first modification, the control means 3 of the terminal 1 shown in FIG. 22 performs predetermined control according to an abnormality flag that is an output of the abnormality detection means 2 (abnormality detection unit).
  • the control means 3 may output a warning to the display device or the like or perform predetermined fail-safe control. Thereby, when abnormality is detected, appropriate control can be performed.
  • Modification 2 In the second modification, the control unit 3 of the terminal 1 performs predetermined control according to the determination result of the abnormality detection unit 2 (abnormality detection unit).
  • FIG. 22 shows a combination of whether or not the operation state of the terminal 1 is in the first normal operation range and whether or not the operation state of the terminal 1 is in the second normal operation range, and the abnormality detection means 2 (abnormality detection unit). It is a figure which shows the table 300 which matches and memorize
  • the table 300 is stored in the storage device 11 of the terminal 1, for example, but may be stored in the storage device 21 of the server 4.
  • the abnormality detection means 2 determines that the control means 3 is normal when the operation state of the terminal 1 is within the first normal operation range and within the second normal operation range.
  • the abnormality detection means 2 determines that the control means 3 is abnormal (abnormal 1) when the operation state of the terminal 1 is outside the first normal operation range and within the second normal operation range.
  • the abnormality detection unit 2 determines that the control unit 3 is abnormal (abnormal 2) when the operation state of the terminal 1 is within the first normal operation range and out of the second normal operation range.
  • the abnormality detecting means 2 determines that the control means 3 is abnormal (abnormal 3) when the operating state of the terminal 1 is outside the first normal operating range and outside the second normal operating range.
  • the control unit 3 (control unit) identifies an identifier (normal control, control 1 to 3) identifier corresponding to the determination result (normal, abnormality 1 to 3) of the abnormality detection unit 2 (abnormality detection unit) from the table 300. ID) is read out, and control corresponding to this identifier is executed.
  • control 1 a warning may be output to a display device or the like, in control 2 fail-safe control may be performed, and in control 3 control may be stopped. Thereby, when abnormality is detected, appropriate control according to abnormality can be performed.
  • Control 1 to control 3 may be the same control (for example, output of warning).
  • the present invention is not limited to the above-described embodiment, and includes various modifications.
  • the above-described embodiment has been described in detail for easy understanding of the present invention, and is not necessarily limited to the one having all the configurations described.
  • a part of the configuration of an embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of an embodiment.
  • each of the above-described configurations, functions (means), etc. may be realized by hardware by designing a part or all of them, for example, by an integrated circuit.
  • Each of the above-described configurations, functions (means), and the like may be realized by software by interpreting and executing a program that realizes each function (means) or the like by a processor (CPU).
  • Information such as programs, tables, and files that realize each function (means) is stored in a memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC card, SD card, or DVD. be able to.
  • control system including a plurality of terminals each controlling a device and communicating with a server
  • the control system learns a first normal operation range common to the plurality of terminals.
  • an abnormality determination unit that determines that the operation of the one terminal is abnormal when an operation state of one of the plurality of terminals is not in the first normal operation range.
  • a control system comprising a normal operation range learning means for learning a second normal operation range unique to a terminal caused by individual differences, environment, user characteristics, changes with time, and the like.
  • a normal operation range learning means for learning a second normal operation range unique to the terminal due to individual differences, environment, user characteristics, changes with time, etc.
  • the second normal operation range is A control system that learns based on at least a parameter value related to control of each terminal, information from a sensor provided in each terminal, or information obtained by communication at the terminal.
  • At least the first normal range or the second normal range sets a past operation state of the terminal based on a result of processing by statistical processing or machine learning. And control system.
  • control system is a device for controlling a robot.
  • control system is a device for controlling an autonomous driving vehicle.
  • control system is a device for controlling a flying object such as a drone.
  • an abnormality taking into account that there is an abnormal range common to each terminal and a specific abnormal range due to individual differences, environmental differences, user characteristic differences, changes with time, etc. Since detection is performed, the reliability of a control system having a plurality of terminals can be improved.
  • Processing of abnormality detection means (based on normal operating range common to terminals) 52. Processing of abnormality detection means (based on normal operation range unique to the terminal) 61 ... Processing of abnormality detection means (based on normal operating range common to terminals) 62 ... Processing of abnormality detection means (based on normal operation range unique to terminal) 100: Data division means (normal operation range learning unit common to terminals) 101: Processing of data dividing means (normal operation range learning unit common to terminals) 102: Data dividing means (terminal-specific normal operating range learning unit) 103 ...
  • Normal operation range setting means normal operation range learning unit common to terminals
  • Processing of normal operation range setting means normal operation range learning unit common to terminals
  • 112 Processing of normal operation range setting means (normal operation range learning unit common to terminals)
  • Normal operation range setting means terminal-specific normal operation range learning unit
  • 113 Processing of normal operation range setting means (terminal-specific normal operation range learning unit) 121 ...
  • Rule-based threshold learning normal operation range learning unit common to terminals
  • Rule-based threshold learning a condition-based

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Automation & Control Theory (AREA)
  • Testing And Monitoring For Control Systems (AREA)

Abstract

Provided are a control device, a control system, and a server capable of enhancing reliability of control. A terminal (1) (control device) is provided with an input circuit (16) (reception unit), a control means (3) (control unit), and an abnormality detection means (2) (abnormality detection unit). The input circuit (16) (reception unit) receives a normal operation range (first normal operation range) common to a plurality of terminals (1). The control means (3) (control unit) controls a machine such as a robot (201), a self-driving car (202), and a drone (203) (flying body). In the case where an operation state of the control means (3) (control unit) is outside of the normal operation range (first normal operation range) common to the terminals, the abnormality detection means (2) (abnormality detection unit) determines that the control means (3) is abnormal.

