Control model generation device and control model generation method

Abstract

This control model generation device is configured to comprise: an observation value acquisition unit (1) that acquires a plurality of observation values that are outputs of a controlled object having nonlinear characteristics; a state space model estimation unit (2) that estimates a state space model representing a linear approximation curve for the plurality of observation values acquired by the observation value acquisition unit (1); and an upper-bound model estimation unit (3) that calculates estimation errors, which are errors between the observation values acquired by the observation value acquisition unit (1) and the linear approximation curve represented by the state space model estimated by the state space model estimation unit (2), and estimates an upper-bound model representing the upper bounds of the estimation errors. The control model generation device also comprises a control model generation unit (4) that uses the state space model estimated by the state space model estimation unit (2) and the upper-bound model estimated by the upper-bound model estimation unit (3) to generate a control model representing an equation of motion of the controlled object.

Description

[Technical Field]

[0001] This disclosure relates to a control model generation device and a control model generation method. [Background technology]

[0002] There is a control model generation device that generates a control model representing the equations of motion of a controlled object with nonlinear characteristics. As an example of such a control model generation device, Patent Document 1 discloses a control device that includes multiple control model candidates that are all different from each other. Each control model candidate is prepared in advance. The control device calculates the error between the output of each control model candidate and the output of the controlled object, and based on the error calculation result, selects one of the control model candidates from among the multiple control model candidates. [Prior art documents] [Patent Documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 5-303408 [Overview of the project] [Problems that the invention aims to solve]

[0004] If the person attempting to prepare a candidate control model does not possess sufficient knowledge of the operation of the controlled object, it is generally difficult to prepare a candidate control model that represents the equations of motion of the controlled object. The control device disclosed in Patent Document 1 had the problem that multiple control model candidates had to be prepared in advance. For example, if none of the multiple control model candidates represent the equations of motion of the controlled object, the control device may not be able to control the controlled object accurately even if it selects one of the control model candidates from among the multiple candidates.

[0005] This disclosure was made to solve the above-mentioned problems, and aims to provide a control model generation device that allows a person with sufficient knowledge of the operation of the controlled object to generate a control model representing the equations of motion of the controlled object without having to prepare multiple control model candidates in advance. [Means for solving the problem]

[0006] The control model generation device according to this disclosure comprises: an observation value acquisition unit that acquires multiple observation values ​​which are the output of a controlled object having nonlinear characteristics; a state space model estimation unit that estimates a state space model representing a linear approximation curve related to the multiple observation values ​​acquired by the observation value acquisition unit; and an upper bound model estimation unit that calculates an estimation error which is the error between each observation value acquired by the observation value acquisition unit and the linear approximation curve represented by the state space model estimated by the state space model estimation unit, and estimates an upper bound model which represents an upper bound of the estimation error. Furthermore, the control model generation device uses the state space model estimated by the state space model estimation unit and the upper bound model estimated by the upper bound model estimation unit to generate a control model that represents the equation of motion of the controlled object. As such, a control model in which the error between the output of the control model and the linear approximation curve is below the upper bound. It includes a control model generation unit that generates the following. [Effects of the Invention]

[0007] According to this disclosure, a person with sufficient knowledge of the operation of the controlled object can generate a control model that represents the equations of motion of the controlled object without having to prepare multiple candidate control models in advance. [Brief explanation of the drawing]

[0008] [Figure 1] This is a configuration diagram showing a control model generation device according to Embodiment 1. [Figure 2] This is a hardware configuration diagram showing the hardware of the control model generation device according to Embodiment 1. [Figure 3] This is a hardware configuration diagram of a computer when a control model generation device is implemented using software or firmware. [Figure 4]This diagram illustrates the relationship between a controller implementing the control model CM and a linear system that includes both the state-space model Jz and the upper bound model Hz. [Figure 5] Figure 5A is a graph showing an example of multiple observed values ​​f(z), a state-space model Jz, and an upper bound model Hz, which are the output of the controlled object OB. Figure 5B is a graph obtained by rotating the graph of Figure 5A so that the state-space model Jz is on the horizontal axis. [Figure 6] This flowchart shows the control model generation method, which is the processing procedure of the control model generation device. [Figure 7] This is a configuration diagram showing a control model generation device according to Embodiment 2. [Figure 8] This is a hardware configuration diagram showing the hardware of the control model generation device according to Embodiment 2. [Figure 9] Figures 9A and 9B are explanatory diagrams showing the model selection unit 9, which selects one of the M controllers 31-1 to 31-M, specifically controller 31-m. [Figure 10] This is an explanatory diagram showing the M substate spaces JK1 to JKM contained within the state space JK in which multiple observed values ​​f(z) exist. [Modes for carrying out the invention]

