Control input generating device, control device, control input generating method, actuator, arithmetic circuit, control board, control input generating program, and storage medium having control input generating program stored therein

Abstract

[Problem] To provide a control input generating device that generates a control input for controlling a pneumatic actuator with high precision. [Solution] This device generates a control input for controlling a pneumatic actuator, and includes: a first control input generation function for generating a first control input if there is a deviation of a controlled quantity from a target value; a second control input generation function for generating a second control input that increases as the deviation of the controlled quantity from the target value decreases, if there is a change in the controlled quantity over time; and an output generation function for generating the control input on the basis of the first control input and the second control input. If the generated control input is used to control the pneumatic actuator, the responsiveness of the pneumatic actuator in the vicinity of the target value can be improved, and overshoot of the target value can be suppressed, thereby making it possible to increase the accuracy with which the pneumatic actuator is operated to reach the target value. Moreover, overshoot of the target value can be prevented.

Description

Control input generation device, control device, control input generation method, actuator, arithmetic circuit, control board, control input generation program, and storage medium storing the control input generation program

[0001] The present invention relates to a control input generating device, a control device, a control input generating method, an actuator, an arithmetic circuit, a control board, a control input generating program, and a control input generating program. More specifically, the present invention relates to a control input generating device, a control device, a control input generating method, an actuator, an arithmetic circuit, a control board, a control input generating program, and a storage medium storing a control input generating program that are suitable for controlling a pneumatic actuator that uses air pressure as a working fluid, such as a pneumatic cylinder.

[0002] Various devices are used as actuators for driving robots, mechatronic devices, etc., and hydraulic cylinders and pneumatic cylinders are commonly used as actuators that utilize fluid pressure. For example, when an object (actuation object) operated by such an actuator is moved to a target position and placed at that position, that is, when the actuation object is moved a predetermined amount, the actuator is controlled based on the difference (deviation) between the current position of the actuation object and the target position. In other words, the current position of the actuation object is measured by a sensor, and the actuator is controlled so that the deviation between the measured position and the target position is zero.

[0003] Among these actuators, pneumatic cylinders have the advantages of being able to drive an actuated object at high speed and having a high power-to-weight ratio. Furthermore, compared to hydraulic cylinders, they also have the advantages of being relatively simple and inexpensive to construct, being explosion-proof, clean, environmentally friendly, and easy to maintain.

[0004] On the other hand, pneumatic cylinders use compressible air as the working fluid, which results in strong nonlinearity in the pressure response section. Furthermore, due to their low rigidity, they are susceptible to the effects of frictional forces (such as the Coulomb friction between the head and cylinder and the maximum static friction) in the various components that make up the motion control system. For these reasons, pneumatic cylinders have the disadvantage of being inferior in positioning accuracy compared to other actuators, such as hydraulic cylinders, when stopping an actuated object at a target position. Therefore, conventional control methods using PID control make it difficult to achieve micron-order motion control using pneumatic cylinders.

[0005] In an actuator using a pneumatic cylinder that has the above-mentioned problems, there are techniques for improving the positioning accuracy, as disclosed in Non-Patent Documents 1 and 2.

[0006] Non-Patent Documents 1 and 2 disclose a device that moves an object on a plane using a pneumatic cylinder, and a technology that controls the amount of movement of the object by controlling the drive of the pneumatic cylinder based on a signal (position signal) that senses the position of the object. Both of the technologies in Non-Patent Documents 1 and 2 improve positioning accuracy by suppressing the effect on control of the maximum static friction force acting between the object and the plane.

[0007] First, in Non-Patent Document 1, a control signal is formed by adding a compensation signal that compensates for a main signal that moves an object based on the difference between the current position obtained from a position signal and a target position. Specifically, the influence of the maximum static friction force on control is suppressed by adjusting the compensation signal based on the natural frequency of the feedback control system. Non-Patent Document 1 also describes that by using this technology, the moving position of an object can be adjusted with an accuracy of about 20 nm relative to the target position, even when the object is moved by a pneumatic cylinder.

[0008] Furthermore, Non-Patent Document 2 discloses a technique for controlling the drive of a pneumatic cylinder using a pulse signal with an adjusted pulse width. Non-Patent Document 2 also describes that by using this technique, even if an object is moved by a pneumatic cylinder, the position of the object can be adjusted with an accuracy of approximately 50 nm relative to the target position.

[0009] Kei-Ren PAI and Ming-Qiang SHTH, nanoaccuracy Position Control of a Pneumatic Cylinder Driven Table, JSME International Journal, Series C, Vol.46, No. 3,1062- 1067, 2003Kaiji Sato, Yusuke Sano, Practical and intuitive controller design method for precision positioning of a pneumatic cylinder actuator stage, Precision Engineering 38 703-710, 2014

[0010] However, the technology of Non-Patent Document 1 adjusts the compensation signal based on the natural frequency of the feedback control system, and therefore high-precision position adjustment is only possible with a feedback control system whose natural frequency is known. Typically, the natural frequency of a feedback control system including an actuator changes depending on the mass of the object moved by the actuator, aging deterioration of the actuator, and other factors. Therefore, while the technology of Non-Patent Document 1 can be used to control experimental equipment in laboratories, it is difficult to adopt it for controlling actuators and other devices used in actual robots, mechatronic equipment, and the like. Furthermore, Non-Patent Document 1 presumably uses air bearings to achieve positioning control by minimizing the effects of maximum static friction, etc., which also makes it difficult to adopt it in actual robots, mechatronic equipment, and the like to perform high-precision position adjustment.

[0011] Furthermore, the technology of Non-Patent Document 2 cannot achieve high-precision position adjustment unless it is possible to determine the pulse width that allows the device including the pneumatic cylinder to be positioned appropriately. Such pulse width also changes depending on the mass of the object moved by the actuator, deterioration of the actuator over time, etc. Therefore, although the technology of Non-Patent Document 2 can be used to control experimental devices in laboratories, it is difficult to use it to control actuators used in actual robots, mechatronics equipment, etc.

[0012] In view of the above circumstances, the present invention aims to provide a control input generation device, an arithmetic circuit, and a control board that generate control inputs for controlling a pneumatic actuator with high precision. It also aims to provide a control device equipped with a control input generation device that generates control inputs for controlling a pneumatic actuator with high precision, and an actuator equipped with such a control device. It also aims to provide a control input generation method for controlling the drive of a pneumatic actuator with high precision, a control input generation program for executing such a control input generation method, and a storage medium on which the control input generation program is stored.