Description

制御装置、制御システム及びサーバーControl device, control system and server
　本発明は、制御装置、制御システム及びサーバーに関する。 The present invention relates to a control device, a control system, and a server.
　本技術分野の背景技術として、特開2014-186631（特許文献１）がある。この文献には、「自走式機械に搭載された端末装置と管理センタに設置されたサーバとを無線通信回線によって接続した診断処理システムであって、前記端末装置は、前記自走式機械に設けられたセンサからデータを受信するデータ受信部と、前記自走式機械の異常を診断する第１診断部と、を備え、前記サーバは、前記自走式機械の異常を診断する第２診断部を備え、前記第１診断部と前記第２診断部のうち一方が、前記データ受信部にて受信された前記データに基づいて前記自走式機械の異常の一次診断を行うと共に当該一次診断の結果を他方に送信し、前記一次診断の結果を受信した他方が、当該一次診断の結果に基づいて二次診断を行うようにしたことを特徴とする診断処理システム。」と記載されている（〔請求項１〕参照）。 As a background art in this technical field, there is JP 2014-186631 (Patent Document 1). This document describes a diagnosis processing system in which a terminal device mounted on a self-propelled machine and a server installed in a management center are connected by a wireless communication line, and the terminal device is connected to the self-propelled machine. A data receiving unit that receives data from a sensor provided; and a first diagnosis unit that diagnoses an abnormality of the self-propelled machine, wherein the server diagnoses an abnormality of the self-propelled machine. One of the first diagnosis unit and the second diagnosis unit performs a primary diagnosis of the abnormality of the self-propelled machine based on the data received by the data receiving unit and the primary diagnosis. The diagnostic processing system is characterized in that the result of the above is transmitted to the other and the other receiving the result of the primary diagnosis performs the secondary diagnosis based on the result of the primary diagnosis. " (Refer to [Claim 1]).
　また、特開2016-128971（特許文献２）がある。この文献には、「予兆診断システムは、機械設備に設置した複数のセンサからのセンサデータを時系列データとして取得し、時系列データを学習データとした統計的手法により、機械設備の異常や性能などの状態を示す指標である状態測度を算出する状態測度算出部と、過去から現在までの状態測度の推移を、多項式により近似した近似式を算出する近似式算出部と、近似式を用いて、将来の所定の時点までの状態測度を推定する状態測度推定部と、近似式を算出するために参照する状態測度の期間を設定する参照期間設定部とを備え、参照期間設定部は、参照期間として最新の時系列データの取得時刻を含む第１期間か、第１期間よりも短く、最新の時系列データの取得時刻を含む第２期間かを設定する。」と記載されている（〔解決手段〕参照）。 There is also JP-A-2016-128971 (Patent Document 2). This document states that “The predictive diagnosis system acquires sensor data from multiple sensors installed in mechanical equipment as time-series data, and uses statistical techniques using the time-series data as learning data. A state measure calculation unit that calculates a state measure that is an index indicating a state, an approximate expression calculation unit that calculates an approximate expression that approximates a transition of the state measure from the past to the present by a polynomial, and an approximate expression A state measure estimating unit that estimates a state measure up to a predetermined time in the future, and a reference period setting unit that sets a period of a state measure to be referred to calculate an approximate expression. As the period, the first period including the acquisition time of the latest time-series data or the second period including the acquisition time of the latest time-series data shorter than the first period is set. Solution Reference).
特開2014-186631号公報JP 2014-186631 特開2016-128971号公報Japanese Unexamined Patent Publication No. 2016-128971
　しかしながら、前述の先行技術（特許文献１）は、端末とサーバーで、診断処理を分担するものであり、前述の先行技術（特許文献２）は、単一の機械装置の異常を検出するものであり、複数の端末がある場合、個々の端末の異常範囲には、各端末共通の異常範囲と、各端末の個体差、環境差、ユーザー特性差、経時変化等に起因する固有の異常範囲があることを考慮した診断方式あるいは異常検知方式ではない。 However, the above-described prior art (Patent Document 1) shares diagnosis processing between a terminal and a server, and the above-mentioned prior art (Patent Document 2) detects an abnormality of a single mechanical device. Yes, if there are multiple terminals, the abnormal range of each terminal includes an abnormal range common to each terminal and a specific abnormal range caused by individual differences, environmental differences, user characteristics differences, changes over time, etc. It is not a diagnosis method or an abnormality detection method that takes into account certain things.
　本発明の目的は、制御の信頼性を向上することができる制御装置等を提供することにある。 An object of the present invention is to provide a control device or the like that can improve control reliability.
　上記目的を達成するために、本発明は、複数の端末に共通の第１の正常動作範囲を受信する受信部と、機械の制御を行う制御部と、前記制御部の動作状態が前記第１の正常動作範囲にない場合、前記制御部が異常であると判定する異常検知部と、を備えることを特徴とする制御装置。 In order to achieve the above object, the present invention provides a receiving unit that receives a first normal operating range common to a plurality of terminals, a control unit that controls a machine, and an operation state of the control unit is the first operating range. An abnormality detection unit that determines that the control unit is abnormal when it is not in the normal operation range.
　本発明によれば、制御の信頼性を向上することができる。上記した以外の課題、構成及び効果は、以下の実施形態の説明により明らかにされる。 According to the present invention, the reliability of control can be improved. Problems, configurations, and effects other than those described above will be clarified by the following description of embodiments.
実施形態１～３における制御システムの全体図である。FIG. 3 is an overall view of a control system in Embodiments 1 to 3. 実施形態１における端末（制御装置）と制御対象を示した図である。It is the figure which showed the terminal (control apparatus) and controlled object in Embodiment 1. 実施形態１～５における端末（制御装置）のシステム構成図である。6 is a system configuration diagram of a terminal (control device) in Embodiments 1 to 5. FIG. 実施形態１～５におけるサーバーのシステム構成図である。6 is a system configuration diagram of a server in Embodiments 1 to 5. FIG. 実施形態１～４における端末共通の正常動作範囲学習部の処理を示した図である。FIG. 6 is a diagram showing processing of a normal operation range learning unit common to terminals in Embodiments 1 to 4. 実施形態１～４におけるデータ分割手段の処理を示した図である。FIG. 6 is a diagram showing processing of data dividing means in the first to fourth embodiments. データ分割手段の処理結果例を示す図である。It is a figure which shows the example of a process result of a data division | segmentation means. 実施形態１～４における端末共通の正常動作範囲設定手段の処理を示した図である。FIG. 6 is a diagram showing processing of a normal operation range setting unit common to terminals in Embodiments 1 to 4. 正常動作範囲設定手段の処理結果例を示す図である。It is a figure which shows the example of a process result of a normal operation range setting means. 実施形態１～３における異常検知手段の処理を示した図である。FIG. 5 is a diagram illustrating processing of an abnormality detection unit in the first to third embodiments. 実施形態２における端末（制御装置）と制御対象を示した図である。It is the figure which showed the terminal (control apparatus) and control object in Embodiment 2. FIG. 実施形態３における端末（制御装置）と制御対象を示した図である。It is the figure which showed the terminal (control apparatus) and control object in Embodiment 3. 実施形態４～５における制御システムの全体図である。6 is an overall view of a control system in Embodiments 4 to 5. FIG. 実施形態４～５における端末（制御装置）と制御対象を示した図である。FIG. 6 is a diagram showing terminals (control devices) and controlled objects in Embodiments 4 to 5. 実施形態４における端末固有の正常動作範囲学習部の処理を示した図である。FIG. 10 is a diagram illustrating processing of a normal operation range learning unit unique to a terminal in the fourth embodiment. 実施形態４におけるデータ分割手段の処理を示した図である。It is the figure which showed the process of the data division | segmentation means in Embodiment 4. 実施形態４における端末固有の正常動作範囲設定手段の処理を示した図である。FIG. 10 is a diagram illustrating processing of a normal operation range setting unit unique to a terminal in the fourth embodiment. 実施形態４における異常検知手段の処理を示した図である。It is the figure which showed the process of the abnormality detection means in Embodiment 4. 実施形態５における端末共通の正常動作範囲学習部の処理を示した図である。FIG. 10 is a diagram illustrating processing of a normal operation range learning unit common to terminals in the fifth embodiment. 実施形態５における端末固有の正常動作範囲学習部の処理を示した図である。FIG. 10 is a diagram illustrating processing of a normal operation range learning unit unique to a terminal in the fifth embodiment. 実施形態５における異常検知手段の処理を示した図である。It is the figure which showed the process of the abnormality detection means in Embodiment 5. 変形例１における端末を示した図である。It is the figure which showed the terminal in the modification 1. 変形例２における端末に用いられるテーブルを示す図である。It is a figure which shows the table used for the terminal in the modification 2.
　以下、図面を用いて、本発明の実施形態１～５による端末（制御装置）を含む制御システムの構成及び動作について説明する。端末は、ロボット、自動運転車、ドローン（飛行体）等の機械を制御する。なお、各図において、同一符号は同一部分を示す。 Hereinafter, the configuration and operation of a control system including a terminal (control device) according to Embodiments 1 to 5 of the present invention will be described with reference to the drawings. The terminal controls machines such as a robot, an autonomous driving vehicle, and a drone (aircraft). In each figure, the same numerals indicate the same parts.
　（実施形態１）
　本実施形態においては、各々が機器を制御し、サーバーと通信を行う複数の端末を備えた制御システムにおいて、前記複数の端末共通の第１の正常動作範囲を学習する正常動作範囲学習部と、前記複数の端末中の一の端末の動作状態が前記第１の正常動作範囲にないとき、前記一の端末の動作が異常と判断する異常判断部とを備えた形態について示す。 (Embodiment 1)
In the present embodiment, in a control system including a plurality of terminals each controlling a device and communicating with a server, a normal operation range learning unit that learns a first normal operation range common to the plurality of terminals; An embodiment including an abnormality determination unit that determines that the operation of one terminal among the plurality of terminals is abnormal when the operation state of the one terminal is not in the first normal operation range.
　また、第１の正常範囲学習はサーバーで実施し、異常検知は、各端末で実施する。 Also, the first normal range learning is performed on the server, and the abnormality detection is performed on each terminal.
　また、第１の正常範囲は、過去のデータに基づいて、機械学習の結果に基づいて設定する。 Also, the first normal range is set based on the result of machine learning based on past data.
　また、制御システムは、複数のロボットを制御する装置である。 The control system is a device that controls a plurality of robots.
　図１は、制御システムの全体を表した図である。複数の端末１は、サーバー４を介して、情報の送受信を行う。サーバー４では、端末共通の正常動作範囲学習部５を備える。換言すれば、端末共通の正常動作範囲学習部５（共通正常動作範囲学習部）は、複数の端末に共通の正常動作範囲（第１の正常動作範囲）を学習する。なお、端末共通の正常動作範囲学習部５の詳細については、図５を用いて後述する。 FIG. 1 is a diagram showing the entire control system. The plurality of terminals 1 transmit and receive information via the server 4. The server 4 includes a normal operation range learning unit 5 common to terminals. In other words, the normal operation range learning unit 5 (common normal operation range learning unit) common to terminals learns a normal operation range (first normal operation range) common to a plurality of terminals. The details of the normal operation range learning unit 5 common to the terminals will be described later with reference to FIG.
　端末１は、異常検知手段２と制御手段３とを備える。異常検知手段２は、制御パラメータ値、センサ出力などが端末共通の正常動作範囲にあるか否かで異常検知を行い、異常と判定されたときは、異常フラグをオン（＝１）にする。 The terminal 1 includes an abnormality detection means 2 and a control means 3. The abnormality detection means 2 performs abnormality detection based on whether or not the control parameter value, sensor output, etc. are within the normal operation range common to the terminals, and when it is determined that there is an abnormality, the abnormality flag is turned on (= 1).
　図２は、端末１によって制御される生産ラインのロボット２０１を示している。制御手段３は、ロボット２０１を制御するための操作量（例えば、目標角度、目標速度、目標トルクなど）を演算する。換言すれば、制御手段３（制御部）は、ロボット２０１（機械）の制御を行う。なお、ロボット２０１を制御する制御手段３の詳細仕様については、多くの公知技術があるので、ここでは詳述しない。 FIG. 2 shows a robot 201 on the production line controlled by the terminal 1. The control unit 3 calculates an operation amount (for example, a target angle, a target speed, a target torque, etc.) for controlling the robot 201. In other words, the control means 3 (control unit) controls the robot 201 (machine). The detailed specification of the control means 3 for controlling the robot 201 has many known techniques and will not be described in detail here.
　図３は、端末１のシステム構成図である。端末１には、外部からの信号を処理する入力回路１６が設けてある。ここでいう外部からの信号とは、例えば、端末に設置されているセンサ信号とサーバーから送信される端末共通の正常範囲などを指す。これら外部からの信号は、入力回路１６を経て、入力信号となり入出力ポート１７へ送られる。入出力ポートに送られた各入力情報は、データバス１５を通って、ＲＡＭ１４（Random Access Memory）に書き込まれる。あるいは、記憶装置１１に記憶される。 FIG. 3 is a system configuration diagram of the terminal 1. The terminal 1 is provided with an input circuit 16 for processing an external signal. The signal from the outside here refers to, for example, a sensor signal installed in the terminal and a normal range common to the terminal transmitted from the server. These external signals are sent to the input / output port 17 through the input circuit 16 as input signals. Each input information sent to the input / output port is written into the RAM 14 (Random Access Memory) through the data bus 15. Alternatively, it is stored in the storage device 11.
　例えば、入力回路１６（受信部）は、端末共通の正常動作範囲（第１の正常動作範囲）を受信する。 For example, the input circuit 16 (receiver) receives a normal operating range common to terminals (first normal operating range).
　ＲＯＭ１３（Read Only Memory）もしくは記憶装置１１には、後述の処理が書き込まれていて、ＣＰＵ１２（Central Processing Unit）で実行される。その際、ＲＡＭ１４あるいは記憶装置１１に書き込まれた値を、適宜、用いて演算を行う。演算結果の内、外部へ送り出す情報（値）は、データバス１５を通って、入出力ポート１７に送られ、出力信号として、出力回路１８に送られる。出力信号は、出力回路１８から外部への信号として、外部に出力される。ここでいう外部への信号とは制御対象を所望の動きをさせるためのアクチュエータ信号、サーバー４へ送信する各端末の動作範囲を指す。 The ROM 13 (Read Only Memory) or the storage device 11 stores processing described later and is executed by the CPU 12 (Central Processing Unit). At that time, the value written in the RAM 14 or the storage device 11 is used as appropriate for calculation. Of the calculation results, information (value) to be sent to the outside is sent to the input / output port 17 through the data bus 15 and sent to the output circuit 18 as an output signal. The output signal is output to the outside as a signal from the output circuit 18 to the outside. The signal to the outside here refers to an actuator signal for causing the control target to make a desired movement, and an operation range of each terminal to be transmitted to the server 4.