[0009] To provide a more detailed explanation of this disclosure, the forms for implementing this disclosure will be described below with reference to the attached drawings.

[0010] Embodiment 1. Figure 1 is a configuration diagram showing a control model generation device according to Embodiment 1. Figure 2 is a hardware configuration diagram showing the hardware of the control model generation device according to Embodiment 1. The control model generation device shown in Figure 1 comprises an observation value acquisition unit 1, a state space model estimation unit 2, an upper bound model estimation unit 3, and a control model generation unit 4.

[0011] The observation value acquisition unit 1 is implemented, for example, by the observation value acquisition circuit 11 shown in Figure 2. The observation unit 1 acquires multiple observation values ​​f(z), which are the output of the controlled object OB having nonlinear characteristics. The controlled object OB may be a known controlled object or an unknown controlled object. The observation value acquisition unit 1 outputs multiple observation values ​​f(z) to the state space model estimation unit 2 and the upper bound model estimation unit 3, respectively.

[0012] The state-space model estimation unit 2 is implemented, for example, by the state-space model estimation circuit 12 shown in Figure 2. The state-space model estimation unit 2 acquires multiple observed values ​​f(z) from the observed value acquisition unit 1. The state-space model estimation unit 2 estimates a state-space model Jz that represents a linear approximation curve relating to multiple observed values ​​f(z). The state-space model estimation unit 2 outputs the state-space model Jz to the upper bound model estimation unit 3 and the control model generation unit 4, respectively.

[0013] The upper bound model estimation unit 3 is implemented, for example, by the upper bound model estimation circuit 13 shown in Figure 2. The upper bound model estimation unit 3 obtains multiple observed values ​​f(z) from the observed value acquisition unit 1 and obtains the state-space model Jz from the state-space model estimation unit 2. The upper bound model estimation unit 3 calculates the estimation error |f(z)-Jz|, which is the error between each observed value f(z) and the linear approximation curve represented by the state-space model Jz. The upper bound model estimation unit 3 estimates the upper bound model Hz, which represents the upper bound of the estimation error |f(z)-Jz|. Specifically, the upper bound model estimation unit 3 estimates the upper bound model Hz using a loss function ζ(t) which represents the difference between the result of multiplying the squared value of the upper bound by a constant α less than or equal to 1, and the squared value of the estimation error |f(z)-Jz|. The upper bound model estimation unit 3 outputs the upper bound model Hz to the control model generation unit 4.

[0014] The control model generation unit 4 is implemented, for example, by the control model generation circuit 14 shown in Figure 2. The control model generation unit 4 obtains the state-space model Jz from the state-space model estimation unit 2 and the upper bound model Hz from the upper bound model estimation unit 3. The control model generation unit 4 uses the state-space model Jz and the upper bound model Hz to generate a control model CM that represents the equations of motion of the controlled object OB. Specifically, the control model generation unit 4 generates a control model CM such that the error between the output of the control model and the linear approximation curve is below the upper limit. The control model CM generated by the control model generation unit 4 is implemented in the controller that controls the controlled object OB.

[0015] In Figure 1, the control model generation device is assumed to be implemented by dedicated hardware as shown in Figure 2, with each of its components—the observation acquisition unit 1, the state-space model estimation unit 2, the upper bound model estimation unit 3, and the control model generation unit 4—being realized by dedicated hardware. Specifically, the control model generation device is assumed to be implemented by an observation acquisition circuit 11, a state-space model estimation circuit 12, an upper bound model estimation circuit 13, and a control model generation circuit 14. Each of the observation value acquisition circuit 11, state space model estimation circuit 12, upper bound model estimation circuit 13, and control model generation circuit 14 can be, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a combination thereof.