[0013] <Control Input Generating Device> A control input generating device of a first invention is a device that generates a control input for controlling a pneumatic actuator, and is characterized in having: a first control input generating function that generates a first control input when there is a deviation of a controlled variable from a target value; a second control input generating function that generates a second control input that increases as the deviation of the controlled variable from the target value decreases when there is a time change in the controlled variable; and an output generating function that generates the control input based on the first control input and the second control input. A control input generating device of a second invention is the first invention, characterized in that the first control input generating function has a function of generating a steady-state control input as the first control input, or has a function of generating, as the first control input, a control input that combines a steady-state control input and a variable control input that decreases as the deviation of the controlled variable from the target value decreases when there is a time change in the controlled variable. A third aspect of the control input generating device is the first aspect, wherein the second control input generating function calculates an expected number of steps N until the controlled variable matches the target value based on the deviation of the controlled variable from the target value and the operating state of the pneumatic actuator, and generates the second control input based on the expected number of steps N so that the second control input increases as the expected number of steps N decreases. A fourth aspect of the control input generating device is the second or third aspect, wherein the first control input generating function generates the first control input S1 based on Equation 1, the second control input generating function generates the second control input S2 based on Equation 2, and the output generating function generates the control input based on Equation 3. Equation 1 S1=K １ ・sgn(e) Formula 2 S2=K2 / N α ・sgn(v) Formula 3 Yc=K １ ・sgn(e)-K2 / N α sgn(v) e: Deviation of the controlled variable from the target value v: Displacement speed of the pneumatic actuator α: Index for adjusting the timing at which the suppression force that suppresses the operation of the pneumatic actuator begins to act K １ = Gain that generates thrust K ２= Gain that generates a restraining force N = Number of predicted steps The control input generating device of a fifth aspect of the present invention is the fourth aspect of the present invention, wherein the first control input generating function compensates for nonlinear friction of the pneumatic actuator based on Equation 2-1 and generates a gain K １ and the second control input generation function generates a gain K ２ The feature is to set the following equation: １ = Ka | e | > Eth K １ = Ka / (Eth·|e|) when |e|≦Eth Ka: specified control input Eth: specified displacement |e|: absolute value of deviation e Formula 2-2 K ２ =Kb |v|>Vth K ２ = 0 |v| ≦ Vth Kb: specified control input Vth: specified displacement velocity |v|: absolute value of velocity v A control input generating device according to a sixth aspect of the present invention is a third aspect, wherein the second control input generating function has an arrival time calculating function that calculates the time Ta required for the controlled variable to match the target value based on the deviation of the controlled variable from the target value, and an expected step number calculating function that calculates an expected step number N = Ta / Δt, which is the value obtained by dividing the time Ta by a sampling period Δt that detects the deviation of the controlled variable from the target value, and wherein the expected step number calculating function sets the expected step number N as a threshold value when the expected step number N is smaller than a predetermined threshold. A control input generating device according to a seventh aspect of the present invention is a sixth aspect, wherein the arrival time calculating function calculates the arrival time Ta based on Equation 4: Equation 4 (1 / 2) a Ta ２+v・Ta-e=0 a: displacement acceleration of the pneumatic actuator v: displacement velocity of the pneumatic actuator Ta: arrival time (arrival time - current time) e: deviation (target position - current position) A control input generating device of an eighth aspect of the invention is a third aspect of the invention, wherein the second control input generation function calculates the predicted number of steps N based on the following equation 5: Equation 5 N=e / ΔL=e / (v・Δt) e: deviation of the controlled variable from the target value v: displacement velocity of the pneumatic actuator ΔL: amount of change in displacement from one sampling time before A control input generating device of a ninth aspect of the invention is a first aspect of the invention, wherein the target value is the target position of an object operated by the pneumatic actuator, the drive velocity of the pneumatic actuator, or the pressure generated by the pneumatic actuator. <Control Device> A control device of a tenth aspect of the invention is a control device for controlling a pneumatic actuator, and is characterized by having the control input generating device of the first aspect of the invention. A control device of an eleventh invention is a control device for controlling a pneumatic actuator, and includes a main control input generation unit that generates a main control input to control the pneumatic actuator, and a compensation control input generation unit that generates a compensation control input to correct the main control input, wherein the compensation control input generation unit is the control input generation device of the first invention. A control device of a twelfth invention is the tenth or eleventh invention, characterized in that the control device is a type 2 control system having a disturbance observer. A control device of a thirteenth invention is the twelfth invention, characterized in that the control device performs feedforward control. <Actuator> An actuator of a fourteenth invention is characterized in that it has an operating unit that operates by air pressure and a control unit that controls the operation of the operating unit, wherein the control unit is the control device of the tenth or eleventh invention.<Arithmetic Circuit> The arithmetic circuit of the fifteenth invention is a circuit to which a target value and a controlled variable are input as input signals, and is characterized by having: a first control input generation unit that generates a first control input when there is a deviation of the controlled variable from the target value; a second control input generation unit that generates a second control input that increases as the deviation of the controlled variable from the target value decreases when there is a change in the controlled variable over time; and an output generation unit that generates and outputs a control input based on the first control input and the second control input. <Control Board> A control board of a sixteenth invention is a control board having an input terminal to which a target value and a control amount are input as input signals, and an output terminal that generates and outputs a control input, the control board being provided with: a main controller that generates a main control input based on the signal input from the input terminal; a compensator that generates a compensation control input based on the signal input from the input terminal; and an output generation unit that is an adder circuit that combines the main control input and the compensation control input to generate a control input and supplies the control input to the output terminal, the compensator having a first control input generation function that generates a first control input when there is a deviation of the controlled amount from the target value; a second control input generation function that generates a second control input that increases as the deviation of the controlled amount from the target value decreases when there is a time change in the controlled amount; and an output generation function that generates and outputs a control input based on the first control input and the second control input. A control board of the seventeenth invention is a control board having an input terminal to which a target value and a control amount are input as input signals, and an output terminal to generate and output a control input, wherein the control board is provided with a compensator that generates a compensation control input based on the signal input from the input terminal, and an output generation unit that is an adder circuit that combines the compensation control input with a control input supplied from outside to generate a control input and supplies the control input to the output terminal, wherein the compensator has a first control input generation function that generates a first control input when there is a deviation of the control amount from the target value, a second control input generation function that generates a second control input that increases as the deviation of the control amount from the target value decreases when there is a time change in the control amount, and an output generation function that generates and outputs a control input based on the first control input and the second control input.<Control Input Generation Method> A control input generation method of an 18th invention is a method for generating a control input for controlling a pneumatic actuator, the method comprising: generating a first control input when a deviation of a controlled variable from a target value exists; generating a second control input that increases as the deviation of the controlled variable from the target value decreases when the controlled variable varies over time; and generating the control input based on the first control input and the second control input. A control input generation method of a 19th invention is the 18th invention, characterized in that the first control input is a steady-state control input, or, when the controlled variable varies over time, the first control input is a combination of a steady-state control input and a variable control input that decreases as the deviation of the controlled variable from the target value decreases. A control input generation method of a 20th invention is the 18th invention, characterized in that the method calculates an expected number of steps N until the controlled variable matches the target value based on the deviation of the controlled variable from the target value and the operating state of the pneumatic actuator, and generating the second control input based on the expected number of steps N so that the second control input increases as the expected number of steps N decreases. A control input generation method of a 21st invention is the 18th or 19th invention, characterized in that the first control input S1 is generated based on Equation 1, the second control input S2 is generated based on Equation 2, and the control input is generated based on Equation 3. Equation 1 S1=K. １ ・sgn(e) Formula 2 S2=K2 / N α ・sgn(v) Formula 3 Yc=K １ ・sgn(e)-K2 / N α sgn(v) e: Deviation of the controlled variable from the target value v: Displacement speed of the pneumatic actuator α: Index for adjusting the timing at which the suppression force that suppresses the operation of the pneumatic actuator begins to act K １ = Gain that generates thrust K ２ = Gain that generates a restraining force N = Number of predicted steps The control input generation method of the 22nd invention is the 21st invention, in which a gain K that compensates for nonlinear friction of the pneumatic actuator based on Equation 2-1 and generates a thrust force is calculated. １ and a gain K that generates a suppression force that reduces overshooting relative to the target value based on Equation 2-2.２ The feature is to set the following equation: １ = Ka | e | > Eth K １ = Ka / (Eth·|e|) when |e|≦Eth Ka: specified control input Eth: specified displacement |e|: absolute value of deviation e Formula 2-2 K ２ =Kb |v|>Vth K ２ = 0 |v| ≦ Vth Kb: specified control input Vth: specified displacement velocity |v|: absolute value of velocity v A control input generation method according to a 23rd aspect of the present invention is characterized in that, in the 20th aspect, a time Ta required for the controlled variable to match the target value is calculated based on the deviation of the controlled variable from the target value, and an expected number of steps N = Ta / Δt is calculated by dividing the time Ta by a sampling period Δt for detecting the deviation of the controlled variable from the target value, and if the expected number of steps N is smaller than a predetermined threshold, the expected number of steps N is used as the threshold. A control input generation method according to a 24th aspect of the present invention is characterized in that, in the 23rd aspect, the arrival time Ta is calculated based on Equation 4: Equation 4 (1 / 2) a Ta ２+v・Ta-e=0 a: displacement acceleration of the pneumatic actuator v: displacement velocity of the pneumatic actuator Ta: arrival time (arrival time - current time) e: deviation (target position - current position) A control input generation method of a 25th aspect of the present invention is the 20th aspect, wherein the second control input generation function calculates the predicted number of steps N based on the following formula 5: Formula 5 N=e / ΔL=e / (v・Δt) e: deviation of the controlled variable from the target value v: displacement velocity of the pneumatic actuator ΔL: amount of change in displacement from one sampling time before. A control input generation method of a 26th aspect of the present invention is the 18th aspect, wherein the target value is the target position of an object operated by the pneumatic actuator, the drive velocity of the pneumatic actuator, or the pressure generated by the pneumatic actuator. A control input generation program of a 27th aspect of the present invention is a program that causes a control device that controls the operation of a pneumatic actuator to execute any of the control input generation methods of the 18th to 25th aspects. A storage medium storing the control input generation program of the 28th aspect is the storage medium storing the control input generation program of the 27th aspect.

[0014] <Control Input Generator> According to the first to ninth aspects of the present invention, using the generated control input to control a pneumatic actuator can improve the responsiveness of the pneumatic actuator near the target value and suppress overshooting of the target value, thereby increasing the accuracy of operating the pneumatic actuator to the target value. Moreover, overshooting of the target value can be prevented. <Controller> According to the tenth and eleventh aspects of the present invention, using the generated control input to control a pneumatic actuator can improve the responsiveness of the pneumatic actuator near the target value and suppress overshooting of the target value, thereby increasing the accuracy of operating the pneumatic actuator to the target value. Moreover, overshooting of the target value can be prevented. According to the twelfth aspect of the present invention, even when there is load fluctuation, the accuracy of operating the pneumatic actuator to the target value can be further increased. According to the thirteenth aspect of the present invention, the pneumatic actuator can be moved to match the target trajectory. <Actuator> According to the fourteenth aspect of the present invention, the responsiveness is high near the target value and overshooting of the target value is suppressed, thereby operating the pneumatic actuator to the target value with high accuracy. <Calculation circuit> According to the fifteenth aspect of the present invention, simply by installing the control circuit in an existing pneumatic actuator control device, the responsiveness of the pneumatic actuator near the target value can be improved and overshooting of the target value can be suppressed, thereby increasing the accuracy of operating the pneumatic actuator to the target value. <Control board> According to the sixteenth and seventeenth aspects of the present invention, simply by installing the control circuit in an existing pneumatic actuator control device, the responsiveness of the pneumatic actuator near the target value can be improved and overshooting of the target value can be suppressed, thereby increasing the accuracy of operating the pneumatic actuator to the target value. <Control input generation device> According to the eighteenth to twenty-sixth aspects of the present invention, using the generated control input to control the pneumatic actuator can improve the responsiveness of the pneumatic actuator near the target value and suppress overshooting of the target value, thereby increasing the accuracy of operating the pneumatic actuator to the target value. Moreover, overshooting of the target value can be prevented.<Control input generation program, storage medium storing the control input generation program> According to the twenty-seventh and twenty-eighth aspects, by installing the program in a control device that controls a pneumatic actuator and executing the program, it is possible to improve the responsiveness of the pneumatic actuator near the target value and suppress overshooting of the target value, thereby increasing the accuracy with which the pneumatic actuator is operated to achieve the target value. Moreover, it is possible to prevent overshooting of the target value.

[0015] 1 is a block diagram of a control system 1 having a control device 10 of this embodiment. It is an explanatory diagram of the operation of a pneumatic cylinder 2 controlled by the control device 10 of this embodiment. It is a diagram showing an example of a position control system having the control device 10 of this embodiment. It is a table showing the relationship between the compensation input Yc, the deviation e, and the velocity v. It is an explanatory diagram of the state change of the compensation input Yc until the object M comes to rest at the position of the target value Lr. It is an explanatory diagram of the flow until the object M comes to rest at the position of the target value Lr. It is a table showing the relationship between the control amount and the parameters. It is a diagram showing an example of a conventional arithmetic circuit of this embodiment. It is a schematic explanatory diagram of a conventional positioning control system to which a disturbance observer is added. (A) is a photograph of the experimental apparatus, and (B) is a circuit diagram of the experimental apparatus. It is a diagram showing the experimental results of Comparative Example 1, where (A) is a graph of the response for 3 to 6 seconds, and (B) is a graph of the response in the first step (3 to 6 seconds) of the experimental results of (A), with the scale of the position deviation enlarged. 12A and 12B are diagrams showing the experimental results of Example 1, where (A) is a graph of the response for 3 to 6 seconds, and (B) is a graph of the response in the first step (3 to 6 seconds) of the experimental results of (A), with the scale of the position deviation enlarged.

[0034] Fig. 12A is a diagram showing the experimental results of Example 1, enlarging the range of 10.0 s to 10.2 s in Fig. 12A to show the change in the target generated force relative to the deviation, and (B) shows the change in the control input relative to the deviation in the same time period as (A).