　図４は、サーバー４のシステム構成図である。サーバー４には、外部からの信号を処理する入力回路２６が設けてある。ここでいう外部からの信号とは、各端末からの動作情報である。外部からの信号は、入力回路２６を経て、入力信号となり入出力ポート２７へ送られる。入出力ポートに送られた入力情報は、データバス２５を通って、ＲＡＭ２４に書き込まれる。あるいは、記憶装置２１に記憶される。 FIG. 4 is a system configuration diagram of the server 4. The server 4 is provided with an input circuit 26 for processing an external signal. The signal from the outside here is operation information from each terminal. An external signal is sent to the input / output port 27 as an input signal through the input circuit 26. The input information sent to the input / output port is written into the RAM 24 through the data bus 25. Alternatively, it is stored in the storage device 21.
　例えば、入力回路２６（受信部）は、ロボット２０１（機械）の制御を行う制御手段３（制御部）を有する複数の端末１からそれぞれの端末１の動作情報（動作状態）を受信する。なお、端末１の動作状態は、それぞれの端末１の制御手段３（制御部）が行う制御へ入力されるセンサ出力（入力値）、当該制御から出力される操作量（出力値）、制御パラメータ値のうち少なくとも１つによって示される。ここで、制御パラメータ値は、入力値（センサ出力）から出力値（操作量）を決定する関数のパラメータである。 For example, the input circuit 26 (reception unit) receives operation information (operation state) of each terminal 1 from a plurality of terminals 1 having a control unit 3 (control unit) that controls the robot 201 (machine). Note that the operation state of the terminal 1 includes the sensor output (input value) input to the control performed by the control means 3 (control unit) of each terminal 1, the operation amount (output value) output from the control, and the control parameter. Indicated by at least one of the values. Here, the control parameter value is a parameter of a function that determines an output value (operation amount) from an input value (sensor output).
　ＲＯＭ２３もしくは記憶装置２１には、後述の処理が書き込まれていて、ＣＰＵ２２で実行される。その際、ＲＡＭ２４あるいは記憶装置２１に書き込まれた値を、適宜、用いて演算を行う。演算結果の内、外部へ送り出す情報（値）は、データバス２５を通って、入出力ポート２７に送られ、出力信号として、出力回路２８に送られる。出力信号は、出力回路２８から外部への信号として、外部に出力される。ここでいう外部への信号とは、端末共通の正常動作範囲であり、複数の端末１へ送られる。例えば、出力回路２８（送信部）は、端末共通の正常動作範囲（第１の正常動作範囲）を複数の端末１に送信する。以下、各処理の詳細を説明する。 Processing described below is written in the ROM 23 or the storage device 21 and is executed by the CPU 22. At that time, the value written in the RAM 24 or the storage device 21 is used as appropriate for calculation. Of the calculation results, information (value) to be sent to the outside is sent to the input / output port 27 through the data bus 25 and sent to the output circuit 28 as an output signal. The output signal is output to the outside as a signal from the output circuit 28 to the outside. The signal to the outside here is a normal operation range common to terminals, and is sent to a plurality of terminals 1. For example, the output circuit 28 (transmission unit) transmits the normal operation range common to the terminals (first normal operation range) to the plurality of terminals 1. Details of each process will be described below.
　＜端末共通の正常動作範囲学習部（図５）＞
　図５は端末共通の正常動作範囲学習部５の全体を表した図であり、以下の演算手段（演算部）から構成される。 <Normal operation range learning unit common to terminals (FIG. 5)>
FIG. 5 is a diagram showing the entire normal operation range learning unit 5 common to terminals, and includes the following calculation means (calculation unit).
　・データ分割手段１００
　・正常動作範囲設定手段１１０
　データ分割手段１００は、制御パラメータ値、センサ出力値の分布を所定の塊毎(クラスタ毎)に分割する。正常動作範囲設定手段１１０は、分割した各クラスタを所定の範囲で規定する。ここで用いる制御パラメータ値、センサ出力値は、端末共通の正常動作範囲を規定するものとして選ぶ。例えば、システム（ソフト）を稼働させる前の事前検証で得られるデータを用いることが考えられる。 Data dividing means 100
Normal operation range setting means 110
The data dividing unit 100 divides the distribution of control parameter values and sensor output values into predetermined chunks (clusters). The normal operation range setting unit 110 defines each divided cluster within a predetermined range. The control parameter value and the sensor output value used here are selected to define a normal operation range common to terminals. For example, it is conceivable to use data obtained by prior verification before operating the system (software).
　＜データ分割手段（図６）＞
　本処理では、制御パラメータ値、センサ出力値のデータを用いて、分割情報を演算する。具体的には、図６に示される。 <Data division means (FIG. 6)>
In this process, the division information is calculated using the data of the control parameter value and the sensor output value. Specifically, it is shown in FIG.
　ベクトル化されたデータ群を機械学習k-meansを用いてクラスタリングする。k-means法により、得られた分割情報を出力する。ここでいう分割情報とは、以下の情報を指す。 Clustering vectorized data group using machine learning k-means. The obtained division information is output by the k-means method. The division information here refers to the following information.
　・k-means法によって分割されたデータが属するクラスタ番号
　・各クラスタに属するデータの平均値（中心ベクトル）
　なお、k-means法の詳細については、多くの文献、書籍で述べてあるので、ここでは詳述しない。図７は、ある２次元のベクトルからなるデータ群をk-means法により、クラスタ数４でクラスタリングした結果例である。 ・ Cluster number to which data divided by k-means method belongs ・ Average value of data belonging to each cluster (center vector)
The details of the k-means method are described in many literatures and books, and therefore will not be described in detail here. FIG. 7 shows an example of the result of clustering a data group consisting of a certain two-dimensional vector with 4 clusters by the k-means method.
　＜正常動作範囲設定手段（図８）＞
　本処理では、データ分割手段１００で演算した上述の分割情報を用いて、分割されたデータ集合毎に、データ範囲を設定し、その結果を範囲情報として出力する。具体的には、図８に示される。 <Normal operation range setting means (FIG. 8)>
In this processing, a data range is set for each divided data set using the above-described division information calculated by the data division unit 100, and the result is output as range information. Specifically, it is shown in FIG.
　・各クラスタに属するデータの各次元の最小値を各クラスタに対応する範囲の各次元の下限とする。 ・ The minimum value of each dimension of the data belonging to each cluster is set as the lower limit of each dimension in the range corresponding to each cluster.
　・各クラスタに属するデータの各次元の最大値を各クラスタに対応する範囲の各次元の上限とする。 ・ The maximum value of each dimension of data belonging to each cluster is set as the upper limit of each dimension in the range corresponding to each cluster.
　ここでいう範囲情報とは、各クラスタ（中心ベクトル）に対応する範囲を規定する各次元の下限値と上限値を指す。 The range information here refers to the lower limit value and the upper limit value of each dimension that define the range corresponding to each cluster (center vector).
　このようにして、端末共通の正常動作範囲学習部５（共通正常動作範囲学習部）は、機械学習により、複数の端末１の過去の動作状態から正常動作範囲（第１の正常動作範囲）を設定する。なお、機械学習に代えて統計処理を用いてもよい。これにより、端末共通の正常動作範囲の信頼性が向上する。 In this way, the normal operation range learning unit 5 (common normal operation range learning unit) common to the terminals obtains the normal operation range (first normal operation range) from the past operation states of the plurality of terminals 1 by machine learning. Set. Note that statistical processing may be used instead of machine learning. Thereby, the reliability of the normal operation range common to the terminals is improved.
　中心ベクトルの座標と範囲情報は、サーバー４から各端末１に送信される。 The center vector coordinates and range information are transmitted from the server 4 to each terminal 1.
　図９は、図７で示した分割結果を基に、本設定手段により範囲を設定した結果例である。 FIG. 9 shows an example of the result of setting the range by the setting means based on the division result shown in FIG.
　＜異常検知手段（図１０＞
　制御の動作異常を検知する。具体的には、図１０に示される。 <Abnormality detection means (FIG. 10)
Detects abnormal control operation. Specifically, it is shown in FIG.
　・検知対象が２次元の場合を示している。 ・ This shows the case where the detection target is two-dimensional.
　・範囲１～４で示される領域が正常動作範囲である。 · The area indicated by the range 1 to 4 is the normal operating range.
　・検知対象のパラメータ値（ベクトル）に対して、Ｌ２距離の意味で、もっとも近い中心ベクトル（図中の○）を特定する。 ・ For the parameter value (vector) to be detected, specify the closest center vector (O in the figure) in the sense of L2 distance.
　・特定した中心ベクトルに対応する領域の内部に、検知対象である制御パラメータ値（ベクトル）が存在していれば、正常と判定し、存在していなければ，異常と判定する。例えば、図１０のベクトルＡは、正常と判定され、ベクトルＢは、異常と判定される。 ・ If a control parameter value (vector) to be detected exists in the area corresponding to the specified center vector, it is determined to be normal, and if not, it is determined to be abnormal. For example, vector A in FIG. 10 is determined to be normal, and vector B is determined to be abnormal.
　換言すれば、異常検知手段２（異常検知部）は、制御手段３（制御部）の動作状態が端末共通の正常動作範囲（第１の正常動作範囲）にない場合、制御手段３が異常であると判定する。これにより、端末共通の正常動作範囲を考慮した異常検知を行うことができる。 In other words, the abnormality detection unit 2 (abnormality detection unit) indicates that the control unit 3 is abnormal when the operation state of the control unit 3 (control unit) is not in the normal operation range common to the terminals (first normal operation range). Judge that there is. Thereby, the abnormality detection which considered the normal operation | movement range common to a terminal can be performed.
　なお、検知対象のパラメータは、制御手段３の出力に相当する操作量でも良い。また、制御手段３の内部で演算される内部パラメータでも良い。また、制御手段３で用いる制御対象に設置されているセンサ出力でも良い。図１０では、２次元の場合を示しているが、Ｎ次元(Ｎ：自然数)まで拡張可能である。 The parameter to be detected may be an operation amount corresponding to the output of the control means 3. Also, an internal parameter calculated inside the control means 3 may be used. Moreover, the sensor output installed in the control object used with the control means 3 may be sufficient. FIG. 10 shows a two-dimensional case, but it can be expanded to N dimensions (N: natural number).
　以上、本実施形態で示した構成によれば、生産ラインの複数のロボットをそれぞれ制御する端末共通の正常範囲を用いて、各端末の制御に関するパラメータの正常/異常を判定するので、システム全体の信頼性が向上する。すなわち、制御の信頼性を向上することができる。 As described above, according to the configuration shown in this embodiment, the normality / abnormality of the parameters related to the control of each terminal is determined using the normal range common to the terminals that respectively control the plurality of robots in the production line. Reliability is improved. That is, the reliability of control can be improved.
　（実施形態２）
　本実施形態においては、各々が機器を制御し、サーバーと通信を行う複数の端末を備えた制御システムにおいて、前記複数の端末共通の第１の正常動作範囲を学習する正常動作範囲学習部と、前記複数の端末中の一の端末の動作状態が前記第１の正常動作範囲にないとき、前記一の端末の動作が異常と判断する異常判断部とを備えた形態について示す。 (Embodiment 2)
In the present embodiment, in a control system including a plurality of terminals each controlling a device and communicating with a server, a normal operation range learning unit that learns a first normal operation range common to the plurality of terminals; An embodiment including an abnormality determination unit that determines that the operation of one terminal among the plurality of terminals is abnormal when the operation state of the one terminal is not in the first normal operation range.
　特に本実施形態では、制御システムは、複数の自動運転車を制御する装置である。 Particularly in this embodiment, the control system is a device that controls a plurality of autonomous vehicles.
　図１は、制御システムの全体を表した図であるが、実施形態１と同じであるので、詳述しない。 FIG. 1 is a diagram showing the entire control system, but since it is the same as that of the first embodiment, it will not be described in detail.
　図１１は、端末１と端末１によって制御される自動運転車２０２を示している。制御手段３により、自動運転車２０２を制御するための操作量（例えば、目標速度、目標回転速度など）が演算される。なお、自動運転車２０２を制御する制御手段３の詳細仕様については、多くの公知技術があるので、ここでは詳述しない。 FIG. 11 shows the terminal 1 and the automatic driving vehicle 202 controlled by the terminal 1. The control means 3 calculates an operation amount (for example, a target speed, a target rotation speed, etc.) for controlling the autonomous driving vehicle 202. The detailed specifications of the control means 3 for controlling the autonomous vehicle 202 are not described here because there are many known techniques.
　図３は、端末１のシステム構成図であるが、実施形態１と同じであるので、詳述しない。 FIG. 3 is a system configuration diagram of the terminal 1, but since it is the same as that of the first embodiment, it will not be described in detail.
　図４は、サーバー４のシステム構成図であるが、実施形態１と同じであるので、詳述しない。以下、各処理の詳細を説明する。 FIG. 4 is a system configuration diagram of the server 4, which is the same as that of the first embodiment and will not be described in detail. Details of each process will be described below.
　＜端末共通の正常動作範囲学習部（図５）＞
　図５は端末共通の正常動作範囲学習部５の全体を表した図であるが、実施形態１と同じであるので、詳述しない。 <Normal operation range learning unit common to terminals (FIG. 5)>
FIG. 5 is a diagram showing the entire normal operation range learning unit 5 common to terminals, but since it is the same as that of the first embodiment, it will not be described in detail.
　＜データ分割手段（図６）＞
　本処理では、制御パラメータ値、センサ出力値のデータを用いて、分割情報を演算する。具体的には、図６に示されるが、実施形態１と同じであるので、詳述しない。 <Data division means (FIG. 6)>
In this process, the division information is calculated using the data of the control parameter value and the sensor output value. Specifically, although it is shown in FIG. 6, it is the same as that of the first embodiment, and therefore will not be described in detail.
　＜正常動作範囲設定手段（図８）＞
　本処理では、データ分割手段で演算した上述の分割情報を用いて、分割されたデータ集合毎に、データ範囲を設定し、その結果を範囲情報として出力する。具体的には、図８に示されるが、実施形態１と同じであるので、詳述しない。 <Normal operation range setting means (FIG. 8)>