[0016] The components of the control model generation device are not limited to those implemented by dedicated hardware; the control model generation device may also be implemented by software, firmware, or a combination of software and firmware. Software or firmware is stored as a program in the computer's memory. A computer refers to the hardware that executes programs, and includes, for example, a CPU (Central Processing Unit), GPU (Graphics Processing Unit), central processing unit, processing unit, arithmetic unit, microprocessor, microcomputer, processor, or DSP (Digital Signal Processor).

[0017] Figure 3 is a hardware diagram of a computer when the control model generation device is implemented using software or firmware. When the control model generation device is implemented using software or firmware, a program is stored in memory 21 that causes the computer to execute the respective processing procedures in the observation value acquisition unit 1, the state space model estimation unit 2, the upper bound model estimation unit 3, and the control model generation unit 4. The computer's processor 22 then executes the program stored in memory 21.

[0018] Furthermore, Figure 2 shows an example where each component of the control model generation device is implemented by dedicated hardware, and Figure 3 shows an example where the control model generation device is implemented by software or firmware, etc. However, this is only one example, and some components of the control model generation device may be implemented by dedicated hardware, while the remaining components may be implemented by software or firmware, etc.

[0019] Figure 4 is an explanatory diagram showing the relationship between a controller implementing the control model CM and a linear system that includes both the state-space model Jz and the upper bound model Hz. Figure 4 shows that the output x of the state-space model Jz is given to the controller, and the output u of the controller is given to the linear system. Furthermore, Figure 4 shows that the output p of the upper bound model Hz is applied to the disturbance, and the disturbance q is applied to the linear system.

[0020] Figure 5A is a graph showing an example of multiple observed values ​​f(z), a state-space model Jz, and an upper bound model Hz, which are the output of the controlled object OB. Figure 5A shows an example of a two-dimensional space where the state space containing multiple observed values ​​f(z), which are the outputs of the controlled object OB, is represented by the z-axis and the x-axis. However, this is just one example, and the state space may be, for example, a three-dimensional space. Figure 5B is a graph obtained by rotating the graph in Figure 5A so that the state-space model Jz is on the horizontal axis.

[0021] The equation of motion for the controlled object OB, x-dot, is expressed as shown in equation (1) below. Due to the limitations of electronic filing, it is not possible to place a "·" symbol above the letter x in the specification document, so it is written as x-dot.

[0022] In formula (2) of TIFF0007843938000001.tif23166, T is a mathematical symbol indicating transpose.

[0023] The relationship between the multiple observed values ​​f(z), which are the output of the controlled object OB, the state-space model Jz, and the upper bound model Hz is expressed by equation (3) below.

[0024] TIFF0007843938000002.tif15166

[0025] The state-space model Jz is expressed as shown in equation (4) below. The estimated error q, which is the error between the observed value f(z) and the linear approximation curve represented by the state-space model Jz, is expressed as shown in equation (5) below. The upper bound model Hz is expressed as shown in equation (6) below.

[0026] In equations (4) and (6) of TIFF0007843938000003.tif34166, A, B, C, and D are arbitrary matrices.

[0027] Loss function L of state-space model JzJ This can be expressed as shown in equation (7) below. Also, the loss function L of the upper bound model Hz is... H This can be expressed as shown in equation (8) below.

[0028] In equation (9) of TIFF0007843938000004.tif58166, ζ(t) is a loss function that shows the difference between the square of the upper bound p(t) multiplied by the constant α and the square of the estimation error q(t). α is a constant less than or equal to 1.

[0029] Next, we will explain the operation of the control model generation device shown in Figure 1. Figure 6 is a flowchart showing the control model generation method, which is the processing procedure of the control model generation device.

