WO2019207767A1

WO2019207767A1 - Control device and control method

Info

Publication number: WO2019207767A1
Application number: PCT/JP2018/017208
Authority: WO
Inventors: 中川　慎二
Original assignee: 株式会社日立製作所
Priority date: 2018-04-27
Filing date: 2018-04-27
Publication date: 2019-10-31

Abstract

The present invention can suitably normalize an abnormality of a control device by correcting an abnormal parameter on the basis of a past normal parameter, even when an accidental abnormality occurs in the parameter. The present invention is configured to have: a parameter division unit 2 which divides a distribution area of parameter values, which are calculated to control a vehicle 5, into a plurality of division ranges 1-4; a transition frequency calculation unit 3 which calculates a transition frequency between division ranges from and to which the parameter value is transitioned among the divided division ranges 1-4; and a parameter correction unit 4 which, when a parameter value A1 calculated to control the vehicle 5 is included in any of the division ranges 1-4 and a parameter A2, which is calculated after the parameter A1, is not included in any of the division ranges 1-4, corrects, on the basis of the transition frequency, the parameter value A2 to a value within the division range 4 to which a transition frequency from the division range 1, in which the parameter value A1 is included, is the highest.

Description

Control apparatus and control method

The present invention relates to a control device and a control method.

In Patent Literature 1, the health status data of a series of subjects given in the form of a bug of word is clustered as a collection of individual data for each subject and age, and the results of the clustering are as follows. A first clustering unit that calculates transition probabilities between the clusters, and network clustering on the clusters corresponding to the transition probabilities calculated by the first clustering unit, thereby obtaining each cluster as a community and the result of this clustering A second clustering unit that calculates a transition probability between each cluster of the first clustering unit, the cluster and the transition probability output from the first clustering unit, and the cluster and the transition probability output from the second clustering unit, respectively. Prediction model construction device that outputs as a model It has been disclosed.

JP 2016-173728 A

However, in this type of control device, when a sudden abnormality occurs in the parameter output from the control device, there is a problem that the subsequent processing is performed based on the abnormal parameter. What is disclosed in Patent Document 1 does not disclose any countermeasures or countermeasures when a sudden abnormality occurs in the parameter output by the control device.

Therefore, the present invention can properly normalize the abnormal state of the control device by correcting the abnormal parameter based on the past normal parameter even when a sudden abnormality occurs in the parameter output by the control device. An object is to provide a control device.

In order to solve the above problem, a parameter dividing unit that divides a distribution area of parameter values calculated for controlling the controlled device into a plurality of divided ranges, and a plurality of divided ranges divided by the parameter dividing unit, The transition frequency calculation unit that calculates the transition frequency between the divided ranges in which the values have changed, and the first parameter value calculated to control the controlled device are included in any of the divided ranges divided by the parameter dividing unit. The transition calculated by the transition frequency calculation unit when the second parameter value calculated after the first parameter value is not included in any of the division ranges divided by the parameter division unit Based on the frequency, the second parameter value is corrected to a value in the second divided range having the highest transition frequency from the first divided range including the first parameter value. A parameter correction unit that was configured to have.

According to the present invention, even when a sudden abnormality occurs in the parameter output by the control device, the abnormal state of the control device is properly normalized by correcting the abnormal parameter based on the past normal parameter. can do.

It is a block diagram explaining the functional structure of the control apparatus concerning this invention. It is a block diagram explaining the hardware constitutions of a control apparatus. It is a block diagram at the time of applying a control apparatus to an autonomous driving vehicle. It is a block diagram explaining the function structure of a parameter division part. It is a block diagram explaining the function of a 1st division | segmentation process part. It is a figure explaining an example of the result of clustering the parameter value. It is a block diagram explaining the function of a 2nd division | segmentation process part. It is a figure explaining an example of the setting of the division range of parameter value distribution. It is a block diagram explaining the function structure of a transition frequency calculation part. It is a graph showing an example of the result of calculating the transition frequency between division ranges. It is a block diagram explaining the functional structure of a parameter correction | amendment part. It is a block diagram explaining the functional structure of the parameter correction | amendment part concerning 2nd Embodiment. It is a block diagram explaining the functional structure of the parameter correction part concerning 3rd Embodiment. It is a block diagram explaining the functional structure of the control apparatus concerning 4th Embodiment. It is a block diagram explaining the functional structure of the control apparatus concerning 5th Embodiment.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
[First Embodiment]
First, the control device 1 according to the first embodiment of the present invention will be described.
FIG. 1 is a block diagram illustrating a functional configuration of the control device 1 according to the embodiment.