In this process, a data range is set for each divided data set using the above-described division information calculated by the data division means, and the result is output as range information. Specifically, although shown in FIG. 8, it is the same as that of the first embodiment, and thus will not be described in detail.
　＜異常検知手段（図１０＞
　制御の動作異常を検知する。具体的には、図１０に示される、実施形態１と同じであるので、詳述しない。 <Abnormality detection means (FIG. 10)
Detects abnormal control operation. Specifically, since it is the same as the first embodiment shown in FIG. 10, it will not be described in detail.
　以上、本実施形態で示した構成によれば、自動運転車をそれぞれ制御する端末共通の正常範囲を用いて、各端末の制御に関するパラメータの正常/異常を判定するので、システム全体の信頼性が向上する。 As described above, according to the configuration shown in the present embodiment, the normality / abnormality of the parameters related to the control of each terminal is determined using the normal range common to the terminals that respectively control the autonomous driving vehicle. improves.
　（実施形態３）
　本実施形態においては、各々が機器を制御し、サーバーと通信を行う複数の端末を備えた制御システムにおいて、前記複数の端末共通の第１の正常動作範囲を学習する正常動作範囲学習部と、前記複数の端末中の一の端末の動作状態が前記第１の正常動作範囲にないとき、前記一の端末の動作が異常と判断する異常判断部とを備えた形態について示す。 (Embodiment 3)
In the present embodiment, in a control system including a plurality of terminals each controlling a device and communicating with a server, a normal operation range learning unit that learns a first normal operation range common to the plurality of terminals; An embodiment including an abnormality determination unit that determines that the operation of one terminal among the plurality of terminals is abnormal when the operation state of the one terminal is not in the first normal operation range.
　特に本実施形態では、制御システムは、複数のドローンを制御する装置である。 Particularly in this embodiment, the control system is a device that controls a plurality of drones.
　図１は、制御システムの全体を表した図であるが、実施形態１と同じであるので、詳述しない。 FIG. 1 is a diagram showing the entire control system, but since it is the same as that of the first embodiment, it will not be described in detail.
　図１２は、端末１と端末１によって制御されるドローン２０３を示している。制御手段３により、ドローン２０３を制御するための操作量（例えば、各ローターの目標回転速度など）が演算される。なお、ドローン２０３を制御する制御手段３の詳細仕様については、多くの公知技術があるので、ここでは詳述しない。 FIG. 12 shows the terminal 1 and the drone 203 controlled by the terminal 1. An operation amount (for example, a target rotation speed of each rotor) for controlling the drone 203 is calculated by the control means 3. The detailed specification of the control means 3 for controlling the drone 203 is not described in detail here because there are many known techniques.
　図３は、端末１のシステム構成図であるが、実施形態１と同じであるので、詳述しない。 FIG. 3 is a system configuration diagram of the terminal 1, but since it is the same as that of the first embodiment, it will not be described in detail.
　図４は、サーバー４のシステム構成図であるが、実施形態１と同じであるので、詳述しない。以下、各処理の詳細を説明する。 FIG. 4 is a system configuration diagram of the server 4, which is the same as that of the first embodiment and will not be described in detail. Details of each process will be described below.
　＜端末共通の正常動作範囲学習部（図５）＞
　図５は端末共通の正常動作範囲学習部５の全体を表した図であるが、実施形態１と同じであるので、詳述しない。 <Normal operation range learning unit common to terminals (FIG. 5)>
FIG. 5 is a diagram showing the entire normal operation range learning unit 5 common to terminals, but since it is the same as that of the first embodiment, it will not be described in detail.
　＜データ分割手段（図６）＞
　本処理では、制御パラメータ値、センサ出力値のデータを用いて、分割情報を演算する。具体的には、図６に示されるが、実施形態１と同じであるので、詳述しない。 <Data division means (FIG. 6)>
In this process, the division information is calculated using the data of the control parameter value and the sensor output value. Specifically, although it is shown in FIG. 6, it is the same as that of the first embodiment, and therefore will not be described in detail.
　＜正常動作範囲設定手段（図８）＞
　本処理では、データ分割手段で演算した上述の分割情報を用いて、分割されたデータ集合毎に、データ範囲を設定し、その結果を範囲情報として出力する。具体的には、図８に示される、実施形態１と同じであるので、詳述しない。 <Normal operation range setting means (FIG. 8)>
In this process, a data range is set for each divided data set using the above-described division information calculated by the data division means, and the result is output as range information. Specifically, since it is the same as the first embodiment shown in FIG. 8, it will not be described in detail.
　＜異常検知手段（図１０＞
　制御の動作異常を検知する。具体的には、図１０に示される、実施形態１と同じであるので、詳述しない。 <Abnormality detection means (FIG. 10)
Detects abnormal control operation. Specifically, since it is the same as the first embodiment shown in FIG. 10, it will not be described in detail.
　以上、本実施形態で示した構成によれば、ドローンをそれぞれ制御する端末共通の正常範囲を用いて、各端末の制御に関するパラメータの正常/異常を判定するので、システム全体の信頼性が向上する。 As described above, according to the configuration shown in this embodiment, the normality / abnormality of the parameters related to the control of each terminal is determined using the normal range common to the terminals that respectively control the drone, so that the reliability of the entire system is improved. .
　（実施形態４）
　本実施形態においては、各々が機器を制御し、サーバーと通信を行う複数の端末を備えた制御システムにおいて、前記複数の端末共通の第１の正常動作範囲を学習する正常動作範囲学習部と、前記複数の端末中の一の端末の動作状態が前記第１の正常動作範囲にないとき、前記一の端末の動作が異常と判断する異常判断部とを備えた形態について示す。 (Embodiment 4)
In the present embodiment, in a control system including a plurality of terminals each controlling a device and communicating with a server, a normal operation range learning unit that learns a first normal operation range common to the plurality of terminals; An embodiment including an abnormality determination unit that determines that the operation of one terminal among the plurality of terminals is abnormal when the operation state of the one terminal is not in the first normal operation range.
　特に本実施形態では、個体差、環境、ユーザー特性、経時変化等に起因する端末固有の第２の正常動作範囲を学習する正常動作範囲学習手段も備える。 In particular, the present embodiment also includes normal operation range learning means for learning a second normal operation range unique to the terminal due to individual differences, environment, user characteristics, changes with time, and the like.
　また、各端末の制御に関するパラメータ値もしくは各端末に備えられているセンサ情報に基づいて、第２の正常動作範囲を学習する。 Also, the second normal operation range is learned based on the parameter value relating to the control of each terminal or the sensor information provided in each terminal.
　また、第２の正常範囲学習は、各端末で実施する。 Also, the second normal range learning is performed at each terminal.
　また、端末の動作状態が、少なくとも前記第１の正常動作範囲もしくは前記第２の正常動作範囲にないとき、当該端末の動作異常と判断する。 Also, when the operating state of the terminal is not at least in the first normal operating range or the second normal operating range, it is determined that the terminal is operating abnormally.
　図１３は、制御システムの全体を表した図である。複数の端末１は、サーバー４を介して、情報の送受信を行う。サーバー４では、端末共通の正常動作範囲学習部５を備える。端末１は、端末固有の正常動作範囲学習部６と異常検知手段７と制御手段３とを備える。 FIG. 13 is a diagram showing the entire control system. The plurality of terminals 1 transmit and receive information via the server 4. The server 4 includes a normal operation range learning unit 5 common to terminals. The terminal 1 includes a normal operation range learning unit 6, an abnormality detection unit 7, and a control unit 3 unique to the terminal.
　なお、換言すれば、端末固有の正常動作範囲学習部６（固有正常動作範囲学習部）は、制御手段３（制御部）に固有の正常動作範囲（第２の正常動作範囲）を学習する。詳細には、端末固有の正常動作範囲学習部６は、個体差、環境、ユーザー特性、経時変化のうち少なくとも１つに起因する正常動作範囲（第２の正常動作範囲）を学習する。なお、端末固有の正常動作範囲学習部６の詳細については、図１５を用いて後述する。 In other words, the normal operation range learning unit 6 (specific normal operation range learning unit) unique to the terminal learns the normal operation range (second normal operation range) unique to the control means 3 (control unit). Specifically, the terminal-specific normal operation range learning unit 6 learns a normal operation range (second normal operation range) caused by at least one of individual differences, environment, user characteristics, and changes with time. Details of the normal operation range learning unit 6 unique to the terminal will be described later with reference to FIG.
　異常検知手段７は、制御パラメータ値、センサ出力などが端末共通の正常動作範囲もしくは端末固有の正常動作範囲にあるか否かで異常検知を行い、異常と判定されたときは、異常フラグをオン（＝１）にする。 The abnormality detection means 7 performs abnormality detection based on whether the control parameter value, the sensor output, etc. are within the normal operation range common to the terminals or the normal operation range specific to the terminals, and when it is determined as abnormal, the abnormality flag is turned on. (= 1).
　図１４は、端末１によって制御される生産ラインのロボット２０１を示している。制御手段３により、ロボット２０１を制御するための操作量（例えば、目標角度、目標速度、目標トルクなど）が演算される。なお、ロボット２０１を制御する制御手段３の詳細仕様については、多くの公知技術があるので、ここでは詳述しない。 FIG. 14 shows a robot 201 on the production line controlled by the terminal 1. An operation amount (for example, a target angle, a target speed, a target torque, etc.) for controlling the robot 201 is calculated by the control means 3. The detailed specification of the control means 3 for controlling the robot 201 has many known techniques and will not be described in detail here.
　図３は、端末１のシステム構成図であるが、実施形態１と同じであるので、詳述しない。 FIG. 3 is a system configuration diagram of the terminal 1, but since it is the same as that of the first embodiment, it will not be described in detail.
　図４は、サーバー４のシステム構成図であるが、実施形態１と同じであるので、詳述しない。以下、各処理の詳細を説明する。 FIG. 4 is a system configuration diagram of the server 4, which is the same as that of the first embodiment and will not be described in detail. Details of each process will be described below.
　＜端末共通の正常動作範囲学習部（図５）＞
　図５は端末共通の正常動作範囲学習部５の全体を表した図であるが、実施形態１と同じであるので、詳述しない。 <Normal operation range learning unit common to terminals (FIG. 5)>
FIG. 5 is a diagram showing the entire normal operation range learning unit 5 common to terminals, but since it is the same as that of the first embodiment, it will not be described in detail.
　＜データ分割手段（図６）＞
　本処理では、制御パラメータ値、センサ出力値のデータを用いて、分割情報を演算する。具体的には、図６に示されるが、実施形態１と同じであるので、詳述しない。 <Data division means (FIG. 6)>
In this process, the division information is calculated using the data of the control parameter value and the sensor output value. Specifically, although it is shown in FIG. 6, it is the same as that of the first embodiment, and therefore will not be described in detail.
　＜正常動作範囲設定手段（図８）＞
　本処理では、データ分割手段１００で演算した上述の分割情報を用いて、分割されたデータ集合毎に、データ範囲を設定し、その結果を範囲情報として出力する。具体的には、図８に示されるが、実施形態１と同じであるので、詳述しない。 <Normal operation range setting means (FIG. 8)>
In this processing, a data range is set for each divided data set using the above-described division information calculated by the data division unit 100, and the result is output as range information. Specifically, although shown in FIG. 8, it is the same as that of the first embodiment, and thus will not be described in detail.
　＜端末固有の正常動作範囲学習部（図１５）＞
　図１５は端末固有の正常動作範囲学習部６の全体を表した図であり、以下の演算部から構成される。 <Terminal-specific normal operating range learning unit (FIG. 15)>
FIG. 15 is a diagram showing the entire normal operation range learning unit 6 unique to the terminal, and includes the following calculation units.
　・データ分割手段１０２
　・正常動作範囲設定手段１１２
　データ分割手段１０２は、制御パラメータ値、センサ出力値の分布を所定の塊毎(クラスタ毎)に分割する。正常動作範囲設定手段１１２は、分割した各クラスタを所定の範囲で規定する。ここで用いる制御パラメータ値、センサ出力値は、端末固有の正常動作範囲を規定するものとして選ぶ。例えば、システム（ソフト）が稼働している時の実績データを用いることが考えられる。 Data dividing means 102
Normal operation range setting means 112
The data dividing unit 102 divides the distribution of control parameter values and sensor output values into predetermined chunks (clusters). The normal operation range setting unit 112 defines each divided cluster within a predetermined range. The control parameter value and the sensor output value used here are selected to define a normal operation range unique to the terminal. For example, it is conceivable to use performance data when the system (software) is operating.
　＜データ分割手段（図１６）＞
　本処理では、制御パラメータ値、センサ出力値のデータを用いて、分割情報を演算する。具体的には、図１６に示される。 <Data division means (FIG. 16)>
In this process, the division information is calculated using the data of the control parameter value and the sensor output value. Specifically, it is shown in FIG.
　ベクトル化されたデータ群を機械学習k-meansを用いてクラスタリングする。k-means法により、得られた分割情報を出力する。ここでいう分割情報とは、以下の情報を指す。 Clustering vectorized data group using machine learning k-means. The obtained division information is output by the k-means method. The division information here refers to the following information.
　・k-means法によって分割されたデータが属するクラスタ番号
　・各クラスタに属するデータの平均値（中心ベクトル）
　なお、k-means法の詳細については、多くの文献、書籍で述べてあるので、ここでは詳述しない。 ・ Cluster number to which data divided by k-means method belongs ・ Average value of data belonging to each cluster (center vector)
The details of the k-means method are described in many literatures and books, and therefore will not be described in detail here.
　＜正常動作範囲設定手段（図１７）＞
　本処理では、データ分割手段１０２で演算した上述の分割情報を用いて、分割されたデータ集合毎に、データ範囲を設定し、その結果を範囲情報として出力する。具体的には、図１７に示される。 <Normal operation range setting means (FIG. 17)>
In this processing, a data range is set for each divided data set using the above-described division information calculated by the data division unit 102, and the result is output as range information. Specifically, it is shown in FIG.
　・各クラスタに属するデータの各次元の最小値を各クラスタに対応する範囲の各次元の下限とする。 ・ The minimum value of each dimension of the data belonging to each cluster is set as the lower limit of each dimension in the range corresponding to each cluster.
　・各クラスタに属するデータの各次元の最大値を各クラスタに対応する範囲の各次元の上限とする。 ・ The maximum value of each dimension of data belonging to each cluster is set as the upper limit of each dimension in the range corresponding to each cluster.