[0035] Fig. 12B is a diagram showing the experimental results of Example 2. (A) is a photograph of the experimental apparatus of Example 3, and (B) is a diagram showing the experimental results of Example 3.

[0036] Fig. 12B is a schematic diagram of the control board B of this embodiment. (A) is a photograph of the experimental apparatus of Example 4. (A) is a schematic diagram of the experimental apparatus of Example 4, and (B) is a graph showing the state of load fluctuation. (A) is a diagram showing the experimental results of the experiment of Example 4(1). (B) is a diagram showing the experimental results of the experiment of Example 4(2). (B) is a diagram showing the experimental results of the experiment of Example 4(3). 10A and 10B are diagrams showing an example of an experiment for setting a constant, and are diagrams showing experimental results when the sensor for detecting the table position in Example 1 is changed, in which (A) is a graph of the response from 0 to 24 seconds, and (B) is a graph of the response in the first step (0 to 4 seconds) of the experimental results of (A), showing the position deviation scale enlarged.10A and 10B are diagrams showing experimental results when the sensor for detecting the table position is changed in Example 1, where (A) is a graph of the response from 0 to 24 seconds, and (B) is a graph of the response in the first step (0 to 4 seconds) of the experimental results of (A), with the scale of the position deviation enlarged. Also, FIG. 10A is a diagram showing the experimental results when the sensor for detecting the table position is changed in Example 2. Also, FIG. 10B is a diagram showing the relationship between the gain K1 and the deviation e, and FIG. 10C is a diagram showing the relationship between the gain K2 and the displacement velocity v.

[0016] Next, an embodiment of the present invention will be described with reference to the drawings. The control input generation device of this embodiment is a device that generates a control input used to control a pneumatic actuator, and is characterized by being able to generate a control input that can improve control accuracy. Furthermore, the control input generation method of this embodiment is a method that generates a control input used to control a pneumatic actuator, and is characterized by being able to generate a control input that can improve control accuracy.

[0017] <Actuator> The actuator (control target) controlled by the control input generating device of this embodiment is not particularly limited as long as it is operated by air pressure. For example, pneumatic actuators such as pneumatic cylinders and rotary actuators operated by air pressure can be cited as actuators controlled by the control input generating device of this embodiment. Note that in this specification, the term "displacement velocity" refers to the movement velocity of the cylinder rod in the case of a pneumatic cylinder, and the angular velocity of the rotation shaft in the case of a rotary actuator, and the term "displacement acceleration" refers to the acceleration of the cylinder rod in the case of a pneumatic cylinder, and the angular acceleration of the rotation shaft in the case of a rotary actuator.

[0018] <Devices Using Actuators> There is no particular limitation on the machines or devices in which the control input generation device of this embodiment can be used. In other words, there is no particular limitation on the machines or devices having actuators controlled by the control input generation device of this embodiment. For example, a robot arm, a single-axis conveying device, a multi-axis XY table, etc., operated by the above-mentioned actuators, etc., correspond to machines or devices in which the control input generation device of this embodiment can be used.

[0019] <Target Value> When an actuator is controlled by the control input generating device of this embodiment, a target value is set to cause the actuator to perform a predetermined operation, but this target value is not particularly limited. For example, when an object is moved by an actuator, the target value can be a target position when the object is moved or kept stationary, or a target speed of the object when the object is moved. Furthermore, the force applied to the object by the actuator when the object is moved or kept stationary can also be a target value.

[0020] In the following description, a case where an object is moved by a pneumatic cylinder and the target value is the position of the object will be described as a representative example. The following description also describes a case where an actuator is controlled by adding a control input generated by the control input generating device of this embodiment as a compensation input to a general control device that controls the actuator. That is, the description describes a case where an actuator is controlled by a control device that has a general control device as a main control input generating unit and the control input generating device of this embodiment as a compensation control input generating unit. Note that this control device corresponds to the control device of claim 11.

[0021] <Control device 10 of this embodiment> Next, the control device of this embodiment will be described. Fig. 1 shows a control system 1 that operates an object by controlling a pneumatic cylinder with the control device 10 of this embodiment, and reference numeral 2 indicates a pneumatic cylinder whose operation is controlled by the control device 10 of this embodiment. In the control system 1, an object M is fixed to a rod 2r of a pneumatic cylinder 2, and when the rod 2r moves (i.e., when the pneumatic cylinder 2 extends or retracts), the object M moves together with the rod 2r.

[0022] 1, in the pneumatic cylinder 2, the inside of the cylinder 2a is airtightly separated into a first chamber h1 and a second chamber h2 by a piston 2p, and the first chamber h1 and the second chamber h2 in the cylinder 2a are each connected to an air pressure source 4 via a pipe 3 and a servo valve 5. A control device 10 of this embodiment (hereinafter simply referred to as the control device 10) is electrically connected to the servo valve 5.

[0023] The control device 10 controls the operation of the servo valve 5 to control the operation of the pneumatic cylinder 2, and has a function of receiving information regarding target values, i.e., the position and amount of movement of the object M (or the position at which the object M is to be stationary), from an input device or the like. The control device 10 receives signals from pressure sensors p1 and p2 that measure the pressure in the first chamber h1 and the second chamber h2 of the pneumatic cylinder 2. Therefore, the control device 10 can detect the force and direction generated by the pneumatic cylinder 2 due to the differential pressure between the pressure sensors p1 and p2. The control device 10 also receives a signal from a position sensor d that measures the position of the object M, i.e., the distance of the object M from a predetermined position. Therefore, the control device 10 can detect the current position of the object M, which is operated by the pneumatic cylinder 2, based on the distance detected by the position sensor d.

[0024] The control device 10 has a function of generating a signal (control input) to control the operation of the pneumatic cylinder 2, i.e., a control input to be supplied to the servo valve 5, based on a target value input from an input means, the current position of the object M determined based on signals from each sensor, and the force and direction generated by the pneumatic cylinder 2. Specifically, the control device 10 includes a main control input generation unit 11, which is a general control device that controls the operation of the servo valve 5, and a compensation control input generation unit 15. The main control input generation unit 11 has a function of generating a main input, which is a control input, based on the input signal using a known control method, for example, linear control theory such as PID control, or nonlinear control theory such as fuzzy logic or neural networks. On the other hand, the compensation control input generation unit 15 has a function of generating a compensation input Yc that corrects the main input generated by the main control input generation unit 11, as will be described in detail below.

[0025] With the above-described configuration, when a target value is input to the control device 10 from an input means or the like, the pneumatic cylinder 2 is controlled as follows.

[0026] 2A, if there is a difference (deviation e) between the current position (control amount L) of the object M and the target value Lr when no external force is being applied to the object M, the control device 10 recognizes the deviation e based on a signal from the position sensor d, and supplies a signal to the servo valve 5 to operate the pneumatic cylinder 2 so that the object M moves toward the target value Lr. The servo valve 5 then adjusts the flow rate of pressurized air supplied to the pneumatic cylinder 2, causing the pneumatic cylinder 2 to extend.

[0027] 2(B), as the extension of the pneumatic cylinder 2 brings the object M closer to the target value Lr, the deviation e becomes smaller, so the control device 10 supplies a signal to the servo valve 5 to operate the pneumatic cylinder 2 so as to reduce the force generated by the pneumatic cylinder 2, i.e., the force that moves the object M toward the target value Lr. Then, the servo valve 5 adjusts the flow rate of pressurized air supplied to the pneumatic cylinder 2, reducing the force generated by the pneumatic cylinder 2 and slowing down the extension speed of the pneumatic cylinder 2.

[0028] 2(C), when the object M reaches the target value Lr (when the deviation e becomes 0), the control device 10 supplies a signal to the servo valve 5 to stop the operation of the pneumatic cylinder 2. Then, the servo valve 5 stops the supply of pressurized air to the pneumatic cylinder 2, and the operation of the pneumatic cylinder 2 stops, so that the object M can be stopped at the target value Lr.

[0029] 2(D), if an external force FR is being applied to the object M, stopping the operation of the pneumatic cylinder 2 when the object M has reached the target value Lr may cause the object M to move from the target value Lr due to the external force FR. Even in this case, when the control device 10 detects the deviation e, it supplies a signal to the servo valve 5 to operate the pneumatic cylinder 2 so as to reduce the deviation e, so that the object M can be kept stationary at the target value Lr.

[0030] <Compensation Control Input Generator 15> In the case of the pneumatic cylinder 2, the effects of air compressibility and friction between the object M and other objects greatly affect the control accuracy (i.e., the difference between the stopping position of the object M and the target value Lr). For this reason, when improving the control accuracy, the main control input generator 11 is generally configured as a control system using a disturbance observer. However, even in a control system using a disturbance observer, it is difficult to increase the control accuracy above a certain level if the main control input generator 11 generates a control input using a general control method. This is for the following reasons.

[0031] In the positioning control system shown in Fig. 9, a controller (having the same function as the main control input generating unit 11) controls the pneumatic cylinder 2 by a general control method (for example, PD control). The positioning control system shown in Fig. 9 includes a generated force control system, and by applying a disturbance observer to each of them, the influence of parameter fluctuations due to load changes, nonlinearity of the pressure response unit, the influence of frictional force, etc. is compensated for. Note that the disturbance observer in Fig. 9 compensates for the influence of the generated force control object P ｆ and the control object P ｋSince the influence of fluctuations from the nominal model is also estimated as a disturbance, in order to ensure the stability of the control system, the disturbance is filtered through a low-pass filter Q to obtain the estimated disturbance D ｆｅ , D ｐｅ These are input to the target generated force Fr and the control input U. The parameters in FIG. 9 indicate the following: F ｒ : Target generated force U: Control input P ｆ : Control target of generated force P ｋ : Control object of movement P ｆｎ －１ : Inverse nominal model of the controlled object of generated force P ｋｎ －１ : Inverse nominal model of the controlled object of motion Q ｆ :D ｐ Low-pass filter for the estimated value of Q ｋ :D ｆ Low-pass filter for the estimated value of D ｆ : Frictional force and external force D ｐ : Disturbance to pressure response section D ｆｅ :D ｆ Estimated value of D ｐｅ :D ｐ Estimated value of

[0032] When the operation of the pneumatic cylinder 2 is controlled by such a positioning control system, even if the object M moves to the target value Lr, it will not stop at the target value Lr but will overshoot, resulting in a typical stick-slip phenomenon. This phenomenon occurs because although the generated force F can be made to follow the target generated force Fr, the target generated force Fr does not change even when the deviation (in the case of Figure 2(B) , the deviation e between the target value Lr and the current position of the object M) is reduced. When the deviation e becomes extremely small, the braking action due to the differential action of the PD control weakens, and in addition, the control system is type 1 and has integral characteristics, so the response to fluctuations in the deviation e is slow. As a measure to reduce this overshoot, the differential gain K ｄ However, in this case, the stability of the transient state is significantly reduced.