<Functional configuration of control device>
As shown in FIG. 1, the control device 1 includes a parameter dividing unit 2, a transition frequency calculating unit 3, and a parameter correcting unit 4.

The parameter dividing unit 2 divides the distribution range V (see FIG. 6) of the past parameter value A for controlling the controlled device (for example, an autonomous driving vehicle) into a predetermined range, which is a divided range 1 to 4 (FIG. 8). Is calculated and stored in the storage device 11 (see FIG. 2).

The transition frequency calculation unit 3 calculates the transition frequency between the divided ranges based on the divided ranges 1 to 4 calculated by the parameter dividing unit 2 and the parameter value A related to control, and stores the calculated frequency in the storage device 11 (see FIG. 2). To do.

The parameter correction unit 4 corrects the parameter value Ax after correction based on the division ranges 1 to 4 calculated by the parameter division unit 2, the transition frequency calculated by the transition frequency calculation unit 3, and the parameter value A related to control. Is calculated. A method for calculating the corrected parameter value Ax by the parameter correction unit 4 will be described later.

<Hardware configuration of control device>
Each function of the parameter dividing unit 2, the transition frequency calculating unit 3, and the parameter correcting unit 4 of the control device 1 described above is exhibited when the CPU 12 described later executes a control program stored in the ROM 13. Next, the hardware configuration of the control device 1 will be described.
FIG. 2 is a block diagram illustrating a hardware configuration of the control device 1.

The control device 1 includes a storage device 11, a CPU (Central Processing Unit) 12, a ROM (Read Only Memory) 13, a RAM (Random Access Memory) 14, an input circuit 16, an input / output port 17, and an output circuit. 18. In each of these devices, information is transmitted / received via the data bus 15.

The input circuit 16 receives signals from various sensors provided in the controlled device (for example, the automatic driving vehicle 5 shown in FIG. 3). For example, when the control target of the control device 1 is the autonomous driving vehicle 5, the driving conditions of the autonomous driving vehicle 5 and the surrounding environmental conditions are detected by various sensors, and signals detected by the various sensors are input to the input circuit 16. Input and processed. The input signal after being processed by the input circuit 16 is transmitted to the input / output port 17. The input signal transmitted to the input / output port 17 is stored in the RAM 14 or the storage device 11 via the data bus 15.

Note that the processing and functions described above are performed by the CPU 12 executing a control program stored in the ROM 13, but the control program may be stored in the storage device 11. Even in this way, the CPU 12 executes the control program stored in the storage device 11 so that the processing and functions described above are performed. At that time, the CPU 12 appropriately reads out the value stored in the RAM 14 or the storage device 11 and performs the calculation.

As a result of the calculation by the CPU 12, information (value) transmitted to the outside of the control device 1 is transmitted to the input / output port 17 via the data bus 15, and then the information transmitted to the input / output port 17 is output as an output signal. It is transmitted to the circuit 18. The output signal transmitted to the output circuit 18 is transmitted to the outside as an external signal. Here, for example, the signal to the outside indicates an actuator signal or the like for causing the control target to perform a desired movement.

<Example of automatic driving vehicle control by control device>
Next, the case where the control device 1 is applied to the control device of the autonomous driving vehicle 5 will be described as an example.
FIG. 3 is a block diagram when the control device 1 is applied to the autonomous driving vehicle 5.
FIG. 4 is a block diagram illustrating a functional configuration of the parameter dividing unit 2.
FIG. 5 is a block diagram illustrating functions of the first division processing unit 21.
FIG. 6 is a diagram for explaining an example of the result of clustering the parameter values, where the horizontal axis indicates the engine speed of the automatic driving vehicle 5 and the vertical axis indicates the engine torque of the automatic driving vehicle 5.
FIG. 7 is a block diagram illustrating functions of the second division processing unit 22.
FIG. 8 is a diagram for explaining an example of setting the division range of the distribution of parameter values, where the horizontal axis indicates the engine speed of the automatic driving vehicle 5 and the vertical axis indicates the engine torque of the automatic driving vehicle 5.

As shown in FIG. 3, the control device 1 (parameter correction unit 4) is connected to an ECU (Electronic Control Unit: not shown) of the autonomous driving vehicle 5 and the like, and the autonomous driving vehicle 5 is used as the corrected parameter value Ax. The amount of operation for controlling is calculated. Here, the operation amount for controlling the autonomous driving vehicle 5 is, for example, a target speed, a target rotational speed of the engine, or the like. In addition, the control apparatus 1 may be provided integrally with ECU (not shown) of the autonomous driving vehicle 5, and may be performed by CPU (not shown) of ECU.