　ここでいう範囲情報とは、各クラスタ（中心ベクトル）に対応する範囲を規定する各次元の下限値と上限値を指す。 The range information here refers to the lower limit value and the upper limit value of each dimension that define the range corresponding to each cluster (center vector).
　このようにして、端末固有の正常動作範囲学習部６（固有正常動作範囲学習部）は、機械学習により、それぞれの端末１の過去の動作状態から正常動作範囲（第２の正常動作範囲）を設定する。なお、機械学習に代えて統計処理を用いてもよい。 In this way, the normal operation range learning unit 6 (inherent normal operation range learning unit) unique to the terminal obtains the normal operation range (second normal operation range) from the past operation state of each terminal 1 by machine learning. Set. Note that statistical processing may be used instead of machine learning.
　中心ベクトルの座標と範囲情報は、正常動作範囲設定手段１１２から異常検知手段７に送られる。 The coordinates and range information of the center vector are sent from the normal operation range setting means 112 to the abnormality detection means 7.
　＜異常検知手段（図１８＞
　制御の動作異常を検知する。具体的には、図１８に示される。 <Abnormality detection means (FIG. 18)
Detects abnormal control operation. Specifically, it is shown in FIG.
　（Ｉ）異常フラグaの演算（端末共通の正常範囲に基づく異常検知）
　・検知対象が２次元の場合を示している。 (I) Calculation of abnormality flag a (abnormality detection based on normal range common to terminals)
-The case where the detection target is two-dimensional is shown.
　・範囲１ａ～４ａで示される領域が端末共通の正常動作範囲である。 · The area indicated by the ranges 1a to 4a is the normal operating range common to terminals.
　・検知対象のパラメータ値（ベクトル）に対して、Ｌ２距離の意味で、もっとも近い中心ベクトル（図中の○）を特定する。 ・ For the parameter value (vector) to be detected, specify the closest center vector (O in the figure) in the sense of L2 distance.
　・特定した中心ベクトルに対応する領域の内部に、検知対象である制御パラメータ値（ベクトル）が存在していれば、正常と判定し（異常フラグaの値を０）、存在していなければ，異常と判定する（異常フラグaの値を１）。 -If the control parameter value (vector) to be detected exists within the area corresponding to the specified center vector, it is determined to be normal (the value of the abnormal flag a is 0), and if it does not exist, It is determined that there is an abnormality (the value of the abnormality flag a is 1).
　（ＩＩ）異常フラグｂの演算（端末固有の正常範囲に基づく異常検知）
　・検知対象が２次元の場合を示している。 (II) Calculation of abnormality flag b (abnormality detection based on normal range unique to terminal)
-The case where the detection target is two-dimensional is shown.
　・範囲１ｂ～４ｂで示される領域が端末固有の正常動作範囲である。 · The area indicated by the ranges 1b to 4b is the normal operating range unique to the terminal.
　・検知対象のパラメータ値（ベクトル）に対して、Ｌ２距離の意味で、もっとも近い中心ベクトル（図中の○）を特定する。 ・ For the parameter value (vector) to be detected, specify the closest center vector (O in the figure) in the sense of L2 distance.
　・特定した中心ベクトルに対応する領域の内部に、検知対象である制御パラメータ値（ベクトル）が存在していれば、正常と判定し（異常フラグｂの値を０）、存在していなければ，異常と判定する（異常フラグｂの値を１）。 If the control parameter value (vector) to be detected exists within the area corresponding to the specified center vector, it is determined as normal (the value of the abnormality flag b is 0), and if it does not exist, It is determined that there is an abnormality (the value of the abnormality flag b is 1).
　（ＩＩＩ）異常フラグの演算
　異常フラグａの値が１もしくは異常フラグｂの値が１のとき、異常フラグの値を１とする。 (III) Calculation of abnormality flag When the value of the abnormality flag a is 1 or the value of the abnormality flag b is 1, the value of the abnormality flag is set to 1.
　換言すれば、異常検知手段７（異常検知部）は、制御手段３（制御部）の動作状態が端末共通の正常動作範囲（第１の正常動作範囲）又は端末固有の正常動作範囲（第２の正常動作範囲）にない場合、制御手段３が異常であると判定する。これにより、端末共通の正常動作範囲及び端末固有の正常動作範囲を考慮した異常検知を行うことができる。 In other words, the abnormality detection unit 7 (abnormality detection unit) is configured such that the operation state of the control unit 3 (control unit) is a normal operation range common to the terminals (first normal operation range) or a normal operation range unique to the terminals (second In the normal operation range), it is determined that the control means 3 is abnormal. Thereby, the abnormality detection which considered the normal operation range common to a terminal and the normal operation range specific to a terminal can be performed.
　なお、検知対象のパラメータは、制御手段３の出力に相当する操作量でも良い。また、制御手段３の内部で演算される内部パラメータでも良い。また、制御手段３で用いる制御対象に設置されているセンサ出力でも良い。図１８では、２次元の場合を示しているが、Ｎ次元(Ｎ：自然数)まで拡張可能である。 The parameter to be detected may be an operation amount corresponding to the output of the control means 3. Also, an internal parameter calculated inside the control means 3 may be used. Moreover, the sensor output installed in the control object used with the control means 3 may be sufficient. FIG. 18 shows a two-dimensional case, but it can be expanded to N dimensions (N: natural number).
　以上、本実施形態で示した構成によれば、生産ラインの複数のロボットをそれぞれ制御する各端末共通の異常範囲と、各端末の個体差、環境差、ユーザー特性差、経時変化等に起因する固有の異常範囲があることを考慮した異常検知を行うので、複数の端末がある制御システムの制御性能と信頼性の双方が向上する。 As described above, according to the configuration shown in the present embodiment, it is caused by an abnormal range common to each terminal that controls a plurality of robots on the production line, individual differences of each terminal, environmental differences, user characteristic differences, temporal changes, and the like. Since abnormality detection is performed considering that there is a unique abnormality range, both control performance and reliability of a control system having a plurality of terminals are improved.
　（実施形態５）
　本実施形態においては、各々が機器を制御し、サーバーと通信を行う複数の端末を備えた制御システムにおいて、前記複数の端末共通の第１の正常動作範囲を学習する正常動作範囲学習部と、前記複数の端末中の一の端末の動作状態が前記第１の正常動作範囲にないとき、前記一の端末の動作が異常と判断する異常判断部とを備えた形態について示す。 (Embodiment 5)
In the present embodiment, in a control system including a plurality of terminals each controlling a device and communicating with a server, a normal operation range learning unit that learns a first normal operation range common to the plurality of terminals; An embodiment including an abnormality determination unit that determines that the operation of one terminal among the plurality of terminals is abnormal when the operation state of the one terminal is not in the first normal operation range.
　また、個体差、環境、ユーザー特性、経時変化等に起因する端末固有の第２の正常動作範囲を学習する正常動作範囲学習手段も備える。 Also provided is a normal operation range learning means for learning a second normal operation range unique to the terminal due to individual differences, environment, user characteristics, changes with time, and the like.
　また、各端末の制御に関するパラメータ値もしくは各端末に備えられているセンサ情報に基づいて、第２の正常動作範囲を学習する。 Also, the second normal operation range is learned based on the parameter value relating to the control of each terminal or the sensor information provided in each terminal.
　また、第２の正常範囲学習は、各端末で実施する。 Also, the second normal range learning is performed at each terminal.
　また、端末の動作状態が、少なくとも第１の正常動作範囲もしくは第２の正常動作範囲にないとき、当該端末の動作異常と判断する
　特に、本実施形態では、第１の正常範囲と第２の正常範囲は、ルールベースで設定する。なお、第１の正常範囲（第１の正常動作範囲）、第２の正常範囲（前記第２の正常動作範囲のうち少なくとも１つが、ルールベースで設定されるようにしてもよい。これにより、経験的に得られたルールに基づいて、異常検知を行うことができる。 Further, when the operating state of the terminal is not at least in the first normal operating range or the second normal operating range, it is determined that the terminal is operating abnormally. In particular, in the present embodiment, the first normal range and the second normal operating range are determined. The normal range is set on a rule basis. The first normal range (first normal operation range) and the second normal range (at least one of the second normal operation ranges may be set on a rule basis. Anomaly detection can be performed based on empirically obtained rules.
　図１３は、制御システムの全体を表した図であるが、実施形態４と同じであるので、詳述しない。 FIG. 13 is a diagram showing the entire control system, but since it is the same as that of the fourth embodiment, it will not be described in detail.
　図１４は、端末１によって制御される生産ラインのロボット２０１を示しているが、実施形態４と同じであるので、詳述しない。なお、ロボット２０１を制御する制御手段３の詳細仕様については、多くの公知技術があるので、ここでは詳述しない。 FIG. 14 shows the robot 201 on the production line controlled by the terminal 1, but since it is the same as that of the fourth embodiment, it will not be described in detail. The detailed specification of the control means 3 for controlling the robot 201 has many known techniques and will not be described in detail here.
　図３は、端末１のシステム構成図であるが、実施形態１と同じであるので、詳述しない。 FIG. 3 is a system configuration diagram of the terminal 1, but since it is the same as that of the first embodiment, it will not be described in detail.
　図４は、サーバー４のシステム構成図であるが、実施形態１と同じであるので、詳述しない。以下、各処理の詳細を説明する。 FIG. 4 is a system configuration diagram of the server 4, which is the same as that of the first embodiment and will not be described in detail. Details of each process will be described below.
　＜端末共通の正常動作範囲学習部（図１９）＞
　図１９は端末共通の正常動作範囲学習部５を表した図であり、以下の演算部から構成される。制御パラメータ値、センサ出力値を用いて、ルールベース用しきい値学習にて、下記処理により、ルールベース用しきい値K1aL,K2aL,・・・,KnaL、K1aH,K2aH,・・・,KnaHを演算する。 <Normal operation range learning unit common to terminals (FIG. 19)>
FIG. 19 is a diagram showing a normal operation range learning unit 5 common to terminals, and includes the following calculation units. The rule base threshold values K1aL, K2aL, ..., KnaL, K1aH, K2aH, ..., KnaH by rule base threshold learning using control parameter values and sensor output values Is calculated.
　制御パラメータ１の値の最小値をK1a_L、最大値をK1a_H
　制御パラメータ２の値の最小値をK2a_L、最大値をK2a_H
　　　・・・
　制御パラメータｎの値の最小値をKna_L、最大値をKna_H
とする。 The minimum value of control parameter 1 is K1a_L and the maximum value is K1a_H
The minimum value of control parameter 2 is K2a_L and the maximum value is K2a_H
...
The minimum value of control parameter n is Kna_L and the maximum value is Kna_H
And
　ここで用いる制御パラメータ値、センサ出力値は、端末共通の正常動作範囲を規定するものとして選ぶ。例えば、システム（ソフト）を稼働させる前の事前検証で得られるデータを用いることが考えられる。 制 御 The control parameter values and sensor output values used here are selected to define the normal operating range common to terminals. For example, it is conceivable to use data obtained by prior verification before operating the system (software).
　＜端末固有の正常動作範囲学習部（図２０）＞
　図２０は端末固有の正常動作範囲学習部６を表した図であり、以下の演算部から構成される。制御パラメータ値、センサ出力値を用いて、ルールベース用しきい値学習にて、下記処理により、ルールベース用しきい値K1bL,K2bL,・・・,KnbL、K1bH,K2bH,・・・,KnbHを演算する。 <Terminal-specific normal operating range learning unit (FIG. 20)>
FIG. 20 is a diagram showing the normal operation range learning unit 6 unique to the terminal, and includes the following calculation units. Rule base threshold K1bL, K2bL, ..., KnbL, K1bH, K2bH, ..., KnbH by rule base threshold learning using control parameter values and sensor output values Is calculated.
　制御パラメータ１の値の最小値をK1b_L、最大値をK1b_H
　制御パラメータ２の値の最小値をK2b_L、最大値をK2b_H
　　　・・・
　制御パラメータｎの値の最小値をKnb_L、最大値をKnb_H
とする。 The minimum value of control parameter 1 is K1b_L and the maximum value is K1b_H
The minimum value of control parameter 2 is K2b_L and the maximum value is K2b_H
...
The minimum value of the control parameter n is Knb_L and the maximum value is Knb_H
And
　ここで用いる制御パラメータ値、センサ出力値は、端末固有の正常動作範囲を規定するものとして選ぶ。例えば、システム（ソフト）が稼働している時の実績データを用いることが考えられる。 * The control parameter values and sensor output values used here are selected to define the normal operating range specific to the terminal. For example, it is conceivable to use performance data when the system (software) is operating.
　＜異常検知手段（図２１）＞
　制御の動作異常を検知する。具体的には、図２１に示される。 <Abnormality detection means (FIG. 21)>
Detects abnormal control operation. Specifically, it is shown in FIG.
　（Ｉ）異常フラグaの演算（端末共通の正常範囲に基づく異常検知）
　（Ｉ－Ａ）下記のいずれかが成立したとき、異常フラグａの値を１とする。 (I) Calculation of abnormality flag a (abnormality detection based on normal range common to terminals)
(IA) The value of the abnormality flag a is set to 1 when any of the following is established.
　　・（制御パラメータ１） ＜K1a_L
　　・K1a_H ＜（制御パラメータ１）
　　・（制御パラメータ２） ＜K2a_L
　　・K2a_H ＜（制御パラメータ２）
　　　　　・・・
　　・（制御パラメータｎ） ＜Kna_L
　　・Kna_H ＜（制御パラメータｎ）
　（Ｉ－Ｂ）それ以外のときは、異常フラグａの値を０とする。 ・ (Control parameter 1) <K1a_L
・ K1a_H <(Control parameter 1)
・ (Control parameter 2) <K2a_L
・ K2a_H <(Control parameter 2)
...
・ (Control parameter n) <Kna_L
・ Kna_H <(Control parameter n)
(IB) Otherwise, the value of the abnormality flag a is set to 0.
　（ＩＩ）異常フラグｂの演算（端末固有の正常範囲に基づく異常検知）
　（ＩＩ－Ａ）下記のいずれかが成立したとき、異常フラグｂの値を１とする。 (II) Calculation of abnormality flag b (abnormality detection based on normal range unique to terminal)
(II-A) The value of the abnormality flag b is set to 1 when any of the following is established.
　　・（制御パラメータ１） ＜K1b_L
　　・K1b_H ＜（制御パラメータ１）
　　・（制御パラメータ２） ＜K2b_L
　　・K2b_H ＜（制御パラメータ２）
　　　　　・・・
　　・（制御パラメータｎ） ＜Knb_L
　　・Knb_H ＜（制御パラメータｎ）
　（ＩＩ－Ｂ）それ以外のときは、異常フラグｂの値を０とする。 ・ (Control parameter 1) <K1b_L
・ K1b_H <(Control parameter 1)
・ (Control parameter 2) <K2b_L
・ K2b_H <(Control parameter 2)
...
・ (Control parameter n) <Knb_L
・ Knb_H <(control parameter n)
(II-B) In other cases, the value of the abnormality flag b is set to 0.
　（ＩＩＩ）異常フラグの演算
　異常フラグａの値が１もしくは異常フラグｂの値が１のとき、異常フラグの値を１とする。 (III) Calculation of abnormality flag When the value of the abnormality flag a is 1 or the value of the abnormality flag b is 1, the value of the abnormality flag is set to 1.