[0033] 1 and 3, the control device 10 of this embodiment configures a positioning control system including a compensation control input generation unit 15 that generates a compensation input Yc that corrects the main input Ms generated by the main control input generation unit 11. That is, in the control device 10 of this embodiment, the compensation input Yc generated by the compensation control input generation unit 15 is added to the main input Ms (Fr obtained by adding an estimated disturbance to the main input Ms in FIG. 3) as a control input, and by inputting a control input including the compensation input Yc, the stick-slip phenomenon of the pneumatic cylinder 2 is suppressed, thereby improving the accuracy of the positioning control of the object M. For example, in the case of using only the control input generated by the main control input generation unit 11 of the control device 10 of this embodiment (the main input Ms or the input obtained by adding an estimated disturbance to the main input Ms), that is, in control by the control system of FIG. 9, the accuracy of the positioning control for placing the object M at a target position is at most about 10 μm. However, in the control device 10 of this embodiment, by adding the compensation input Yc generated by the compensation control input generation unit 15 to the main input Ms (or the input obtained by adding an estimated disturbance to the main input Ms), the positioning control accuracy for placing the object M at a target position can be improved to an accuracy equal to or higher than the detection accuracy of the position sensor that detects the position of the object M. For example, if the detection accuracy of the position sensor that detects the position of the object M is about 50 nm, a positioning accuracy of about 50 nm can be obtained, and if the detection accuracy of the position sensor is about 1 nm, a positioning accuracy of about 1 nm can be obtained.

[0034] The compensation control input generation unit 15 that generates the compensation input Yc has a first control input generation function 16, a second control input generation function 17, and an output generation function 18, as shown in FIG. 1, and has the function of generating the compensation input Yc by receiving the target value Lr, the controlled variable L, and the piston speed v as inputs.

[0035] Specifically, it has a function of generating a compensation input Yc (Yc=S1−S2) using a first control input S1 generated by the first control input generation function 16 and a second control input S2 generated by the second control input generation function 17. The first control input S1 is an input that promotes movement of the pneumatic cylinder 2 toward the target value Lr (a so-called input that causes the pneumatic cylinder 2 to exert an accelerator function), and the second control input S2 is an input that suppresses movement of the pneumatic cylinder 2 toward the target value Lr (a so-called input that causes the pneumatic cylinder 2 to exert a brake function), and by appropriately generating these first control input S1 and second control input S2, the positioning control accuracy for placing the object M at a target position using the compensation input Yc can be improved to the same level as the detection accuracy of the position sensor.

[0036] <First control input generation function 16> The first control input generation function 16 has a function of generating a first control input S1 when there is a deviation e between the target value Lr and the controlled variable L. Specifically, the first control input generation function 16 has a function of generating a steady-state control input as the first control input S1. For example, the first control input generation function 16 generates the first control input S1 based on the following formula 1. Formula 1 S1=K １ sgn(e) e: deviation between target value Lr and controlled variable L K １ = Gain that compensates for the nonlinear friction of the pneumatic cylinder 2 and generates a thrust force. Note that sgn(e) is a function that outputs 1 when e>0, -1 when e<0, and 0 when e=0.

[0037] <Second Control Input Generation Function 17> The second control input generation function 17 has a function of generating a second control input S2 that increases as the deviation e between the target value Lr and the controlled variable L decreases when there is a change in the controlled variable L over time. Specifically, the second control input generation function 17 has an expected step number calculation function 17b that calculates an expected number of steps N until the controlled variable L matches the target value Lr, based on the deviation e between the target value Lr and the controlled variable L and the piston speed v. The second control input generation function 17 has a function of generating the second control input S2 based on the expected number of steps N calculated by the expected step number calculation function 17b, so that the second control input S2 increases as the expected number of steps N decreases.

[0038] For example, the second control input generating function 17 can generate the second control input S2 based on the following formula 2. Note that the displacement speed v of the pneumatic cylinder 2 is the same as the movement speed v of the piston in FIG. 3. Formula 2 S2=K2 / N α sgn(v) e: deviation of the controlled variable from the target value v: operating speed of the pneumatic cylinder 2 α: index for adjusting the timing at which the suppressing force that suppresses the operation of the pneumatic cylinder 2 starts to act K ２ = Gain that generates a suppression force that reduces overshooting of the target value Lr. Note that sgn(v) is a function that outputs 1 when v>0, −1 when v<0, and 0 when v=0.

[0039] <Third control input generating function 18> The third control input generating function 18 has a function of generating a compensation input Yc based on the first control input S1 generated by the first control input generating function 16 and the second control input S2 generated by the second control input generating function 17. The third control input generating function 18 corresponds to the output generating function recited in claim 1 of the claims.

[0040] For example, the compensation control input generating unit 15 can generate the compensation input Yc based on the following equation 3: Yc=K １ ・sgn(e)-K2 / N αsgn(v) e: deviation between target value Lr and controlled variable L v: displacement speed of pneumatic cylinder 2 α: index for adjusting the timing at which the suppression force that suppresses the operation of pneumatic cylinder 2 starts to act K １ = Gain K that compensates for the nonlinear friction of the pneumatic cylinder 2 and generates thrust ２ = Gain that generates a suppression force that reduces overshooting of the target value Lr

[0041] As described above, the third control input generation function 18 generates the compensation input Yc by combining the first control input S1 generated by the first control input generation function 16 and the second control input S2 generated by the second control input generation function 17, thereby obtaining the following advantages.

[0042] First, when the first control input generating function 16 generates the first control input S1 according to the above formula 1, the first control input S1 undergoes on-off control, changing in three stages. Then, when a deviation e occurs between the target value Lr and the controlled variable L, the compensation signal Yc contains a steady-state value with the same sign as the deviation e as the first control input S1, and when the deviation e is 0, the first control input S1 becomes 0. Therefore, the compensation signal Yc, which contains a constant first control input S1 regardless of the magnitude of the deviation e, is added to the main input Ms as a control input, thereby improving response speed in transient states. For example, friction occurring in the piston of the pneumatic cylinder 2 or the object M near the target value Lr can be compensated for, thereby improving response speed near the target value Lr.

[0043] Furthermore, when the second control input generating function 17 generates the second control input S2 according to the above formula 2, if a velocity is occurring, that is, if the piston of the pneumatic cylinder 2 is moving, the compensation signal Yc will be the first control input S1 minus the second control input S2, which is inversely proportional to the predicted number of steps N, and if the velocity is 0, the second control input S2 will be 0, so the first control input S1 will be the compensation signal Yc. Therefore, the effect of suppressing (applying the brakes to) the operation of the pneumatic cylinder 2 near the target value Lr is increased, making it possible to suppress the occurrence of the stick-slip phenomenon and achieving highly accurate positioning at the target value Lr.

[0044] As described above, in the control device 10 of this embodiment, the compensation input Yc generated by the compensation control input generation unit 15 is added to the main input Ms generated by the main control input generation unit 11 to obtain a control input, thereby improving the responsiveness of the pneumatic cylinder 2 in a transient state and achieving high-precision positioning at the target value Lr.

[0045] <State of Compensation Input Yc> The flow until the object M comes to rest at the position of the target value Lr (until the control converges) will be described below with reference to FIGS. 5 and 6. FIG.

[0046] The compensation signal Yc is as shown in the table of Fig. 4, depending on the sign of the deviation e between the target value Lr and the controlled variable L, and the displacement speed (piston speed v) of the pneumatic cylinder 2. Therefore, when the pneumatic cylinder 2 starts to operate from a non-operating state (for example, a state where e>0 and v=0), the compensation signal Yc=K １ As the pneumatic cylinder 2 moves toward the target value Lr (forward movement), the deviation e becomes smaller and the second control input S2 (K2 / N α ) causes the speed of the pneumatic cylinder 2 to decrease, and eventually the operation of the pneumatic cylinder 2 stops (v=0).

[0047] At this time, if the control amount L of the pneumatic cylinder 2 matches the target value Lr (e=0), the pneumatic cylinder 2 remains stationary. On the other hand, if there is a deviation e even when the operation of the pneumatic cylinder 2 is stopped (if it stops before the target value Lr, e>0), the first control input S1(K １ ) causes the air cylinder 2 to start operating and move toward the target value Lr.

[0048] When moving towards the target value Lr, the second control input S2 (K2 / N α ) causes the operation of the pneumatic cylinder 2 to stop (v = 0). At this time, if the pneumatic cylinder 2 stops just before the target value Lr (e > 0), the state of moving in the forward direction and stopping will be repeated until the control amount L of the pneumatic cylinder 2 matches the target value Lr.

[0049] On the other hand, when the control amount L of the pneumatic cylinder 2 exceeds the target value Lr (e<0), the first control input S1(K １ ) and the second control input S2 (K2 / N α ) function as a brake and quickly stop the operation of the pneumatic cylinder 2 (v=0). In this case, since a deviation e occurs (e<0), the first control input S1 (K １ ) causes the pneumatic cylinder 2 to start operating and move toward the target value Lr. That is, the pneumatic cylinder 2 starts moving toward the target value Lr in the direction opposite to the moving direction in the initial state (reverse movement).

[0050] When the pneumatic cylinder 2 starts to operate, the second control input S2 (K2 / N α ), the speed of the pneumatic cylinder 2 decreases, and eventually the operation of the pneumatic cylinder 2 stops (v = 0). At this time, if the control amount L of the pneumatic cylinder 2 matches the target value Lr (e = 0), the pneumatic cylinder 2 remains stationary. On the other hand, if the pneumatic cylinder 2 stops just before the target value Lr (e < 0), the state of reverse movement and stopping is repeated until the control amount L of the pneumatic cylinder 2 matches the target value Lr.

[0051] In addition, even in the reverse direction movement, the control amount L of the pneumatic cylinder 2 may become smaller than the target value Lr (e>0). In this case, as in the case of overshoot, the first control input S1 (K １ ) and the second control input S2 (K2 / N α ) function as a brake and quickly stop the operation of the pneumatic cylinder 2 (v=0). In this case, since a deviation e occurs (e>0), the first control input S1 (K １ ) causes the air cylinder 2 to start operating, and the state of moving in the forward direction and stopping is repeated.