As shown in FIG. 4, the parameter dividing unit 2 of the control device 1 described above is a first division processing unit that obtains the result of dividing the parameter value A from the distribution of the parameter value A related to the control of the autonomous driving vehicle 5 that is the controlled device. 21 and a second division processing unit 22 for determining the division ranges 1 to 4 from the division result of the parameter value A by the first division processing unit. Here, as an example of the parameter value A related to the control when the controlled device is the autonomous driving vehicle 5, a case where the value is a value indicating the relationship between the engine speed and the engine torque will be described.

As shown in FIG. 5, the first division processing unit 21 clusters the parameter values A (vectors) related to the control of the autonomous driving vehicle 5 that is the controlled device using machine learning (for example, k-means method). Then, the first division processing unit 21 outputs the division result of the parameter value A obtained by the k-means method to the second division processing unit 22. Here, the analysis result of the parameter value A indicates the cluster number (see FIG. 6) to which the parameter value after division by the k-means method belongs. The division information may be an average value (center vector) of parameter values belonging to each cluster.

As shown in FIG. 6, the first division processing unit 21 performs a clustering process on the distribution area V of the parameter value A (two-dimensional vector) related to the control of the autonomous driving vehicle 5 with the number of clusters of 4 by the k-means method. In the embodiment, the first division processing unit 21 uses a distribution region V of the parameter value A related to vehicle control as a cluster 1 with an engine speed of about 2000 rpm (Revolutions Per Minute) or less, a cluster 2 with about 2000-3500 rpm, Clustering is performed with cluster number 4 of cluster 3 of about 3500 to 5000 rpm and cluster 4 of about 5000 rpm or more.

Next, referring back to FIG. 4, the second division processing unit 22 uses the division result of the parameter value A related to the control calculated by the first division processing unit 21 to divide the divided range for each distribution of the parameter value A divided. And outputs the result to the transition frequency calculation unit 3 as a divided range. Specifically, as shown in FIG. 7, the second division processing unit 22 performs the following processes (1) and (2).
(1) The minimum value of each dimension of the parameter value A belonging to each of the clusters 1 to 4 divided by the first division processing unit 21 is set as the lower limit value of each dimension of the range corresponding to each cluster.
(2) The maximum value of each dimension of the parameter value A belonging to each of the clusters 1 to 4 is set as the upper limit value of each dimension in the range corresponding to each cluster.
As shown in FIG. 8, the engine speed is about 2000 rpm (Revolutions Per) based on the division result of the distribution of the parameter value A related to control by the processes (1) and (2) by the second division processing unit 22 described above. (Minute) The cluster 1 below is set to the division range 1, the cluster 2 of about 2000 to 3500 rpm is set to the division range 2, the cluster 3 of about 3500 to 5000 rpm is set to the division range 3, and the cluster 4 of about 5000 rpm or more is set. Is set to the division range 4.

Here, the above-described division range of the distribution region V of the parameter value A indicates a range defined by the lower limit value and the upper limit value of each dimension of the division range corresponding to each divided cluster. In this embodiment, the division range is a range defined by the lower limit value and the upper limit value of each dimension of the range corresponding to each cluster. However, a machine using a supervised learning model such as SVM (Support Vector Machine). The division range of the parameter value A may be set based on the learning result by learning.

As described above, the information on the division ranges 1 to 4 set by the parameter division unit 2 and the parameter value A related to the control of the autonomous driving vehicle 5 are output to the transition frequency calculation unit 3, and the transition frequency calculation unit 3 The transition frequency between the divided ranges is calculated from the information of the divided ranges 1 to 4 set by the dividing unit 2 and the parameter value A related to the control, and stored in the storage device 11.