　以上、本実施形態で示した構成によれば、生産ラインの複数のロボットをそれぞれ制御する各端末共通の異常範囲と、各端末の個体差、環境差、ユーザー特性差、経時変化等に起因する固有の異常範囲があることを考慮した異常検知を行うので、複数の端末がある制御システムの制御性能と信頼性の双方が向上する。また、本実施形態では異常検知をルールベースで行うので、異常検知方式が明示的に与えられる異常検知時の説明性も向上する。 As described above, according to the configuration shown in the present embodiment, it is caused by an abnormal range common to each terminal that controls a plurality of robots on the production line, individual differences of each terminal, environmental differences, user characteristic differences, temporal changes, and the like. Since abnormality detection is performed considering that there is a unique abnormality range, both control performance and reliability of a control system having a plurality of terminals are improved. In addition, since the abnormality detection is performed on a rule basis in the present embodiment, the explanation at the time of abnormality detection in which the abnormality detection method is explicitly given is also improved.
　（変形例１）
　本変形１では、図２２に示す端末１の制御手段３は、異常検知手段２（異常検知部）の出力である異常フラグに応じて、所定の制御を行う。 (Modification 1)
In the first modification, the control means 3 of the terminal 1 shown in FIG. 22 performs predetermined control according to an abnormality flag that is an output of the abnormality detection means 2 (abnormality detection unit).
　例えば、制御手段３は、異常フラグがオン（＝１）の場合、表示装置等に警告を出力したり、所定のフェールセーフ制御を行うようにしてもよい。これにより、異常が検知された場合に、適切な制御を行うことができる。 For example, when the abnormality flag is on (= 1), the control means 3 may output a warning to the display device or the like or perform predetermined fail-safe control. Thereby, when abnormality is detected, appropriate control can be performed.
　（変形例２）
　本変形例２では、端末１の制御手段３は、異常検知手段２（異常検知部）の判定結果に応じて、所定の制御を行う。 (Modification 2)
In the second modification, the control unit 3 of the terminal 1 performs predetermined control according to the determination result of the abnormality detection unit 2 (abnormality detection unit).
　図２２は、端末１の動作状態が第１の正常動作範囲にあるか否かと端末１の動作状態が第２の正常動作範囲にあるか否かの組合せと、異常検知手段２（異常検知部）の判定結果と、所定の制御とを対応付けて記憶するテーブル３００を示す図である。なお、テーブル３００は、例えば、端末１の記憶装置１１に記憶されるが、サーバー４の記憶装置２１に記憶されてもよい。 FIG. 22 shows a combination of whether or not the operation state of the terminal 1 is in the first normal operation range and whether or not the operation state of the terminal 1 is in the second normal operation range, and the abnormality detection means 2 (abnormality detection unit). It is a figure which shows the table 300 which matches and memorize | stores the determination result of () and predetermined | prescribed control. The table 300 is stored in the storage device 11 of the terminal 1, for example, but may be stored in the storage device 21 of the server 4.
　異常検知手段２は、端末１の動作状態が第１の正常動作範囲内にあり且つ第２の正常動作範囲内にある場合、制御手段３が正常であると判定する。 The abnormality detection means 2 determines that the control means 3 is normal when the operation state of the terminal 1 is within the first normal operation range and within the second normal operation range.
　異常検知手段２は、端末１の動作状態が第１の正常動作範囲外にあり且つ第２の正常動作範囲内にある場合、制御手段３が異常（異常１）であると判定する。 The abnormality detection means 2 determines that the control means 3 is abnormal (abnormal 1) when the operation state of the terminal 1 is outside the first normal operation range and within the second normal operation range.
　異常検知手段２は、端末１の動作状態が第１の正常動作範囲内にあり且つ第２の正常動作範囲外にある場合、制御手段３が異常（異常２）であると判定する。 The abnormality detection unit 2 determines that the control unit 3 is abnormal (abnormal 2) when the operation state of the terminal 1 is within the first normal operation range and out of the second normal operation range.
　異常検知手段２は、端末１の動作状態が第１の正常動作範囲外にあり且つ第２の正常動作範囲外にある場合、制御手段３が異常（異常３）であると判定する。 The abnormality detecting means 2 determines that the control means 3 is abnormal (abnormal 3) when the operating state of the terminal 1 is outside the first normal operating range and outside the second normal operating range.
　制御手段３（制御部）は、テーブル３００から異常検知手段２（異常検知部）の判定結果（正常、異常１～３）に対応する所定の制御（通常制御、制御１～３）の識別子（ＩＤ）を読み出して、この識別子に対応する制御を実行する。 The control unit 3 (control unit) identifies an identifier (normal control, control 1 to 3) identifier corresponding to the determination result (normal, abnormality 1 to 3) of the abnormality detection unit 2 (abnormality detection unit) from the table 300. ID) is read out, and control corresponding to this identifier is executed.
　例えば、制御１では、表示装置等に警告を出力し、制御２では、フェールセーフ制御を行い、制御３では、制御を停止するようにしてもよい。これにより、異常が検知された場合に、異常に応じた適切な制御を行うことができる。なお、制御１～制御３を同じ制御（例えば、警告の出力）としてもよい。 For example, in control 1, a warning may be output to a display device or the like, in control 2 fail-safe control may be performed, and in control 3 control may be stopped. Thereby, when abnormality is detected, appropriate control according to abnormality can be performed. Control 1 to control 3 may be the same control (for example, output of warning).
　なお、本発明は上記した実施形態に限定されるものではなく、様々な変形例が含まれる。例えば、上述した実施形態は本発明を分かりやすく説明するために詳細に説明したものであり、必ずしも説明した全ての構成を備えるものに限定されるものではない。また、ある実施形態の構成の一部を他の実施形態の構成に置き換えることが可能であり、また、ある実施形態の構成に他の実施形態の構成を加えることも可能である。また、各実施形態の構成の一部について、他の構成の追加・削除・置換をすることが可能である。 Note that the present invention is not limited to the above-described embodiment, and includes various modifications. For example, the above-described embodiment has been described in detail for easy understanding of the present invention, and is not necessarily limited to the one having all the configurations described. Further, a part of the configuration of an embodiment can be replaced with the configuration of another embodiment, and the configuration of another embodiment can be added to the configuration of an embodiment. In addition, it is possible to add, delete, and replace other configurations for a part of the configuration of each embodiment.
　また、上記の各構成、機能（手段）等は、それらの一部又は全部を、例えば集積回路で設計する等によりハードウェアで実現してもよい。また、上記の各構成、機能（手段）等は、プロセッサ（CPU）がそれぞれの機能（手段）等を実現するプログラムを解釈し、実行することによりソフトウェアで実現してもよい。各機能（手段）等を実現するプログラム、テーブル、ファイル等の情報は、メモリや、ハードディスク、SSD（Solid State Drive）等の記録装置、または、ICカード、SDカード、DVD等の記録媒体に置くことができる。 Further, each of the above-described configurations, functions (means), etc. may be realized by hardware by designing a part or all of them, for example, by an integrated circuit. Each of the above-described configurations, functions (means), and the like may be realized by software by interpreting and executing a program that realizes each function (means) or the like by a processor (CPU). Information such as programs, tables, and files that realize each function (means) is stored in a memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC card, SD card, or DVD. be able to.
　なお、本発明の実施形態は、以下の態様であってもよい。 In addition, the following aspects may be sufficient as embodiment of this invention.
　（１）各々が機器を制御し、サーバーと通信を行う複数の端末を備えた制御システムにおいて、前記制御システムは、前記複数の端末共通の第１の正常動作範囲を学習する正常動作範囲学習部と、前記複数の端末中の一の端末の動作状態が前記第１の正常動作範囲にないとき、前記一の端末の動作が異常と判断する異常判断部とを備える制御システム。 (1) In a control system including a plurality of terminals each controlling a device and communicating with a server, the control system learns a first normal operation range common to the plurality of terminals. And an abnormality determination unit that determines that the operation of the one terminal is abnormal when an operation state of one of the plurality of terminals is not in the first normal operation range.
　（２）（１）において、個体差、環境、ユーザー特性、経時変化等に起因する端末固有の第２の正常動作範囲を学習する正常動作範囲学習手段を備えたことを特徴とする制御システム。 (2) A control system according to (1), comprising a normal operation range learning means for learning a second normal operation range unique to a terminal caused by individual differences, environment, user characteristics, changes with time, and the like.
　（３）（１）において、個体差、環境、ユーザー特性、経時変化等に起因する端末固有の第２の正常動作範囲を学習する正常動作範囲学習手段を備え、前記第２の正常動作範囲は、少なくとも各端末の制御に関するパラメータ値もしくは各端末に備えられているセンサからの情報もしくは端末での通信による情報に基づいて、学習することを特徴とする制御システム。 (3) In (1), there is provided a normal operation range learning means for learning a second normal operation range unique to the terminal due to individual differences, environment, user characteristics, changes with time, etc., and the second normal operation range is A control system that learns based on at least a parameter value related to control of each terminal, information from a sensor provided in each terminal, or information obtained by communication at the terminal.
　（４）（１）において、前記端末の動作状態が、少なくとも前記第１の正常動作範囲もしくは前記第２の正常動作範囲にないとき、当該端末の動作異常と判断することを特徴とする制御システム。 (4) In (1), when the operating state of the terminal is not at least in the first normal operating range or the second normal operating range, it is determined that the terminal is operating abnormally. .
　（５）（１）において、前記第１の正常動作範囲の学習は、前記サーバーで行うことを特徴とする制御システム。 (5) The control system according to (1), wherein the learning of the first normal operation range is performed by the server.
　（６）（１）において、前記第２の正常動作範囲の学習は、前記各端末で行うことを特徴とする制御システム。 (6) The control system according to (1), wherein the learning of the second normal operation range is performed by each of the terminals.
　（７）（１）において、前記異常判断部は、各端末で実施することを特徴とする制御システム。 (7) The control system according to (1), wherein the abnormality determination unit is implemented in each terminal.
　（８）（１）において、少なくとも前記第１の正常範囲もしくは前記第２の正常範囲は、前記端末の過去の動作状態を、統計処理もしくは機械学習による処理の結果に基づいて設定することを特徴とする制御システム。 (8) In (1), at least the first normal range or the second normal range sets a past operation state of the terminal based on a result of processing by statistical processing or machine learning. And control system.
　（９）（１）において、少なくとも前記第１の正常範囲もしくは前記第２の正常範囲は、ルールベースで設定することを特徴とする制御システム。 (9) The control system according to (9), wherein at least the first normal range or the second normal range is set on a rule basis.
　（１０）（１）において、前記制御システムは、ロボットを制御する装置であることを特徴とする。 (10) In (1), the control system is a device for controlling a robot.
　（１１）（１）において、前記制御システムは、自動運転車を制御する装置であることを特徴とする。 (11) In (1), the control system is a device for controlling an autonomous driving vehicle.
　（１２）（１）において、前記制御システムは、ドローンなど飛行体を制御する装置であることを特徴とする。 (12) In (1), the control system is a device for controlling a flying object such as a drone.
　上記（１）～（１２）によれば、各端末共通の異常範囲と、各端末の個体差、環境差、ユーザー特性差、経時変化等に起因する固有の異常範囲があることを考慮した異常検知を行うので、複数の端末がある制御システムの信頼性を高めることが可能となる。 According to the above (1) to (12), an abnormality taking into account that there is an abnormal range common to each terminal and a specific abnormal range due to individual differences, environmental differences, user characteristic differences, changes with time, etc. Since detection is performed, the reliability of a control system having a plurality of terminals can be improved.
１…端末
２…異常検知手段
３…制御手段
４…サーバー
５…端末共通の正常動作範囲学習部
６…端末固有の正常動作範囲学習部
７…異常検知手段
１１…端末の記憶装置
１２…端末のＣＰＵ
１３…端末のＲＯＭ
１４…端末のＲＡＭ
１５…端末のデータバス
１６…端末の入力回路
１７…端末の入出力ポート
１８…端末の出力回路
２１…サーバーの記憶装置
２２…サーバーのＣＰＵ
２３…サーバーのＲＯＭ
２４…サーバーのＲＡＭ
２５…サーバーのデータバス
２６…サーバーの入力回路
２７…サーバーの入出力ポート
２８…サーバーの出力回路
５１…異常検知手段の処理（端末共通の正常動作範囲に基づく）
５２…異常検知手段の処理（端末固有の正常動作範囲に基づく）
６１…異常検知手段の処理（端末共通の正常動作範囲に基づく）
６２…異常検知手段の処理（端末固有の正常動作範囲に基づく）
１００…データ分割手段（端末共通の正常動作範囲学習部）
１０１…データ分割手段の処理（端末共通の正常動作範囲学習部）
１０２…データ分割手段（端末固有の正常動作範囲学習部）
１０３…データ分割手段の処理（端末固有の正常動作範囲学習部）
１１０…正常動作範囲設定手段（端末共通の正常動作範囲学習部）
１１１…正常動作範囲設定手段の処理（端末共通の正常動作範囲学習部）
１１２…正常動作範囲設定手段（端末固有の正常動作範囲学習部）
１１３…正常動作範囲設定手段の処理（端末固有の正常動作範囲学習部）
１２１…ルールベース用しきい値学習（端末共通の正常動作範囲学習部）
１２２…ルールベース用しきい値学習（端末固有の正常動作範囲学習部）
２０１…ロボット
２０２…自動運転車
２０３…ドローン
３００…テーブル DESCRIPTION OF SYMBOLS 1 ... Terminal 2 ... Abnormality detection means 3 ... Control means 4 ... Server 5 ... Normal operation range learning part 6 common to terminals ... Normal operation range learning part 7 specific to a terminal ... Abnormality detection means 11 ... Storage device 12 of a terminal ... CPU
13 ... Terminal ROM
14 ... Terminal RAM
DESCRIPTION OF SYMBOLS 15 ... Terminal data bus 16 ... Terminal input circuit 17 ... Terminal input / output port 18 ... Terminal output circuit 21 ... Server storage device 22 ... Server CPU
23 ... Server ROM
24 ... Server RAM
25 ... Server data bus 26 ... Server input circuit 27 ... Server input / output port 28 ... Server output circuit 51 ... Processing of abnormality detection means (based on normal operating range common to terminals)
52. Processing of abnormality detection means (based on normal operation range unique to the terminal)
61 ... Processing of abnormality detection means (based on normal operating range common to terminals)
62 ... Processing of abnormality detection means (based on normal operation range unique to terminal)
100: Data division means (normal operation range learning unit common to terminals)
101: Processing of data dividing means (normal operation range learning unit common to terminals)
102: Data dividing means (terminal-specific normal operating range learning unit)
103 ... Processing of data dividing means (normal operation range learning unit unique to terminal)
110: Normal operation range setting means (normal operation range learning unit common to terminals)
111: Processing of normal operation range setting means (normal operation range learning unit common to terminals)
112 ... Normal operation range setting means (terminal-specific normal operation range learning unit)
113: Processing of normal operation range setting means (terminal-specific normal operation range learning unit)
121 ... Rule-based threshold learning (normal operation range learning unit common to terminals)
122 ... Rule-based threshold learning (terminal-specific normal operation range learning unit)
201 ... Robot 202 ... Automatic driving car 203 ... Drone 300 ... Table