[0030] The observation value acquisition unit 1 acquires multiple observation values ​​f(z), which are the output of the controlled object OB having nonlinear characteristics, from an external source (step ST1 in Figure 6). The observation value acquisition unit 1 outputs multiple observation values ​​f(z) to the state space model estimation unit 2 and the upper bound model estimation unit 3, respectively.

[0031] The state-space model estimation unit 2 acquires multiple observed values ​​f(z) from the observed value acquisition unit 1. As shown in Figures 5A and 5B, the state-space model estimation unit 2 estimates a state-space model Jz that represents a linear approximation curve relating to multiple observed values ​​f(z) (step ST2 in Figure 6). The state-space model Jz, which represents the linear approximation curve relating to multiple observed values ​​f(z), is expressed as shown in equation (4). The process of estimating the state-space model Jz is a well-known technique, so a detailed explanation will be omitted. However, for example, the state-space model Jz can be estimated by using a method that searches for a linear approximation curve that minimizes the expected value of the error vector representing the error between multiple observed values ​​f(z) and the linear approximation curve. The state-space model estimation unit 2 outputs the state-space model Jz to the upper bound model estimation unit 3 and the control model generation unit 4, respectively.

[0032] The upper bound model estimation unit 3 obtains multiple observed values ​​f(z) from the observed value acquisition unit 1 and obtains the state-space model Jz from the state-space model estimation unit 2. The upper bound model estimation unit 3 calculates the estimation error q, which is the error between each observed value f(z) and the linear approximation curve represented by the state-space model Jz, as shown in equation (5) (step ST3 in Figure 6). The upper bound model estimation unit 3 estimates an upper bound model Hz, which represents an upper bound on the estimation error q, as shown in equation (6) (step ST4 in Figure 6). Specifically, the upper bound model estimation unit 3 estimates the upper bound model Hz using a loss function ζ(t) that represents the difference between the square of the estimated error q and the result of multiplying the square of the upper bound p by a constant α less than or equal to 1, as shown in equation (9). The loss function ζ(t) is a function that outputs a vector that suppresses the norm of the error vector representing the difference from above. The upper bound model estimation unit 3 outputs the upper bound model Hz to the control model generation unit 4.

[0033] The control model generation unit 4 obtains the state-space model Jz from the state-space model estimation unit 2 and the upper bound model Hz from the upper bound model estimation unit 3. As shown in equation (3), the control model generation unit 4 generates a control model CM representing the equations of motion of the controlled object OB using the state-space model Jz and the upper bound model Hz (step ST5 in Figure 6). The f(z) that satisfies equation (3) corresponds to the control model CM which represents the equation of motion of the controlled object OB. The control model CM generated by the control model generation unit 4 is implemented in the controller that controls the controlled object OB. This constructs a controller in which the error between the output of the control model CM and the output of the state-space model Jz is considered as an error input to the controlled object OB. The upper bound of the error between the output of the control model CM and the output of the state-space model Jz is the upper bound represented by the upper bound model Hz.

[0034] In the above embodiment 1, the control model generation device is configured to include: an observation value acquisition unit 1 that acquires multiple observation values ​​which are the output of a controlled object having nonlinear characteristics; a state space model estimation unit 2 that estimates a state space model representing a linear approximation curve related to the multiple observation values ​​acquired by the observation value acquisition unit 1; and an upper bound model estimation unit 3 that calculates an estimation error, which is the error between each observation value acquired by the observation value acquisition unit 1 and the linear approximation curve represented by the state space model estimated by the state space model estimation unit 2, and estimates an upper bound model that represents an upper bound of the estimation error. Furthermore, the control model generation device includes a control model generation unit 4 that generates a control model representing the equations of motion of the controlled object using the state space model estimated by the state space model estimation unit 2 and the upper bound model estimated by the upper bound model estimation unit 3.Therefore, the control model generation device can generate a control model representing the equations of motion of the controlled object without requiring a person with sufficient knowledge of the operation of the controlled object to prepare multiple control model candidates in advance.

[0035] Embodiment 2. Embodiment 2 describes a control model generation device that includes a state space division unit 5 that divides a state space containing multiple observed values ​​f(z) into multiple substate spaces.