FIG. 9 is a block diagram illustrating a functional configuration of the transition frequency calculation unit 3. As shown in FIG. 9, when the parameter value related to the control of the autonomous driving vehicle 5 is updated from the parameter value K to the parameter value K + 1, the transition frequency calculation unit 3 determines the division range N corresponding to the cluster to which the parameter value K + 1 belongs. The division range N + 1 corresponding to the cluster to which the parameter value K + 1 belongs is obtained, and the movement occurrence frequency from the division range N to the division range N + 1 is calculated. For example, the occurrence frequency when the parameter value has moved from division range 1 to division range 1 (no movement), the occurrence frequency when movement from division range 1 to division range 2, the occurrence frequency when movement from division range 1 to division range 3, The occurrence frequency of movement from the range 1 to the division range 4, the occurrence frequency of movement from the division range 2 to the division frequency 1, and the occurrence frequency of movement (no movement) from the division range 4 to the division range 4 are calculated. For the calculation of the transition frequency by the transition frequency calculation unit 3, a method that is easy to calculate, such as dividing the number of times of movement from the divided range N to the divided range N + 1 by the number of parameter value updates, can be selected as appropriate. Further, the transition frequency calculation unit 3 calculates the movement pattern of the parameter values K and K + 1 and the transition frequency thereof using the information of the cluster to which the parameter value K and the parameter value K + 1 belong instead of the divided range N and the divided range N + 1. You may do it.

FIG. 10 is a graph showing an example of the result of calculating the transition frequency between the divided ranges. As shown in FIG. 10, the transition frequencies between the divided ranges described above are calculated, and the respective transition frequencies are shown in a graph. In the embodiment, the transition frequency moving from the divided range 1 to the divided range 4 is high. It is shown. The calculation result of the transition frequency between the divided ranges calculated by the transition frequency calculation unit 3 and the parameter value A related to the control of the autonomous driving vehicle 5 are output to the parameter correction unit 4.

FIG. 11 is a block diagram illustrating a functional configuration of the parameter correction unit 4.
As shown in FIG. 11, the parameter correction unit 4 includes the division ranges 1 to 4 of the distribution region V of the parameter value A obtained by the parameter division unit 2, the transition frequency between the division ranges obtained by the transition frequency calculation unit 3, and The corrected parameter value Ax is calculated from the control parameter value A. Specifically, the parameter correction unit 4 sets the newly calculated control parameter value A1 (first parameter value) to the divided range 1 to the distribution range V of the parameter value A calculated by the parameter dividing unit 2. If the parameter value A2 (second parameter value) calculated next is not included in any of the divided ranges 1 to 4, it is calculated by the transition frequency calculation unit 3. In the divided range 4 (second divided range) having the highest transition frequency from the divided range 1 (first divided range) including the parameter value A1 with reference to the transition frequency (for example, the transition frequency shown in FIG. 10). The parameter value A2 is corrected to the boundary point A2a of the division range 4 that is closest to the parameter value A2.

In the above-described embodiment, the control device 1 determines the range of the distribution region V of the past parameter value A related to the control of the autonomous driving vehicle 5 based on the parameter value A during operation of the autonomous driving vehicle 5 that is the controlled device. It is divided into a plurality of divided ranges 1 to 4, and based on the parameter value A, the frequency of transition of the divided ranges divided by the parameter value A is calculated and stored, and newly calculated when the autonomous driving vehicle 5 is operated. When the parameter value A1 is included in any of the divided ranges 1 to 4 and the parameter value A2 calculated next is not included in any of the divided ranges, the parameter value A1 is based on the transition frequency. The parameter value is corrected to the divided range 4 having the highest transition frequency from the divided range 1 including.
In particular, the control device 1 divides the parameter value A into a plurality of ranges based on the result of clustering. Further, the frequency of occurrence of the pattern in which the parameter value A has changed from the division range to which the parameter value A belongs last time to the division range to which the parameter value A belongs is stored. Further, the parameter A2 is corrected to the boundary point A2a of the division range 4 that is closest to the abnormal parameter value A2.
Therefore, when an abnormality in the control parameter value of the control device 1 is detected, the abnormality parameter value A2 is corrected to a control range in which the previous normal parameter value A1 has the highest probability of transition. By correcting to A2a, it is possible to normalize the abnormality of the control device 1 with a high probability, and it is possible to safely control the autonomous driving vehicle.

As explained above, in the first embodiment,
(1) The parameter dividing unit 2 that divides the distribution region V of the parameter value A calculated for controlling the autonomous driving vehicle 5 (controlled device) into a plurality of divided ranges 1 to 4 and the parameter dividing unit 2 In the plurality of divided ranges 1 to 4, the transition frequency calculation unit 3 that calculates the transition frequency between the divided ranges in which the parameter value A has changed, and the parameter value A1 that is calculated to control the autonomous driving vehicle 5 (first 1 parameter value) is included in any of the division ranges 1 to 4 divided by the parameter division unit 2, and the parameter value A2 (second parameter value) calculated after the parameter value A1 is the parameter In the case where it is not included in any of the division ranges 1 to 4 divided by the dividing unit 2, the first parameter value A1 is included based on the transition frequency calculated by the transition frequency calculating unit 3. Split range (e.g., division areas 1) highest second split range transition frequency from (e.g., split range 4) a parameter correction unit 4 corrects the parameter value A2 to a value within, and configured to have.