[0052] As described above, the compensation input Yc causes the pneumatic cylinder 2 to repeatedly move toward the target value Lr and stop, thereby making it possible to make the controlled variable L coincide with the target value Lr. １) is constant, the air cylinder 2 can be restarted at high speed from a stopped state even if the deviation e is small. α ) becomes larger, the occurrence of the stick-slip phenomenon can be suppressed, and positioning to the target value Lr can be performed with high accuracy.

[0053] <Regarding the First Control Input Generating Function 16> In the above description, the first control input generating function 16 generates a steady-state control input as the first control input S1. However, the first control input generating function 16 may have a function to generate, as the first control input S1, a first control input S1 that is a combination of the steady-state control input as the first control input S1 and a variable control input that decreases as the deviation e of the controlled variable L from the target value Lr decreases. For example, in the early stage of a transient response in which the deviation is relatively large, advantages such as improved response speed can be obtained by using the first control input S1 that includes the variable control input.

[0054] <Gain K of first control input S1 １ and the gain K of the second control input S2 ２ About the gain K of the first control input S1 １ and the gain K of the second control input S2 ２ may be the same value regardless of the movement direction of the pneumatic cylinder 2, or may be different depending on the movement direction of the pneumatic cylinder 2. When the pneumatic cylinder 2 is a single-rod cylinder, the pressure-receiving area of ​​the piston differs on both sides of the piston due to the influence of the rod. Therefore, in the case of a single-rod cylinder, a pushing action of the piston (movement opposite to the rod side) is more likely to generate force than a pulling action of the piston (movement toward the rod side). In other words, a pushing action is more likely to move the piston than a pulling action. Therefore, when the pneumatic cylinder 2 is a single-rod cylinder, the gain K of the first control input S1 is set depending on the movement direction of the piston as follows: １ and the gain K of the second control input S2 ２ It is also possible to change the １ =K １ a(e>0) K １ =K １ b(e<0) K ２ =K２ a(v>0) K ２ =K ２ b(v<0) where K １ a<K １ b, K ２ a<K ２ b

[0055] Furthermore, when a force is applied in one direction to the pneumatic cylinder 2, it is also affected by that force. As a result, even when the same amount of movement or speed of movement is to be achieved depending on the direction of movement of the piston, the required force, in other words, the differential pressure generated between the first chamber h1 and the second chamber h2 of the pneumatic cylinder 2, will differ. Therefore, when an external force is applied to the pneumatic cylinder 2 in the direction of the pushing operation, the gain K of the first control input S1 will also differ in the same way as above. １ and the gain K of the second control input S2 ２ On the other hand, when an external force is applied to the pneumatic cylinder 2 in the direction of the pulling motion, K １ a>K １ b, K ２ a>K ２ b, the gain K of the first control input S1 is １ and the gain K of the second control input S2 ２ may be set.

[0056] In addition, the gain K of the first control input S1 １ and the gain K of the second control input S2 ２ may be a fixed value (constant), or may be varied depending on the actual deviation or displacement speed. １ is set as in Equation 2-1, a predetermined input that promotes movement toward the target value can be generated until the deviation e reaches the required positioning accuracy (required deviation), and when the absolute value |e| of the deviation e becomes smaller than the required deviation, an input that attenuates the movement can be generated, thereby further improving the responsiveness near the target value. ２ is set as in Equation 2-2, and when the displacement speed becomes smaller than the required displacement speed (required displacement speed), the gain K ２If K is set to 0, it is possible to prevent the second control input S2 from becoming unstable in the extremely low speed range where the influence of noise becomes large. １ = Ka | e | > Eth K １ = Ka / (Eth·|e|) When |e|≦Eth Ka: Minimum control input that causes displacement Eth: Required positioning accuracy (required deviation) (deviation threshold) |e|: Absolute value of deviation e Formula 2-2 K ２ =Kb |v|>Vth K ２ = 0 |v| ≦ Vth Kb: Minimum control input at which displacement occurs Vth: Threshold value of displacement speed taking into account the influence of observation noise |v|: Absolute value of speed v

[0057] As shown in Equation 2-1, the gain K of the first control input S1 １ When set, the gain K １ The first control input S1 and the deviation e have the relationship shown in FIG. 26(A). Also, as shown in Equation 2-2, the gain K of the second control input S2 ２ When set, the gain K ２ The second control input S2 and the displacement velocity v have a relationship as shown in FIG. 26(B).

[0058] In the present invention, the compensation input Yc is generated from the first control input S1 and the second control input S2 as described above, based on the following technical idea.

[0059] First, the gain K of the first control input S1 is calculated using the above-mentioned formulas 2-1 and 2-2. １ and the gain K of the second control input S2 ２ When the compensation input Yc is generated based on Equation 3, the gain K of the first control input S1, which is the first term, is set as １is designed to have a gain that exceeds the maximum static friction force. Therefore, the first control input S1 is designed to be a control input that moves the controlled object so that its displacement does not stop under any conditions or environments until the deviation reaches a predetermined threshold (until the absolute value of the deviation becomes smaller than the predetermined threshold). Even if the displacement of the controlled object is about to stop during the process of the absolute value of the deviation of the controlled object reaching the predetermined threshold, the second control input S2 of the second term can be set to a value close to zero, thereby allowing the controlled object to continue moving so that the absolute value of the deviation reaches the predetermined threshold. In other words, when the absolute value of the deviation approaches the predetermined threshold and the controlled object is about to stop, the predicted number of steps N increases (in other words, the change in displacement ΔL from one sampling time before becomes small), and therefore the second control input S2 of the second term, which applies a braking force to the movement of the controlled object in the compensation input Yc, becomes a value close to zero. In this case, the compensation input Yc essentially consists of only the first term, allowing the controlled object to overcome the frictional force and continue to operate. On the other hand, when the deviation falls within a predetermined threshold (the absolute value of the deviation becomes smaller than the predetermined threshold) and reaches the target value (e = 0), if the speed (absolute value of the speed) at that time exceeds the predetermined threshold, the second control input S2 is designed to generate a braking force that stops the movement of the controlled object, and an operation to stop the displacement is performed at that point. By designing the first control input S1 and the second control input S2 as described above, when the deviation is within the predetermined threshold (when the absolute value of the deviation is smaller than the predetermined threshold), the function of the first control input S1, which serves as the first term, can quickly return the deviation to within the threshold even if it falls outside the threshold (when the absolute value of the deviation becomes larger than the predetermined threshold). Since the main controller generally has an integrator, a signal obtained by integrating the deviation is output in an attempt to make it zero unless the deviation is zero. As a result, it is possible for the deviation to deviate from the predetermined threshold, but the effect of the first control input S1, the first term described above, quickly returns the deviation to within the threshold.Furthermore, by designing the first control input S1 and the second control input S2 as described above, even if some external force acts and the speed (absolute value of the speed) of the controlled object exceeds a predetermined threshold value, causing the deviation to fall outside the predetermined threshold value (the absolute value of the deviation becomes larger than the predetermined threshold value), the function of the second control input S2, which is the second term, can apply a large braking force to the controlled object to prevent the deviation from deviating significantly from the threshold value, and as a result, the effect of the first control input S1, which is the first term, works to quickly return the deviation to within the threshold value.

[0060] <Regarding the Estimated Step Number Calculation Function 17b> The method by which the estimated step number calculation function 17b of the second control input generation function 17 calculates the estimated step number N is not particularly limited. For example, the estimated step number N may be the time required for the controlled variable L to match the target value Lr based on a sampling period Δt (e.g., a period during which a position sensor detects the controlled variable L) at which the deviation e between the controlled variable L and the target value Lr is detected. That is, as shown in FIG. 1 , the second control input generation function 17 may be provided with an arrival time calculation function 17a that calculates the time required to reach the target position from the current position (arrival time Ta), and the estimated step number N may be calculated by dividing the arrival time Ta calculated by the arrival time calculation function 17a by the sampling period Δt (i.e., N = Ta / Δt). Using this method, the estimated step number calculation function 17b may calculate the estimated step number N, which has the advantage of facilitating calculation of the estimated step number N. Note that N = Ta / Δt may also be calculated using the following equation 5. Note that v·Δt is the same as the movement amount ΔL per sampling period (in other words, the amount of change in displacement from one sampling time before). Equation 5 N(e, v)=e / ΔL=e / (v·Δt) e: deviation between target value Lr and controlled variable L v: displacement speed of pneumatic cylinder 2 ΔL: amount of change in displacement from one sampling time before

[0061] Furthermore, N in equation 5 may be calculated using the following equation 5-1: Equation 5-1 N(Lr, L)=(Lr(k)-L(k)) / (L(k)-L(k-1)) Lr(k): target position L(k): current position L(k-1): current position one step before

[0062] Furthermore, the predicted step number calculation function 17b may also have a function of setting the predicted step number N to the threshold value X when the predicted step number N is smaller than a predetermined threshold value X. For example, when the threshold value X is set to 1, the predicted step number calculation function 17b may also have a function of setting the predicted step number N to 1 when the predicted step number N is smaller than 1.

[0063] The reason for providing such a function is as follows. For example, when the second control input S2 is the gain (K2 / N α ), if the piston speed v is too large with respect to the deviation e between the target value Lr and the controlled variable L, the second control input S2 may become very large and diverge to infinity, potentially causing the control system to go out of control. Therefore, in order to limit the value of the second control input S2 and prevent the control system from going out of control, it is desirable to set the minimum value of the expected number of steps N as the threshold value X.

[0064] Note that, in the above example, a case has been described in which the second control input S2 is generated using Equation 2, but the method by which the expected number of steps calculation function 17b generates the second control input S2, i.e., the equation for generating the second control input S2 (second control input generation equation), is not limited to Equation 2. It is sufficient that the second control input S2 be generated so that it increases as the expected number of steps N decreases. In other words, it is sufficient that the second control input S2 be generated so that the operation of the actuator can be appropriately converged by the compensation input Yc generated based on the first control input S1 and the second control input S2. In other words, it is sufficient to adopt a second control input generation equation that generates the second control input S2 that can cause the convergence of the operation of the actuator to a desired state.