[0036] Figure 7 is a configuration diagram showing a control model generation device according to Embodiment 2. In Figure 7, the same reference numerals as in Figure 1 indicate the same or corresponding parts, so a detailed explanation is omitted. Figure 8 is a hardware configuration diagram showing the hardware of the control model generation device according to Embodiment 2. In Figure 8, the same reference numerals as in Figure 2 indicate the same or corresponding parts, so a detailed explanation is omitted. The control model generation device shown in Figure 7 comprises an observation value acquisition unit 1, a state space division unit 5, a state space model estimation unit 6, an upper bound model estimation unit 7, a control model generation unit 8, and a model selection unit 9.

[0037] The state space division unit 5 is implemented, for example, by the state space division circuit 15 shown in Figure 8. The state space division unit 5 acquires a plurality of observation values f(z) from the observation value acquisition unit 1. The state space division unit 5 divides the state space JK in which a plurality of observation values f(z) exist into partial state spaces JK1 to JK which are a plurality of spaces. M M is an integer of 2 or more.

[0038] The state space model estimation unit 6 is realized by, for example, the state space model estimation circuit 16 shown in FIG. 8. The state space model estimation unit 6 acquires a plurality of observation values f(z) from the observation value acquisition unit 1. Among the plurality of observation values f(z), the state space model estimation unit 6 estimates a partial state space model Jz which is a state space model representing a linear approximation curve related to the observation value f(z) existing in each partial state space JK (m = 1, ···, M) after being divided by the state space division unit 5. m (m = 1, ···, M) m (z) m The state space model estimation unit 6 outputs each partial state space model Jz to each of the upper bound model estimation unit 7 and the control model generation unit 8. m

[0039] The upper bound model estimation unit 7 is realized by, for example, the upper bound model estimation circuit 17 shown in FIG. 8. The upper bound model estimation unit 7 acquires a plurality of observation values f(z) from the observation value acquisition unit 1 and acquires each partial state space model Jz (m = 1, ···, M) from the state space model estimation unit 6. m (m = 1, ···, M) The upper bound model estimation unit 7 calculates an estimation error q which is the error between the observation value f(z) existing in each partial state space JK and the linear approximation curve represented by each partial state space model Jz. <严 m m (z) m ... m The upper bound model estimation unit 7 estimates a partial upper bound model Hz (m = 1, ···, M) which is an upper bound model representing the upper bound of each estimation error q. m m (m = 1, ···, M) The upper bound model estimation unit 7 outputs each partial upper bound model Hz to each of the upper bound model estimation unit 7 and the control model generation unit 8. It should be noted that there seems to be some incorrect or incomplete tags in the original text (such as "<严 m "), and the translation is carried out as accurately as possible based on the existing content.m This is output to the control model generation unit 8.

[0040] The control model generation unit 8 is implemented, for example, by the control model generation circuit 18 shown in Figure 8. The control model generation unit 8 generates a substate-space model Jz from the state-space model estimation unit 6. m (m=1,···,M) is obtained, and the upper bound model estimation unit 3 is used to obtain the subupper bound model Hz m Get (m=1,···,M). The control model generation unit 8 generates the partial state space model Jz m and the sub-upper bound model Hz m Using and, the substate space JK m Corresponding control model CM m Generate (m=1,···,M). Control model CM generated by control model generation unit 8 m This will be implemented in the controller 31-m (m=1,···,M) described later.

[0041] The model selection unit 9 is implemented, for example, by the model selection circuit 19 shown in Figure 8. The model selection unit 9 selects M substate spaces JK1 to JK1 generated by the control model generation unit 8. M Control model CM corresponding to each of these m Select one control model from (m=1,···,M). Specifically, the model selection unit 9 selects M substate spaces JK1 to JK M Control model CM corresponding to each of these m The output and the substate space model Jz m The error Δe from the linear approximation curve represented by m Calculate (m=1,···,M). The model selection unit 9 selects M errors Δe1~Δe M Based on the calculation results, one control model is selected.