With this configuration, when the control device 1 detects an abnormality in the parameter value A2, the abnormal parameter value A2 is corrected to a control range where the probability that the previous normal parameter value A1 transitions is highest. Therefore, when a sudden abnormality occurs, it is possible to normalize the abnormal state of the control device with a high probability by correcting it to the most plausible value, and the autonomous driving vehicle 5 can be controlled safely. .

(2) Further, the parameter dividing unit 2 is configured to divide the parameter value distribution area calculated for controlling the autonomous driving vehicle 5 into a plurality of divided ranges by clustering.

With this configuration, the parameter dividing unit 2 can appropriately divide the parameter value distribution region by clustering processing using a computer.

(3) Moreover, the transition frequency calculation part 3 represents the transition frequency (occurrence frequency) of the pattern in which the parameter value calculated in order to control the autonomous driving vehicle 5 represents within the division | segmentation range where a parameter value is included. It was set as the structure to calculate (refer FIG. 10).

With this configuration, the transition frequency calculation unit 3 can appropriately calculate the transition frequency of the parameter value as a representative value within the divided range including the parameter value.

(4) Further, the transition frequency calculation unit 3 calculates the occurrence frequency of the pattern in which the parameter value calculated for controlling the autonomous driving vehicle 5 has changed from the divided range 1 belonging to the past to the divided range 4 belonging to this time. It was set as the structure to do.

With this configuration, the transition frequency calculation unit 3 can appropriately calculate the transition frequency based on the previous transition between the previous parameter value and the current parameter value.

(5) In addition, the parameter correction unit 4 converts the second parameter value A2 not included in any of the division ranges 1 to 4 into the first division range (for example, the division range) including the first parameter value A1. The configuration is such that the value is corrected to any value within the second divided range (for example, divided range 4) having the highest transition frequency from range 1).

With this configuration, the parameter correction unit 4 sets the abnormal parameter value A2 that is not included in any of the divided ranges 1 to 4 from the divided range 1 that includes the normal parameter value A1 immediately before the parameter value A2. Since the correction is made to any value in the divided range 4 having a high transition frequency, the abnormal parameter value A2 can be corrected to the most likely normal parameter value.

(6) The parameter correction unit 4 sets the parameter value A2 that is not included in any of the divided ranges 1 to 4 to the transition frequency from the divided range 1 that includes the normal parameter value A1 in the past in the straight line. It was set as the structure corrected to the boundary value of the high division range 4.

With this configuration, the parameter correction unit 4 corrects the parameter value A2a as close as possible to the parameter value A2 because the parameter correction unit 4 corrects the parameter value A2 to the parameter value closest to the parameter A2 in the divided range 4 including the normal parameter A1. It can be corrected. Therefore, since the control device 1 corrects the abnormal parameter value A2 to be close to the abnormal parameter value A2 and to the normal parameter value A2a as compared with the case where the abnormal parameter value A2 is corrected to a completely different value, the automatic operation after the parameter value correction is performed. The vehicle 5 can be controlled smoothly.

<Second Embodiment>
Next, a control device according to a second embodiment of the present invention will be described.
FIG. 12 is a block diagram illustrating a functional configuration of a parameter correction unit 4A according to the second embodiment. In addition, about the same structure and function as the control apparatus 1 concerning 1st Embodiment, the same code | symbol is attached | subjected and it demonstrates as needed.

As shown in FIG. 12, the parameter correction unit 4A of the control device is any one in which the parameter value A1 related to the control of the newly calculated controlled device (for example, the autonomous driving vehicle 5) is divided by the parameter dividing unit 2. When the parameter value A2 relating to the control calculated next is not included in any of the division ranges 1 to 4 and is included in the division ranges 1 to 4, the transition frequency calculated by the transition frequency calculation unit 3 (for example, Referring to (transition frequency shown in FIG. 10), the division range having the highest transition frequency from the division ranges 1 to 4 including the previously calculated parameter value A1 is included in the division range having the highest transition frequency. Of the past parameter values A, the parameter value A2 is corrected to the past actual parameter value A2a closest to the parameter value A2.