[0065] <Method of calculating arrival time Ta> In the second control input generating function 17, the time required for the controlled variable L to match the target value Lr, that is, the method of calculating the arrival time Ta by the arrival time calculation function 17a is not particularly limited. For example, in the following equation 4, a and v are known values, and the arrival time Ta can be calculated using, for example, Newton's method. Equation 4 (1 / 2) a Ta ２+v Ta - e = 0 a: Displacement acceleration of pneumatic cylinder 2 v: Displacement speed of pneumatic cylinder 2 Ta: Arrival time (arrival time - current time) e: Deviation (target position - current position)

[0066] <Regarding the Third Control Input Generation Function 18> In the above example, a case has been described in which the third control input generation function 18 generates the compensation input Yc based on Equation 3, but the equation (compensation input generation equation) by which the third control input generation function 18 generates the compensation input Yc is not particularly limited. Any compensation input generation equation can be adopted that can generate a compensation input Yc that can appropriately converge the operation of the actuator based on the first control input S1 and the second control input S2. In other words, it is sufficient to adopt a compensation input generation equation that can generate a compensation input Yc that can bring the convergence of the operation of the actuator to a desired state based on the first control input S1 and the second control input S2. <When the control input generation device of this embodiment is used as a control device>

[0067] In the above example, the case where the control input generating device of this embodiment is included in the control device as a compensation input generating device has been described, but the operation of the pneumatic cylinder 2 may be controlled by the control input generating device of this embodiment alone. In other words, the control input generating device of this embodiment can be used as a control device that controls the operation of the pneumatic cylinder 2. In this case, the compensation input generated by the control input generating device of this embodiment is input as a control input that controls the operation of the pneumatic cylinder 2.

[0068] <Arithmetic Circuit> The control input generation device of this embodiment may be provided in the form of an arithmetic circuit. That is, an arithmetic circuit may be manufactured that has input terminals to which a target value and a current value are input as input signals, and an output terminal that outputs, as an output signal, a signal equivalent to the control input generated by the control input generation device of this embodiment described above, and the control input generation device of this embodiment may be provided in the form of this arithmetic circuit (the arithmetic circuit of this embodiment).

[0069] In other words, a circuit having the following components can be formed and provided as the arithmetic circuit of this embodiment: a first control input generator that generates a first control input when a deviation of the control variable from the target value exists when the target value and the control variable are input as input signals; a second control input generator that generates a second control input that increases as the deviation of the control variable from the target value decreases when the control variable changes over time; and an output generator that generates and outputs a control input based on the first control input and the second control input. For example, a circuit configured as shown in FIG. 8 can be used to provide the arithmetic circuit of this embodiment having the above-described functions, and therefore a circuit having such a configuration can be provided as the arithmetic circuit of this embodiment.

[0070] When the control input generating device of this embodiment is provided in the form of an arithmetic circuit, the existing control device can be used as is, and the control input generated by the control input generating device of this embodiment can be incorporated into the existing control device so that it becomes a compensation input, making it possible for the existing control device of an existing actuator to perform functions equivalent to those of the control device of this embodiment.

[0071] As described above, the calculation circuit may supply the compensation input to the existing control device, or the control input of the existing control device may also be input to the calculation circuit, and the control input of the existing control device and the compensation input (the control input generated by the control input generation device of this embodiment) may be combined within the calculation circuit to output the control input.

[0072] <Control Board> The control input generation device of this embodiment may be provided in the form of a control board. That is, a control board may be manufactured and provided, having input terminals to which a target value and a current value are input as input signals, and output terminals to which signals equivalent to the control inputs generated by the control input generation device of this embodiment are output as output signals.

[0073] For example, the control board B shown in FIG. 16A includes an input terminal IT to which a target value and a control variable (in FIG. 16A , a signal from a pressure sensor and a signal from a position sensor) are input. The control board B is also equipped with a main controller MC that generates a control input (main control input) equivalent to that of a general control device based on the input target value and control variable, and a compensator CM (i.e., a compensator having the function of generating a compensation control input) having the same function as the compensation control input generation unit 15 described above. The control board B also includes an output generation unit D3, which is an addition circuit that combines (adds) the main control input and the compensation control input to generate a control input, and an output terminal OT that outputs the control input generated by the output generation unit D3 to the outside. By providing such a control board B, it is possible to introduce control using the control input generation method of this embodiment into the control of existing equipment by replacing an existing control device (control board) with the control board B.

[0074] 16(A), the main control input generated by the main controller MC is first output from the output terminal OT and then input from the input terminal to be supplied to the force generating unit D3, but the main control input generated by the main controller MC may be configured to be input directly to the output generating unit D3 without being output from the output terminal OT. In the configuration of FIG. 16(A), it is also possible to supply only the main control input output from the output terminal OT, that is, without combining it with the compensation control input, as the control input to the controlled object, thereby increasing the degree of freedom in controlling the controlled object.

[0075] Furthermore, when introducing control using the control input generation method of this embodiment into the control of existing equipment, the control board B may be configured to generate a control input including a compensation input using the control input output by the existing control device EM, i.e., an external control input. For example, as shown in FIG. 16(B), the control board B may be provided with an input terminal IT for inputting the control input output by the control device EM of the existing equipment, in addition to an input terminal for inputting a target value or control amount. The output generation unit D3 may then generate a control input using the control input input from this input terminal IT as the main control input. By providing such a control board B, control using the control input generation method of this embodiment can be introduced to the existing control device EM simply by connecting the control board B to the existing control device EM. <Control Input Generation Program>

[0076] In addition, in the above example, a case was described in which control using the control input generation method of this embodiment was introduced by providing the control input generation device of this embodiment in the form of a control board, but it is also possible to introduce control using the control input generation method of this embodiment using a program that realizes the control input generation method of this embodiment.

[0077] That is, it is possible to introduce control by the control input generation method of this embodiment into the controlled object by installing a program that realizes the control input generation method of this embodiment into the actuator to be controlled or into the control device (computer, etc.) of a machine or device that employs such actuator, or by providing the program via an Internet line. When installing the program into the control device, the program may be provided and installed via an Internet line, or the program may be stored in a storage medium readable by the control device and installed from the storage medium.

[0078] For example, when a target value and a controlled variable are input as input signals, a program (the control input generation program of this embodiment) that performs the following functions is installed in a control device of an existing machine or device: a first control input generation function that generates a first control input when there is a deviation of the controlled variable from the target value; a second control input generation function that generates a second control input that increases as the deviation of the controlled variable from the target value decreases when there is a time change in the controlled variable; and an output generation function that generates and outputs a control input based on the first control input and the second control input. Then, by adding the compensation input generated by the control input generation program of this embodiment to the control input (main control input) generated by the control device of the existing machine or device, it becomes possible to implement control using the control input generation method of this embodiment in the existing machine or device.

[0079] Furthermore, the control input generation program of this embodiment may have a function of generating a control input corresponding to a main control input. That is, the control input generation program of this embodiment may have a function of generating a main control input, a function of generating a compensation input, and a function of generating a control input based on the main control input and the compensation input. In this case, by installing the control input generation program of this embodiment in place of an existing control program in a control device of an existing machine or device, it becomes possible to introduce control by the control input generation method of this embodiment into the control device of the existing machine or device.

[0080] The control input generation program of the present embodiment can be stored on a storage medium such as a floppy disk, hard disk, optical disk, magneto-optical disk, CD (Compact Disk)-ROM, magnetic tape, non-volatile memory card, or ROM. The present embodiment is not limited to programs that execute processing on their own, but also includes programs that execute processing by operating on an OS (Operating System) in cooperation with other software or functions of an expansion board.

[0081] <Regarding Control Amount> In the above example, the case where the control amount controlled by the control input and compensation input generated by the control input generating device of this embodiment is the movement amount of the air cylinder 2 has been described, that is, the case where the control input generating device of this embodiment is used in a control system that performs position control. The control amount controlled by the control input and compensation input generated by the control input generating device of this embodiment is not limited to the position described above. For example, by using the following equation 6 and making each parameter correspond to the table in FIG. 7, it becomes possible to generate and control the compensation input Yc (or control input) for the displacement speed of the air cylinder 2 and the force generated in the air cylinder 2 in the same way as position control. Equation 6 Yc=K １ ・sgn(e)-K2 / N α ・sgn(v) N=e / (v・Δt)=(x ｒ (k) - x(k)) / (x(k) - x(k-1)) xr(k): target value x(k): current value x(k-1): current value one step before

[0082] Furthermore, N in Equation 6 may be calculated using the following Equation 6-1: Equation 6-1 N(xr, x)=(xr(k)-x(k)) / (x(k)-x(k-1)) xr(k): target value x(k): current value x(k-1): current value one step before

[0083] For example, if a generated force control system configured inside the position control system shown in FIG. 3 is used, the control input generating device of this embodiment can be used as a device for generating a compensation input in the control system that controls the force generated in the pneumatic cylinder 2 as described above.

[0084] <Regarding Constant Estimation> As described above, the equation by which the compensation control input generation unit 15 generates the compensation input Yc requires the setting of constants such as various gains. Such constants are usually determined by trial and error. That is, the control input generation device 1 of this embodiment is used to operate the controlled object (a device operated by an air cylinder, etc.) while changing the constants, and the operating status is confirmed. The constants are then changed so as to achieve an appropriate operating state, and the constants to be finally adopted are set.

[0085] For example, when generating the compensation input using the above-mentioned formula 3, the gain K１ , K2, power exponent α, etc., are set, the control object is operated using the generated compensation input, the operation results are evaluated, and the combination of constants that gives an appropriate evaluation is finally set as the constant.

[0086] However, when setting constants in this way, although it may be possible to find appropriate constants in a relatively short time, it is usually necessary to conduct an extremely large number of experiments before finding constants that will provide the desired operating accuracy.

[0087] Therefore, if an estimation device is created that can estimate constants using a machine learning model, appropriate constants can be estimated even with a small number of experiments, based on the constants used in the experiment and the operating results.

[0088] For example, an evaluation function is set to evaluate the operation results when the controlled object is operated, and learning data is collected that associates the constants set when the controlled object is operated, the results of evaluating the operation of the controlled object using the set constants using the evaluation function (evaluation results), and evaluations of experimental results.Then, by having a machine learning model, which is an estimation device, learn the learning data, evaluation results can be obtained simply by inputting the constants into the estimation device, without having to conduct experiments using the controlled object.

[0089] There are no particular limitations on the machine learning model used in the estimation device, but support vector machines (SVM), neural networks (NN), random forests (RF), etc. can be used.