[0042] In Figure 7, the control model generation device is assumed to be implemented by dedicated hardware as shown in Figure 8, with each of its components—the observation acquisition unit 1, the state space division unit 5, the state space model estimation unit 6, the upper bound model estimation unit 7, the control model generation unit 8, and the model selection unit 9—being realized by dedicated hardware. Specifically, the control model generation device is assumed to be implemented by the observation acquisition circuit 11, the state space division circuit 15, the state space model estimation circuit 16, the upper bound model estimation circuit 17, the control model generation circuit 18, and the model selection circuit 19. Each of the observation acquisition circuit 11, state space partitioning circuit 15, state space model estimation circuit 16, upper bound model estimation circuit 17, control model generation circuit 18, and model selection circuit 19 can be, for example, a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, an ASIC, an FPGA, or a combination thereof.

[0043] The components of the control model generation device are not limited to those implemented by dedicated hardware; the control model generation device may also be implemented by software, firmware, or a combination of software and firmware. When the control model generation device is implemented using software or firmware, a program is stored in the memory 21 shown in Figure 3 that causes the computer to execute the respective processing procedures in the observation value acquisition unit 1, the state space division unit 5, the state space model estimation unit 6, the upper bound model estimation unit 7, the control model generation unit 8, and the model selection unit 9. Then, the processor 22 shown in Figure 3 executes the program stored in the memory 21.

[0044] Furthermore, Figure 8 shows an example where each component of the control model generation device is implemented by dedicated hardware, while Figure 3 shows an example where the control model generation device is implemented by software or firmware. However, this is merely one example, and some components of the control model generation device may be implemented by dedicated hardware, while the remaining components may be implemented by software or firmware.

[0045] Figures 9A and 9B are explanatory diagrams showing the model selection unit 9, which selects one of the M controllers 31-1 to 31-M, specifically controller 31-m. Each of the controllers 31-1 to 31-M is a controller that controls the controlled object OB. Controller 31-m (m=1,···,M) has a substate space Jz m It supports this. Figure 10 shows the M substate spaces JK1~JK contained within the state space JK where multiple observed values ​​f(z) exist. M This is an explanatory diagram illustrating this. Figure 10 shows an example where M=3.

[0046] Next, we will explain the operation of the control model generation device shown in Figure 7. The observation value acquisition unit 1 acquires multiple observation values ​​f(z) from an external source, which are the output of the controlled object OB having nonlinear characteristics. The observation value acquisition unit 1 outputs multiple observation values ​​f(z) to the state space partitioning unit 5, the state space model estimation unit 6, and the upper bound model estimation unit 7, respectively.

[0047] The state space division unit 5 acquires multiple observed values ​​f(z) from the observed value acquisition unit 1. As shown in Figure 10, the state space partitioning unit 5 divides the state space JK in which multiple observed values ​​f(z) exist into multiple substate spaces JK1 to JK. M Divide it into parts.

[0048] The state-space model estimation unit 6 acquires multiple observed values ​​f(z) from the observed value acquisition unit 1. The state-space model estimation unit 6 determines the substate space JK of the multiple observed values ​​f(z). m Observed values ​​f present at (m=1,···,M) m Identify (z). The state space model estimation unit 6 determines the substate space JK m Observed value f present m The substate-space model Jz is a state-space model that represents the linear approximation curve related to (z). m We estimate this. Substate space model Jz m This can be expressed as shown in equation (10) below. Substate space model Jz m Since the process of estimating this is a publicly known technique, a detailed explanation will be omitted. The state-space model estimation unit 6 uses the substate-space model Jz m (m=1,···,M) is output to the upper bound model estimation unit 7 and the control model generation unit 8, respectively.

[0049] In formula (10) TIFF0007843938000005.tif13166, A m ,B m This is an arbitrary matrix.

[0050] The upper bound model estimation unit 7 obtains multiple observed values ​​f(z) from the observed value acquisition unit 1 and estimates each substate-space model Jz from the state-space model estimation unit 6. m Get (m=1,···,M). The upper bound model estimation unit 7 determines each substate space JK m Observed value f present m (z) and their respective substate space models Jz m The estimated error q is the error between the linear approximation curve represented by . m Calculate the estimated error q. m The calculation process is performed according to the following formula (11). The upper bound model estimation unit 7 calculates the respective estimation error q m The subupper bound model Hz is an upper bound model that represents an upper bound. m We estimate (m=1,···,M). Subupper bound model Hz m The estimation process for (m=1,···,M) is performed according to the following equation (12). The upper bound model estimation unit 7 estimates the subupper bound model Hz. m The control model generation unit 8 outputs (m=1,···,M).