In the embodiment, the parameter correction unit 4A includes the newly calculated parameter value A1 in the divided range 1, and the next calculated parameter value A2 is not included in any of the divided ranges 1 to 4. In this case, referring to the transition frequency (transition frequency shown in FIG. 10) calculated by the transition frequency calculation unit 3, in the divided range 4 having the highest transition frequency from the divided range 1 including the parameter A1, this divided range 4 The parameter value A2 is corrected to the past actual parameter value A2a that is closest to the parameter value A2 among the past parameter values A included in.

In the second embodiment described above, the control device 1 distributes the past parameter value A related to the control of the autonomous driving vehicle 5 based on the parameter value A during the operation of the autonomous driving vehicle 5 that is the controlled device. Is divided into a plurality of divided ranges 1 to 4, and based on the parameter value A, the frequency of transition of the plurality of divided ranges divided by the parameter value A is calculated and stored. When the calculated parameter value A1 is included in any of the divided ranges 1 to 4 and the next calculated parameter value A2 is not included in any of the divided ranges, based on the transition frequency The parameter value is corrected to the divided range 4 having the highest transition frequency from the divided range 1 including the parameter value A1.
In particular, the control device 1 divides the parameter value A into a plurality of ranges based on the result of clustering. Further, the frequency of occurrence of the pattern in which the parameter value A has changed from the division range to which the parameter value A belongs last time to the division range to which the parameter value A belongs is stored. Further, the parameter value A2 is corrected to the parameter value A2a closest to the abnormal parameter value A2 among the past parameter values A belonging to the division range 4.
Therefore, when an abnormality in the control parameter value of the control device 1 is detected, the abnormality parameter value A2 is corrected to a control range in which the previous normal parameter value A1 has the highest probability of transition. By correcting to A2a, it is possible to normalize the abnormality of the control device 1 with a high probability, and it is possible to safely control the autonomous driving vehicle.

As explained above, in the second embodiment,
(8) The parameter correction unit 4 converts the parameter value A2 not included in any of the division ranges 1 to 4 into the second in the division range 4 having the highest transition frequency from the division range 1 including the parameter value A1. The parameter value is corrected to the closest past parameter value A2a.

If comprised in this way, since the parameter correction | amendment part 4 correct | amends the abnormal parameter value A2 to the past normal actual parameter value A2a, based on the past corrected parameter value A2a which has the past track record, the automatic driving vehicle 5 Can be controlled safely.

<Third Embodiment>
Next, a control device according to a third embodiment of the present invention will be described.
FIG. 13 is a block diagram illustrating a functional configuration of a parameter correction unit 4B according to the third embodiment. In addition, about the same structure and function as the control apparatus 1 concerning 1st Embodiment, the same code | symbol is attached | subjected and it demonstrates as needed.

As shown in FIG. 13, the parameter correction unit 4B of the control device is any of the parameter values A1 related to the control of the newly calculated controlled device (for example, the autonomous driving vehicle 5) divided by the parameter dividing unit 2. When the parameter value A2 relating to the control calculated next is not included in any of the division ranges 1 to 4 and is included in the division ranges 1 to 4, the transition frequency calculated by the transition frequency calculation unit 3 (for example, Referring to (transition frequency shown in FIG. 10), the division range having the highest transition frequency from the division ranges 1 to 4 including the previously calculated parameter value A1 is included in the division range having the highest transition frequency. The parameter value A2 is corrected to the representative vector of the past parameter values. Here, the representative vector of the parameter value A included in the predetermined divided range is a center vector of the distribution of past actual parameter values included in the predetermined divided range, and the past actual parameter included in the divided range. The average value of the distribution of value A.

Specifically, the parameter correction unit 4B includes the newly calculated parameter value A1 in the divided range 1 and the next calculated parameter value A2 in any of the divided ranges 1 to 4. If there is not, in the divided range 4 having the highest transition frequency from the divided range 1 including the parameter A1 with reference to the transition frequency calculated by the transition frequency calculating unit 3 (the transition frequency shown in FIG. 10), this divided range 4, the parameter value A2 is corrected to the representative vector A2a that is the average value of the distribution of the actual parameter value A in the past.