Claims (11)

  1. 　複数の端末に共通の第１の正常動作範囲を受信する受信部と、
    　機械の制御を行う制御部と、
    　前記制御部の動作状態が前記第１の正常動作範囲にない場合、前記制御部が異常であると判定する異常検知部と、
    　を備えることを特徴とする制御装置。     A receiving unit that receives a first normal operating range common to a plurality of terminals;
    A control unit for controlling the machine;
    An abnormality detection unit that determines that the control unit is abnormal when the operation state of the control unit is not in the first normal operation range;
    A control device comprising:
  2. 　請求項１に記載の制御装置であって、
    　前記制御部に固有の第２の正常動作範囲を学習する固有正常動作範囲学習部を備え、
    　前記異常検知部は、
    　前記制御部の動作状態が前記第１の正常動作範囲又は前記第２の正常動作範囲にない場合、前記制御部が異常であると判定する
    　ことを特徴とする制御装置。     The control device according to claim 1,
    A normal normal operation range learning unit that learns a second normal operation range specific to the control unit;
    The abnormality detection unit
    When the operation state of the control unit is not in the first normal operation range or the second normal operation range, it is determined that the control unit is abnormal.
  3. 　請求項２に記載の制御装置であって、
    　前記固有正常動作範囲学習部は、
    　個体差、環境、ユーザー特性、経時変化のうち少なくとも１つに起因する前記第２の正常動作範囲を学習する
    　ことを特徴とする制御装置。     The control device according to claim 2,
    The inherent normal operation range learning unit
    A control device that learns the second normal operation range caused by at least one of individual differences, environment, user characteristics, and changes over time.
  4. 　請求項２に記載の制御装置と同じ構成の複数の端末とサーバーを含む制御システムであって、
    　前記サーバーは、
    　複数の端末に共通の前記第１の正常動作範囲を学習する共通正常動作範囲学習部を備える
    　ことを特徴とする制御システム。     A control system including a plurality of terminals and a server having the same configuration as the control device according to claim 2,
    The server
    A control system, comprising: a common normal operation range learning unit that learns the first normal operation range common to a plurality of terminals.
  5. 　請求項４に記載の制御システムであって、
    　前記サーバーの前記共通正常動作範囲学習部は、
    　統計処理又は機械学習により、複数の前記端末の過去の動作状態から前記第１の正常動作範囲を設定する
    　ことを特徴とする制御システム。     The control system according to claim 4,
    The common normal operation range learning unit of the server is
    The control system is characterized in that the first normal operation range is set from past operation states of the plurality of terminals by statistical processing or machine learning.
  6. 　請求項４に記載の制御システムであって、
    　それぞれの前記端末の前記固有正常動作範囲学習部は、
    　統計処理又は機械学習により、それぞれの前記端末の過去の動作状態から前記第２の正常動作範囲を設定する
    　ことを特徴とする制御システム。     The control system according to claim 4,
    The inherent normal operation range learning unit of each terminal is
    The control system is characterized in that the second normal operation range is set from past operation states of the terminals by statistical processing or machine learning.
  7. 　請求項４に記載の制御システムであって、
    　前記第１の正常動作範囲、前記第２の正常動作範囲のうち少なくとも１つは、
    　ルールベースで設定される
    　ことを特徴とする制御システム。     The control system according to claim 4,
    At least one of the first normal operating range and the second normal operating range is:
    A control system characterized by being set on a rule basis.
  8. 　請求項４に記載の制御システムであって、
    　それぞれの前記端末の前記動作状態は、
    　それぞれの前記端末の前記制御部が行う制御の入力値、出力値、前記入力値から前記出力値を決定する関数のパラメータを示す制御パラメータ値のうち少なくとも１つによって示される
    　ことを特徴とする制御システム。     The control system according to claim 4,
    The operating state of each of the terminals is
    Control represented by at least one of an input value, an output value, and a control parameter value indicating a parameter of a function for determining the output value from the input value, performed by the control unit of each terminal. system.
  9. 　請求項４に記載の制御システムであって、
    　前記異常検知部の判定結果に対応する所定の制御を対応付けて記憶するテーブルを有し、
    　それぞれの前記端末の前記制御部は、
    　それぞれの前記端末の前記異常検知部の判定結果に対応する前記所定の制御を行う
    　ことを特徴とする制御システム。     The control system according to claim 4,
    A table for storing predetermined control corresponding to the determination result of the abnormality detection unit,
    The control unit of each terminal is
    The control system characterized by performing the predetermined control corresponding to the determination result of the abnormality detection unit of each of the terminals.
  10. 　請求項４に記載の制御システムであって、
    　前記機械は、
    　ロボット、自動運転車、又は飛行体である
    　ことを特徴とする制御システム。     The control system according to claim 4,
    The machine is
    A control system characterized by being a robot, an autonomous vehicle, or a flying object.
  11. 　機械の制御を行う制御部を有する複数の端末からそれぞれの前記端末の動作状態を受信する受信部と、
    　複数の前記端末に共通の第１の正常動作範囲を学習する共通正常動作範囲学習部と、
    　前記第１の正常動作範囲を複数の前記端末に送信する送信部と、
    　を備えることを特徴とするサーバー。     A receiving unit for receiving the operating state of each of the terminals from a plurality of terminals having a control unit for controlling the machine;
    A common normal operation range learning unit that learns a first normal operation range common to the plurality of terminals;
    A transmitter that transmits the first normal operation range to the plurality of terminals;
    A server characterized by comprising:
PCT/JP2017/009572 2017-03-09 2017-03-09 Control device, control system, and server WO2018163375A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/JP2017/009572 WO2018163375A1 (en) 2017-03-09 2017-03-09 Control device, control system, and server