[0090] The evaluation function for evaluating the above-mentioned operational accuracy can be appropriately selected depending on the controlled object and its control amount. For example, when the target is the position control of a pneumatic cylinder, the operational accuracy can be evaluated using the evaluation function shown in Equation 7 (see FIG. 22). In this case, the smaller the evaluation score S1, the more appropriately the position control is performed (the closer the target trajectory R and actual behavior A match). Equation 7 S1 = w1 OR + w2 PA S1: Evaluation score OR: Ratio of overshoot amount OS to target movement amount DA (OS / DA) PA: Position accuracy polarity ACC > E: 1 ACC ≤ E: 0 w1, w2: Weights The position accuracy polarity PA is set to 0 when the difference ACC between the target movement amount DA and the actual movement amount RA in the stable state is within a predetermined value E when the actual behavior A is stable, and is set to 1 when it is greater than E.

[0091] Furthermore, when an actuation command (target trajectory R) such as that shown in Fig. 22 is executed, the following formula 8 can also be used as the evaluation function. With this evaluation function, the smaller the evaluation score S2, the more appropriately the position control is being performed (the closer the target trajectory R and actual behavior A match). Formula 8 S2 = w3 OS + w4 ACC + w5 TM S1: Evaluation score OS: Amount of overshoot relative to the target movement amount DA ACC: Difference between the target movement amount DA and the actual movement amount RA in a stable state TM: Time from the start of control until movement stabilizes w3, w4, w5: Weights

[0092] <Regarding the Target Control System> By employing the control input generation device 1 of this embodiment described above, the positioning control accuracy for placing the object M at a target position can be improved to the same level as the detection accuracy of the position sensor. On the other hand, if there is a change in the external environment that affects the movement of the object M, the positioning control accuracy may decrease when responding to the change in the external environment (transient response). However, if the control input generation device 1 of this embodiment configures a type 2 control system having a disturbance observer and performs feedforward control to cancel its dynamics, the accuracy of tracking the target trajectory during transient response can be improved.

[0093] There is no particular limitation on the method of designing the control input generating device 1 of this embodiment to configure a type 2 control system having a disturbance observer. For example, in the case of the position control system shown in FIG. 3, Gf(s) / (1-Q(s)·Gf(s)) (Gf(s) is the transfer function of the force control system) is 1 / s ２ If the control system is designed to independently include the integrator (1 / s is an integrator), the control input generating device 1 of this embodiment can be made into a type 2 control system having a disturbance observer.

[0094] In this way, when the control input generating device 1 of this embodiment is designed to be a type 2 control system having a disturbance observer, it is possible to suppress a decrease in the accuracy of tracking the target trajectory during a transient response caused by a change in the external environment, even if a change in the external environment occurs that affects the movement of the object M. In other words, when controlling the movement of the object M, it is possible to maintain stable movement of the object M even if a change in the external environment, such as a load fluctuation, occurs during the control.

[0095] The effect of applying a compensator (compensation control input generating unit of the present invention) that generates a compensation input by the control input generating method of the present invention to a control device that uses a general control method was confirmed.

[0096] The experiment was carried out using the experimental device shown in FIG. 10(A) which comprises the pneumatic drive circuit shown in FIG. 10(B), and the accuracy of positioning the table to a predetermined position by the operation of the pneumatic cylinder was confirmed.

[0097] The pneumatic cylinder used in the experimental equipment is a general-purpose double-acting pneumatic cylinder (SMC, CDJ2E16-150Z-B, inner diameter 16 mm, rod length 150 mm). The table is supported by a linear guide (THK, SHW12CA1UUM190LM), and moves while being guided by the linear guide.

[0098] The operation of the double-acting pneumatic cylinder was controlled by a flow-control type servo valve (MPYE-5-M5-010-B, manufactured by Festo). Specifically, the pressure in both cylinder chambers was measured by a pressure sensor (PA-500-103G, manufactured by Nidek Components), and the pressure difference between the two cylinder chambers was adjusted by the servo valve based on this measurement value. The supply pressure of air supplied to the servo valve was 0.3 MPa.

[0099] The position of the table was determined by measuring the displacement of the rod of a double-acting air cylinder with a linear encoder (iC-Haus, iC-PZ205 EVAL PZ1M_Z, resolution 12.5 nm).

[0100] The control system that generates the control signal supplied to the servo valve is a type 1 feedforward control system with a disturbance observer, as shown in Figure 3. It was implemented on a PC using RTAI, a real-time extension of the Linux kernel. The control system's pressure data acquisition and voltage application to the servo valve were achieved via SPI communication between various IC chips and a Raspberry Pi 4B. The control system's pressure data acquisition was achieved by A / D converting the signal supplied from the pressure sensor, and the control system's voltage application to the servo valve was achieved by D / A converting the control signal generated by the control system. Displacement data from the control system's linear encoder was acquired via UART communication after filtering and noise removal using an STM32 microcontroller.

[0101] The sampling period of the control system is 2 ms, and socket communication is performed using UDP between the Raspberry Pi 4B and the PC. The time from A / D conversion request to obtaining the converted value is achieved in 0.5 ms, which is sufficient to execute real-time tasks.

[0102] Furthermore, because the displacement data from the linear encoder used in this experiment contains noise, the deviation between the target value and the measured value is not exactly zero. As a result, the compensator continues to output a compensation input, causing chattering near the target value. To address this, the control system used in this experiment sets a threshold value for the deviation, and if the deviation is below the threshold, the compensation input is not output.

[0103] Furthermore, in the extremely low speed range, the effect of noise on the speed parameter becomes large, making the second control input generation part of the compensator unstable. As a countermeasure, the control system used in this experiment sets a threshold value for speed, and if the speed is below the threshold value, the second control input is not output.

[0104] Example 1: The control performance of the control input generation method of the present invention was verified when a compensation input generated by a compensator employing the control input generation method of the present invention was incorporated into a target generated force and a control input. In the experiment, the response (position deviation) when step inputs of 10 mm and 50 mm were repeatedly given as target values ​​was confirmed for the case where the compensation input generated by the control input generation method of the present invention was input (Example 1, see FIG. 3 ) and the experimental results (Comparative Example 1) where the compensation input was not input (a state in which the compensation control input generation unit 15 in FIG. 3 was not present).

[0105] First, the experimental results of Comparative Example 1 are shown in Figure 11. The experimental results in Figure 11(A) are graphs of responses from 3 to 24 seconds, and Figure 11(B) is a graph of responses in the first step (3 to 6 seconds) of the experimental results in Figure 11(A), showing an enlarged scale of the position deviation. Also, the parameters of the main control input generator 11 in Figure 3 are Kp = 100 [N / m], Kd = 10 [N / m], K ｆｕ = 12.5 [N －１As shown in Fig. 11, when no compensation input was applied, an overshoot of about 2 mm occurred, and a steady-state deviation of at least about 0.05 mm remained.

[0106] Next, the experimental results of Example 1 are shown in Figures 12 and 13. Note that Figure 12(A) is a graph of the experimental result for 3 to 6 seconds, and Figure 12(B) is a graph of the response in the first step (3 to 6 seconds) of the experimental result of Figure 12(A), showing the scale of the position error enlarged. Also, the parameter of the compensation control input generating unit 15 incorporated for the target generated force is K １ = 0.1 N, K ２ = 2.0N, α = 0.8, and the parameters of the compensation control input generator 15 incorporated into the control input are K １ = 20, K ２ = 100, α = 0.8, the deviation threshold was set to 20 nm, and the speed threshold was set to 40 nm / ms. The other parameters were the same as those in the comparative example. As shown in Figure 12, when the compensation input was applied, the overshoot was significantly reduced compared to when the compensation input was not applied, and it was confirmed that the deviation in the steady state was kept below 50 nm.

[0107] Figure 13(A) shows the change in target generated force relative to deviation, enlarging the range from 10.0 s to 10.2 s in Figure 12(A). As shown in Figure 13(A), the input obtained by adding the compensation input Fc (Yc in Figure 3) to the main input Fr is given to the generated force control unit. As the deviation approaches 0, the compensation input Fc increases in the negative direction, acting as a powerful brake. Furthermore, after the overshoot, the compensation input Fc continues to reach its maximum value, and when the speed reaches 0, only the first control input Fr acts, immediately displacing the target value.

[0108] Figure 13(B) shows the change in control input relative to the deviation over the same time period as Figure 13(A). The control input is shown as a 12-bit digital value, which corresponds to the voltage applied to the valve. The input given to the valve is the main input U plus the compensation input Uc (Yc in Figure 3). As in the case of the target generated force, it can be seen that the compensation input Uc acts as a powerful brake, reducing overshoot and immediately displacing the force toward the target value (Figure 13(B)).

[0109] From the above results, it was confirmed that by incorporating a compensator employing the control input generation method of the present invention into the target generated force and control input, it is possible to position to the target value with high accuracy even when using a pneumatic actuator.

[0110] Furthermore, in the above example, it was confirmed that the deviation in the steady state can be reduced to 50 nm or less by measuring the table position with a linear encoder having a resolution of 12.5 nm. On the other hand, when a compensator employing the control input generation method of the present invention is used, if the detection accuracy of the sensor is improved, it is possible to obtain a positioning accuracy that corresponds to the detection accuracy of the sensor.

[0111] Below, we confirmed that by changing the sensor that detects the table position and using a compensator that employs the control input generation method of the present invention, it is possible to obtain positioning accuracy that corresponds to the detection accuracy of the sensor.

[0112] The experimental results are shown in Figures 23 and 24. Figures 23 and 24 show the results of an experiment similar to that of Example 1, using a linear encoder (BL50H manufactured by Magnescale) with a resolution of 6.1 pm as the sensor for detecting the table position. The experiment in Figure 23 was conducted under the same conditions as in Example 1, except that a control system that generates a main control input signal by PID control rather than a feedforward system having a disturbance observer was used as the control system, and a compensation input signal was added to this control system to form the control input signal, and the above-mentioned sensor was used. Figure 24 was also conducted under the same conditions as in Example 1, except that the above-mentioned sensor was used.

[0113] Note that Fig. 23(A) is a graph of the response from 0 to 24 seconds, and Fig. 23(B) is a graph of the response in the first step (0 to 4 seconds) of the experimental results of Fig. 23(A), showing an enlargement of part B. Similarly, Fig. 24(A) is a graph of the response from 0 to 24 seconds, and Fig. 24(B) is a graph of the response in the first step (0 to 4 seconds) of the experimental results of Fig. 24(A), showing an enlargement of part B.