In the third embodiment described above, the control device 1 distributes the past parameter value A distribution region V related to the control of the autonomous driving vehicle 5 based on the parameter value A during operation of the autonomous driving vehicle 5 that is the controlled device. Is divided into a plurality of divided ranges 1 to 4, and based on the parameter value A, the frequency of transition of the plurality of divided ranges divided by the parameter value A is calculated and stored. When the calculated parameter value A1 is included in any of the divided ranges 1 to 4 and the next calculated parameter value A2 is not included in any of the divided ranges, based on the transition frequency The parameter value is corrected to the divided range 4 having the highest transition frequency from the divided range 1 including the parameter value A1.
In particular, the control device 1 divides the parameter value A into a plurality of ranges based on the result of clustering. Further, the frequency of occurrence of the pattern in which the parameter value A has changed from the division range to which the parameter value A belongs last time to the division range to which the parameter value A belongs is stored. Also, the parameter A2 is corrected to the representative vector (center vector) A2a of the experienced parameter value A belonging to the division range 4.
Therefore, when an abnormality in the control parameter value of the control device 1 is detected, the abnormality parameter value A2 is corrected to a control range in which the previous normal parameter value A1 has the highest probability of transition. By correcting to A2a, it is possible to normalize the abnormality of the control device 1 with a high probability, and it is possible to safely control the autonomous driving vehicle.

As explained above, in the second embodiment,
(9) The parameter correction unit 4 converts the parameter value A2 not included in any of the divided ranges 1 to 4 into the parameter value in the divided range 4 having the highest transition frequency from the divided range 1 including the parameter value A1. It was set as the structure correct | amended to the representative vector (representative value) of A2a.

With this configuration, the parameter correction unit 4 converts the abnormal parameter value A2 that is not included in any of the division ranges 1 to 4 into the division that has the highest transition frequency from the division range 1 that includes the normal parameter value A1. Since it is corrected to the representative value (average value) of the parameter value A2a in the range 4, it can be corrected to a more appropriate normal parameter value A2a.

<Fourth embodiment>
Next, the control apparatus 1 concerning the 4th Embodiment of this invention is demonstrated.
FIG. 14 is a block diagram illustrating a functional configuration of the control device 1 according to the fourth embodiment. The fourth embodiment is different from the above-described embodiment in that the controlled device controlled by the control device 1 is a robot 6. In addition, about the same structure and function as the control apparatus 1 concerning 1st Embodiment, the same code | symbol is attached | subjected and it demonstrates as needed.

As shown in FIG. 14, the control device 1 is connected to the robot 6, and the newly calculated parameter value A1 related to the control of the robot 6 is divided into any one of the divided ranges 1 divided by the parameter dividing unit 2 described above. When the parameter value A2 calculated next is not included in any of the divided ranges 1 to 4, the parameter correction unit 4 makes a transition from the divided range including the parameter value A1. In the division range with the highest frequency, the parameter value A2 is corrected to the boundary point A2a of the division range closest to the parameter value A2. Alternatively, the parameter correction unit 4A sets the parameter value A2a closest to the parameter value A2 among the past parameter values included in the division range in the division range having the highest transition frequency from the division range including the parameter value A1. The parameter value A2 is corrected. Alternatively, the parameter correction unit 4B sets the parameter value at the point A2a of the representative vector (center vector) of the past parameter value included in the divided range in the divided range having the highest transition frequency from the divided range including the parameter value A1. Correct A2

As described above, by applying the control device 1 to the control of the robot 6, even if a sudden abnormality occurs in the control parameter of the robot 6, the control parameter is appropriately corrected to prevent malfunction of the robot 6. And can be controlled safely. Further, it is possible to suitably prevent the production line from being stopped due to a malfunction of the robot 6.

<Fifth embodiment>
Next, the control apparatus 1 concerning the 5th Embodiment of this invention is demonstrated.
FIG. 15 is a block diagram illustrating a functional configuration of the control device 1 according to the fifth embodiment. The fifth embodiment differs from the above-described embodiment in that the controlled device controlled by the control device 1 is a drone 7. In addition, about the same structure and function as the control apparatus 1 concerning 1st Embodiment, the same code | symbol is attached | subjected and it demonstrates as needed.

As shown in FIG. 15, the control device 1 is connected to the drone 7, and the newly calculated parameter value A1 related to the control of the drone 7 is any one of the division ranges 1 divided by the parameter division unit 2 described above. When the parameter value A2 calculated next is not included in any of the divided ranges 1 to 4, the parameter correction unit 4 makes a transition from the divided range including the parameter value A1. In the division range with the highest frequency, the parameter value A2 is corrected to the boundary point A2a of the division range closest to the parameter value A2. Alternatively, the parameter correction unit 4A sets the parameter value A2a closest to the parameter value A2 among the past parameter values included in the division range in the division range having the highest transition frequency from the division range including the parameter value A1. The parameter value A2 is corrected. Alternatively, the parameter correction unit 4B sets the parameter value at the point A2a of the representative vector (center vector) of the past parameter value included in the divided range in the divided range having the highest transition frequency from the divided range including the parameter value A1. Correct A2

As described above, by applying the control device 1 to the control of the drone 7, even if a sudden abnormality occurs in the control parameter of the drone 7, the control parameter can be appropriately corrected, and the drone 7 malfunctions. It is possible to prevent the danger of a fall caused by, and to control it safely.