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/JP2017/009572 WO2018163375A1 (en) 2017-03-09 2017-03-09 Control device, control system, and server

Publications (1)

Publication Number Publication Date
WO2018163375A1 true WO2018163375A1 (en) 2018-09-13

Family

ID=63449032

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/JP2017/009572 WO2018163375A1 (en) 2017-03-09 2017-03-09 Control device, control system, and server

Country Status (1)

Country Link
WO (1) WO2018163375A1 (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2020046964A (en) * 2018-09-19 2020-03-26 ファナック株式会社 Characteristic determination device, characteristic determination method, and characteristic determination program
JP2021149315A (en) * 2020-03-17 2021-09-27 富士電機株式会社 Information processing device, information processing method, and information processing program
JP7500980B2 (en) 2020-02-04 2024-06-18 富士電機株式会社 Information processing device, information processing method, and information processing program

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH02173809A (en) * 1988-12-27 1990-07-05 Toshiba Corp Trouble diagnosing device
JP2014186631A (en) * 2013-03-25 2014-10-02 Hitachi Constr Mach Co Ltd Diagnosis processing system, terminal equipment and server
WO2016067852A1 (en) * 2014-10-29 2016-05-06 株式会社日立製作所 Diagnostic job generating system, diagnostic job generating method, and diagnostic job generating display method

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH02173809A (en) * 1988-12-27 1990-07-05 Toshiba Corp Trouble diagnosing device
JP2014186631A (en) * 2013-03-25 2014-10-02 Hitachi Constr Mach Co Ltd Diagnosis processing system, terminal equipment and server
WO2016067852A1 (en) * 2014-10-29 2016-05-06 株式会社日立製作所 Diagnostic job generating system, diagnostic job generating method, and diagnostic job generating display method

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2020046964A (en) * 2018-09-19 2020-03-26 ファナック株式会社 Characteristic determination device, characteristic determination method, and characteristic determination program
JP7500980B2 (en) 2020-02-04 2024-06-18 富士電機株式会社 Information processing device, information processing method, and information processing program
JP2021149315A (en) * 2020-03-17 2021-09-27 富士電機株式会社 Information processing device, information processing method, and information processing program
JP7484261B2 (en) 2020-03-17 2024-05-16 富士電機株式会社 Information processing device, information processing method, and information processing program

Similar Documents

Publication Publication Date Title
WO2018179191A1 (en) Control device and control system
US11190533B2 (en) Anomaly detection electronic control unit, onboard network system, and anomaly detection method
US11438350B2 (en) Unauthorized communication detection method, unauthorized communication detection system, and non-transitory computer-readable recording medium storing a program
US11423145B2 (en) Methods and arrangements for multi-layer in-vehicle network intrusion detection and characterization
US11591070B2 (en) Methods and apparatus for a distributed aircraft actuation system
US9008808B2 (en) Control system for safely operating at least one functional component
WO2018163375A1 (en) Control device, control system, and server
US20220368708A1 (en) Unauthorized communication detection method, unauthorized communication detection system, and non-transitory computer-readable recording medium storing a program
EP3742295A1 (en) Automatic firmware rollback
US9840124B2 (en) Electronic control system for a motor vehicle
JP2019008618A (en) Information processing apparatus, information processing method, and program
CN110455549B (en) Diagnostic device, diagnostic system, and diagnostic method
US20190035174A1 (en) Vehicle calibration based upon performance product detection
KR101601074B1 (en) ECU Updating Apparatus, ECU Updating Method and ECU Updating Network using the same
EP3972217A1 (en) Ml-based voltage fingerprinting for ground truth and controlled message error for message and ecu mapping for can bus
EP3806428B1 (en) Transformation device, transformation method and storage medium
WO2021095491A1 (en) In-vehicle relay device and information processing method
WO2021127700A1 (en) Reduced number of states in a tire pressure monitoring system
US20230202246A1 (en) Persistent alarm transmission associated with a vehicle mounted wireless sensor device
US11682217B2 (en) Surveillance device, learning device, surveillance method and storage medium
US20240086541A1 (en) Integrity verification device and integrity verification method
US10936397B2 (en) Hybrid control module status communication system and method
US12007836B2 (en) Parameterization process and system for a vehicle
US20240187448A1 (en) Method for Detecting a Manipulation of a Message of a Bus System of a Vehicle
US11750424B2 (en) Electronic control unit and determination method

Legal Events

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

Ref document number: 17899857

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17899857

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: JP