[0051] In formula (12) of TIFF0007843938000006.tif27166, C m ,Dm This is an arbitrary matrix.

[0052] The control model generation unit 8 generates a substate-space model Jz from the state-space model estimation unit 6. m (m=1,···,M) is obtained, and the upper bound model estimation unit 3 is used to obtain the subupper bound model Hz m Get (m=1,···,M). The control model generation unit 8 generates a partial state space model Jz as shown in equation (13) below. m and the sub-upper bound model Hz m Using and, the substate space JK m Corresponding control model CM m Generates. f that satisfies equation (13) m (z) is a sub-state space JK m The control model CM represents the equations of motion for the controlled object OB. m It corresponds to this. Control model CM generated by control model generation unit 8 m This will be implemented in controller 31-m (m=1,···,M).

[0053] TIFF0007843938000007.tif12166

[0054] The model selection unit 9 consists of M substate spaces JK1 to JK M Control model CM corresponding to each of these m Choose one control model CM from among (m=1,···,M) m Select this option. Specifically, the model selection unit 9 has M substate spaces JK1 to JK as shown in equation (14) below. M Control model CM corresponding to each of these m Output u m And, the substate space model Jz m The error Δe from the linear approximation curve represented by m Calculate (m=1,···,M).

[0055] TIFF0007843938000008.tif14166

[0056] The model selection unit 9 selects M errors Δe1~Δe M Based on the calculation results, M control models CM1 to CM M Choose one control model CM from the following: m Select this option. Specifically, the model selection unit 9 calculates the error Δe1~Δe M Among them, the smallest error Δe MIN Identify. The model selection unit 9 selects M control models CM1 to CM M From among them, the smallest error Δe MIN Corresponding control model CM m Select this option. Of the M controllers 31-1 to 31-M, the control model CM selected by the model selection unit 9 is selected. m Controller 31-m, which implements this, controls the controlled object OB.

[0057] In the control model generation device shown in Figure 7, the model selection unit 9 selects M control models CM1 to CM M From among them, the smallest error Δe MIN Corresponding control model CM m This is the selection. However, this is just one example, and the model selection unit 9 selects the smallest error Δe within a range that does not cause practical problems. MIN Control model CM for errors other than those mentioned above. m It may also be possible to select the following: Specifically, the model selection unit 9 selects the control model CM with the second smallest error. m Select the control model CM with the third smallest error. m You may choose to select that option.

[0058] In the above-described Embodiment 2, the control model generation device shown in FIG. 7 is configured to include a model selection unit 9 that selects any one of the control models corresponding to each of the plurality of partial state spaces generated by the control model generation unit 8. Therefore, similar to the control model generation device shown in FIG. 1, the control model generation device shown in FIG. 7 enables a person with sufficient knowledge of the operation of the control target to generate a control model representing the equation of motion of the control target without preparing a plurality of control model candidates in advance. In addition, the control accuracy by the controller can be improved more than that of the control model generation device shown in FIG. 1.

[0059] In the control model generation device shown in FIG. 7, the model selection unit 9 calculates the error Δe M (m = 1, ···, M) between the output of the control model CM m corresponding to each of the M partial state spaces JK1 to JK m and the linear approximation curve represented by the partial state space model Jz m and selects any one of the control models based on the calculation results of the M errors Δe1 to Δe M . However, this is merely an example. The model selection unit 9 may identify the partial upper bound model Hz M with the smallest upper bound slope, which is the gain of the partial upper bound model, among the M partial upper bound models Hz1 to Hz m estimated by the upper bound model estimation unit 7, and select the control model CM M corresponding to the identified partial upper bound model Hz m from among the M control models CM1 to CM m .