In the above, an example of the embodiment of the present invention has been described. However, the present invention may combine all of the above-described embodiments, and may arbitrarily combine any two or more embodiments.

Further, the present invention is not limited to the one having all the configurations of the above-described embodiment, and a part of the configuration of the above-described embodiment is replaced with the configuration of another embodiment. In addition, the configuration of the above-described embodiment may be replaced with the configuration of another embodiment.

Further, a part of the configuration of the above-described embodiment may be added to, deleted from, or replaced with the configuration of another embodiment.

1: control device, 11: storage device, 12: CPU, 13: ROM, 14: RAM, 15: data bus, 16: input circuit, 17: input / output port, 18: output circuit, 2: parameter dividing unit, 21 : First division processing unit, 22: second division processing unit, 3: transition frequency calculation unit, 4: parameter correction unit, 5: vehicle, 6: robot, 7: drone

Claims

A parameter dividing unit that divides the distribution region of the parameter values calculated for controlling the controlled device into a plurality of divided ranges;
In a plurality of divided ranges divided by the parameter dividing unit, a transition frequency calculating unit that calculates a transition frequency between divided ranges in which the parameter value has changed,
The first parameter value calculated for controlling the controlled device is included in any of the division ranges divided by the parameter dividing unit, and the first parameter value calculated after the first parameter value is included. When the parameter value of 2 is not included in any of the division ranges divided by the parameter dividing unit, the first parameter value is included based on the transition frequency calculated by the transition frequency calculating unit. And a parameter correction unit that corrects the second parameter value to a value in the second divided range having the highest transition frequency from the first divided range.
The control device according to claim 1, wherein the parameter dividing unit divides a distribution region of parameter values calculated for controlling the controlled device into a plurality of divided ranges by clustering.
The said transition frequency calculation part calculates the occurrence frequency of the pattern in which the parameter value calculated in order to control the said to-be-controlled apparatus represents in the said division | segmentation range in which the said parameter value is included. Control device.
The said transition frequency calculation part calculates the occurrence frequency of the pattern in which the parameter value calculated in order to control the said to-be-controlled device changed to the division range which belongs this time from the division range which belongs to the past. Control device.
The parameter correction unit converts the second parameter value that is not included in any of the divided ranges into a second division that has the highest transition frequency from the first divided range that includes the first parameter value. The control device according to claim 1, wherein the control device corrects any value within the range.
The parameter correction unit converts the second parameter value that is not included in any of the divided ranges into a second division that has the highest transition frequency from the first divided range that includes the first parameter value. The control device according to claim 1, wherein the control device corrects the boundary value of the range.
The parameter correction unit converts the second parameter value that is not included in any of the divided ranges into a second division that has the highest transition frequency from the first divided range that includes the first parameter value. The control device according to claim 1, wherein the control is performed to correct the past parameter value closest to the second parameter value within the range.
The parameter correction unit converts the second parameter value that is not included in any of the divided ranges into a second division that has the highest transition frequency from the first divided range that includes the first parameter value. The control device according to claim 1, wherein the control device corrects the parameter value within a range to a representative value.
The control device according to claim 1, wherein the controlled device is an autonomous vehicle, and the control device controls the autonomous vehicle.
The control device according to claim 1, wherein the controlled device is a robot, and the control device controls the robot.
The control device according to claim 1, wherein the controlled device is a flying object, and the control device controls the flying object.
Dividing the distribution region of the parameter values calculated for controlling the controlled device into a plurality of divided ranges;
Calculating a transition frequency between division ranges in which the parameter value has changed in a plurality of division ranges divided by the parameter division unit;
The first parameter value calculated for controlling the controlled device is included in any of the division ranges divided by the parameter dividing unit, and the first parameter value calculated after the first parameter value is included. When the parameter value of 2 is not included in any of the division ranges divided by the parameter dividing unit, the first parameter value is included based on the transition frequency calculated by the transition frequency calculating unit. And a step of correcting the second parameter value to a value in the second divided range having the highest transition frequency from the first divided range.