[0060] Note that in the present disclosure, free combinations of the embodiments, modifications of any components of the embodiments, or omissions of any components in the embodiments are possible.

Industrial Applicability

[0061] The present disclosure is suitable for a control model generation device and a control model generation method.

Explanation of Signs

[0062] 1 Observation value acquisition unit, 2 State space model estimation unit, 3 Upper bound model estimation unit, 4 Control model generation unit, 5 State space partitioning unit, 6 State space model estimation unit, 7 Upper bound model estimation unit, 8 Control model generation unit, 9 Model selection unit, 11 Observation value acquisition circuit, 12 State space model estimation circuit, 13 Upper bound model estimation circuit, 14 Control model generation circuit, 15 State space partitioning circuit, 16 State space model estimation circuit, 17 Upper bound model estimation circuit, 18 Control model generation circuit, 19 Model selection circuit, 21 Memory, 22 Processor, 31-1~31-M Controller.

Claims

1. An observation value acquisition unit that acquires multiple observation values ​​which are the output of a controlled object having nonlinear characteristics, A state-space model estimation unit estimates a state-space model that represents a linear approximation curve related to a plurality of observed values ​​obtained by the observation value acquisition unit, An upper bound model estimation unit calculates an estimation error, which is the error between each observed value obtained by the observation value acquisition unit and the linear approximation curve represented by the state space model estimated by the state space model estimation unit, and estimates an upper bound model that represents an upper bound of the estimation error. A control model generation unit generates a control model that represents the equations of motion of the controlled object, using the state-space model estimated by the state-space model estimation unit and the upper bound model estimated by the upper bound model estimation unit, such that the error between the output of the control model and the linear approximation curve is less than or equal to the upper bound. A control model generation device equipped with the following features.

2. The aforementioned upper bound model estimation unit, The control model generation apparatus according to claim 1, characterized in that it estimates the upper bound model using a loss function that shows the difference between the result of multiplying the squared value of the upper bound by a constant of 1 or less and the squared value of the estimation error.

3. The state space division unit divides the state space containing multiple observed values ​​acquired by the observation value acquisition unit into multiple substate spaces, which are substate spaces. The state-space model estimation unit, From among the multiple observed values ​​acquired by the observation value acquisition unit, a substate-space model is estimated, which is a state-space model representing a linear approximation curve related to the observed values ​​present in each substate-space after division by the state-space division unit. The aforementioned upper bound model estimation unit, We calculate the estimation error, which is the error between the observed values ​​in each sub-state space and the linear approximation curve represented by each sub-state space model, and estimate a sub-upper bound model, which is an upper bound model representing the upper bound of each estimation error. The control model generation unit, The control model generation device according to claim 1, characterized in that it generates a control model representing the equations of motion of the controlled object, corresponding to each of the substate space models estimated by the state space model estimation unit and each of the subupper bound models estimated by the upper bound model estimation unit.

4. The control model generation apparatus according to claim 3, further comprising a model selection unit that selects one control model from among the control models corresponding to each of the plurality of partial state spaces generated by the control model generation unit.

5. The aforementioned model selection unit is The control model generation apparatus according to claim 4, characterized in that it calculates the error between the output of each of the plurality of substate spaces and the linear approximation curve, and selects one of the control models based on the calculation result of the error between the output of each control model and the linear approximation curve.

6. The observation value acquisition unit acquires multiple observation values, which are the output of a controlled object having nonlinear characteristics. The state-space model estimation unit estimates a state-space model that represents a linear approximation curve relating to a plurality of observed values ​​obtained by the observation value acquisition unit. The upper bound model estimation unit calculates the estimation error, which is the error between each observed value obtained by the observed value acquisition unit and the linear approximation curve represented by the state space model estimated by the state space model estimation unit, and estimates an upper bound model that represents the upper bound of the estimation error. The control model generation unit uses the state-space model estimated by the state-space model estimation unit and the upper bound model estimated by the upper bound model estimation unit to generate a control model that represents the equations of motion of the controlled object, such that the error between the output of the control model and the linear approximation curve is less than or equal to the upper bound. A method for generating control models.

Images