[0114] First, as shown in FIG. 23, it was confirmed that even in a control system that generates a main control input signal by PID control, the deviation in the steady state can be kept to 40 nm or less simply by employing a compensator that employs the control input generation method of the present invention.

[0115] Furthermore, as shown in FIG. 24, it was confirmed that even with the same control system as that of Example 1, the deviation in the steady state could be kept to ±5 nm or less by improving the resolution of the sensor.

[0116] From the above results, it was confirmed that by incorporating a compensator employing the control input generation method of the present invention into the target generated force and control input, it is possible to position to the target value with high accuracy, even when using a pneumatic actuator, regardless of the control system. Furthermore, it was confirmed that even when the same control system and the same compensator are used, by changing the accuracy (resolution) of the sensor that detects the position, it is possible to achieve positioning accuracy that corresponds to the accuracy of the sensor.

[0117] <Example 2> In Example 2, it was confirmed that even for minute changes in the target value, by incorporating a compensator that employs the control input generation method of the present invention into the target generated force and control input, it is possible to position the target value with high accuracy.

[0118] In Example 2, the response (position deviation) was confirmed when a step input of 0.1 μm in width was continuously applied every 3 s in a steady state where a step response of a target value of 50 mm was performed. The experiment was conducted under the same conditions as Example 1, except that a step input of 0.1 μm in width was continuously applied every 3 s.

[0119] The results are shown in Fig. 14. As shown in Fig. 14, by continuously applying step inputs with a width of 0.1 µm every 3 seconds, a larger overshoot (about 1.5 µm) occurs compared to Example 1, but it can be confirmed that the deviation in the steady state is within ±50 nm.

[0120] The results of the experiment are the same as those of Example 2, except that a linear encoder (BL50H manufactured by Magnescale) with a resolution of 6.1 pm was used to detect the position of the table, and that step inputs with a width of 1 nm were continuously applied every 2 seconds.

[0121] The results are shown in Figure 25. In Figure 25, the vertical axis has the target value of 50 mm set to 0. As shown in Figure 25, even when step inputs of 1 nm width were continuously applied every 2 seconds, it was confirmed that the target trajectory could be closely followed, although there was a fluctuation of about ±1 nm from the step input that formed the target trajectory.

[0122] From the above results, it was confirmed that by incorporating a compensator employing the control input generation method of the present invention into the target generated force and control input, it is possible to use a pneumatic actuator to position to the target value with high precision even in response to minute changes in the target value, and to accurately track fluctuations in the target value.

[0123] <Example 3> In Example 3, it was confirmed that even when an external force is applied, positioning to the target value can be performed with high accuracy by incorporating a compensator that employs the control input generation method of the present invention into the target generated force and control input.

[0124] In Example 3, the response (position deviation) was confirmed when a 0.1 μm step feed was performed with a 3 kg load suspended via a pulley as shown in Figure 15(A) . Note that the experiment was carried out under the same conditions as Example 2, except that a 3 kg load was suspended and the step feed was 0.1 μm.

[0125] The results are shown in Figure 15(B). The tensile force was equivalent to approximately 50% of the maximum force calculated from the supply pressure and the pressure-receiving area of ​​the cylinder, but even under such an environment, positioning was possible with the same accuracy as when there was no tensile force.

[0126] From the above results, it was confirmed that by incorporating a compensator employing the control input generation method of the present invention into the target generated force and control input, it is possible to position the target value with high precision using a pneumatic actuator, even when an external force is applied.

[0127] <Example 4> In Example 4, it was confirmed that when a control system to which a compensator employing the control input generation method of the present invention is applied is designed to be a type 2 control system having a disturbance observer, it is possible to position to a target value with high accuracy even when a load fluctuation occurs during operation.

[0128] In Example 4, the load applied to the pneumatic cylinder was changed as follows to confirm the movement state when the table was lifted to a predetermined position by the operation of the pneumatic cylinder. In the experiment, the following cases were confirmed before the table was lifted to the predetermined position: (1) a load was applied during movement from an unloaded state, (2) an additional load was applied during movement from a state where a certain initial load was applied, and (3) a first load was applied during movement from an unloaded state and then an additional second load was applied during movement (see FIG. 18(B)). Note that the experiment was conducted under the same conditions as Example 1, except that the pneumatic cylinder was extended and retracted in the vertical direction and the applied load was changed.

[0129] The no-load state refers to a state in which only the load of the plate attached to the tip of the cylinder ("m1" in FIG. 18(B)) is applied. The load applied during the movement in experiment (1), the first load in experiment (3), and the initial load in experiment (2) correspond to "m2" in FIG. 18(B). The additional load applied during the movement in experiment (2) and the second load in experiment (3) correspond to "m3" in FIG. 18(B).

[0130] In experiments (1) to (3), the apparatus was configured as follows. First, as shown in Figures 17 and 18(A), two tables (tables 1 and 2: each weighing 3 kg) supported by a frame were installed at different heights on the path of movement of the pneumatic cylinder. These two tables were guided by the frame and could move up and down, but were mounted on the frame so that they could not move below a certain height. The pneumatic cylinder was installed so that it contacted the two tables from below. In other words, the pneumatic cylinder was installed so that its axial direction was parallel to the vertical direction and the cylinder could extend and retract along the vertical direction. A plate was attached to the tip of the pneumatic cylinder rod to stably support the tables.

[0131] <Experiments (1) to (3)> Experiments (1) to (3) were each conducted as follows. In the experiments, as shown in Figures 19 to 21, when a control signal (Reference, target value) controlling the operation of the pneumatic cylinder was input, the position control status of the pneumatic cylinder was confirmed in two cases: when the control system was designed to be a type 1 control system with a disturbance observer and a compensator employing the control input generation method of the present invention was applied (Conventional + FF, Device 1), and when the control system was designed to be a type 2 control system with a disturbance observer and a compensator employing the control input generation method of the present invention was applied (Improved + FF, Device 2). Both Device 1 and Device 2 implemented feedforward (FF) control.

[0132] <Experiment (1)> In experiment (1), the experiment begins with the pneumatic cylinder in a contracted state (the state shown in FIG. 18(A)). When the experiment begins, the pneumatic cylinder is extended. Then, the pneumatic cylinder extends with a load of m1 (the weight of the plate) (no load state) (period X0 in FIG. 18(B)), and the plate comes into contact with table 1, lifting table 1. That is, the weight m2 of table 1 is applied as a load at the installed position of table 1, and the load applied to the pneumatic cylinder changes from m1 to m1 + m2. When the pneumatic cylinder further extends from the state in contact with table 1, the pneumatic cylinder extends with a load of m1 + m2 (period X1 in FIG. 18(B)). Then, before table 1 comes into contact with table 2, the pneumatic cylinder is contracted. When the pneumatic cylinder contracts, table 1 is held by the frame in its original position, and the load applied to the pneumatic cylinder changes. In other words, when the air cylinder contracts, the load is reduced by the weight m1 of the table 1 at the original position of the table 1, and the load returns to the initial m1 state and contracts to the original length.

[0133] <Experiment (2)> In experiment (2), the pneumatic cylinder is extended from its contracted state (the state shown in FIG. 18(A)). The experiment begins with the plate coming into contact with table 1 and a load m2 from table 1 being applied to the pneumatic cylinder (i.e., a load of m1 + m2 is being applied). When the experiment begins, the pneumatic cylinder is extended. The pneumatic cylinder then extends with a load of m1 + m2 (period X1 in FIG. 18(B)), causing table 1 to come into contact with table 2 and lifting table 2. In other words, the weight m3 of table 2 is added as an additional load at the installed position of table 2, and the load applied to the pneumatic cylinder changes from m1 + m2 to m1 + m2 + m3. When the pneumatic cylinder is further extended from the state in which table 1 is in contact with table 2, the pneumatic cylinder extends with a load of m1 + m2 + m3 (period X2 in FIG. 18(B)). Then, when the pneumatic cylinder is extended by a predetermined amount, it is retracted. When the pneumatic cylinder is retracted, the table 2 is held in its original position by the frame, and the load on the pneumatic cylinder changes. In other words, when the pneumatic cylinder is retracted, the load decreases by the weight m3 of the table 2 at the original position of the table 2, and the load returns to the initial state of m1 + m2, and the pneumatic cylinder is retracted to its original length.

[0134] <Experiment (3)> In experiment (3), the experiment begins with the pneumatic cylinder in a contracted state (the state shown in Figure 18(A)). When the experiment begins, the pneumatic cylinder is extended. The pneumatic cylinder then extends with a load of m1 (no load state) (period X0 in Figure 18(B)). Initially, the plate contacts the table 1 below and lifts it. That is, the weight m2 of the table 1 is applied as a load when the table 1 is installed, and the load on the pneumatic cylinder changes from m1 to m1 + m2. Then, as the pneumatic cylinder further extends from the state in which it is in contact with table 1, the pneumatic cylinder continues to extend with a load of m1 + m2 (period X1 in Figure 18(B)). Eventually, table 1 comes into contact with table 2, and with tables 1 and 2 overlapping, the pneumatic cylinder lifts both tables 1 and 2. In other words, when table 2 is in its installed position, the weight m3 of table 2 is added as an additional load, and the load on the pneumatic cylinder changes from m1 + m2 to m1 + m2 + m3. When the pneumatic cylinder further extends from its contact with table 2, the pneumatic cylinder extends with a load of m1 + m2 + m3 during this period (period X2 in Figure 18(B)). When the pneumatic cylinder extends by a predetermined amount, the pneumatic cylinder retracts and contracts, and tables 1 and 2 are held by the frame in their original positions, changing the load on the pneumatic cylinder. In other words, when the pneumatic cylinder retracts and contracts, the load decreases by the weight m2 of table 2 when table 2 is in its original position, and decreases by the weight m1 of table 1 when table 1 is in its original position, returning the load to the initial m1 state and contracting to its original length.

[0135] The results are shown in Figures 19 to 21.

[0136] 19 to 21, in all of experiments (1) to (3), although there was a slight delay before the target value was reached when the load changed, it was confirmed that Device 1 showed fluctuations that generally followed the target value. On the other hand, Device 2 showed very little delay from the target value, and showed fluctuations that were almost the same as the target value.

[0137] From the above results, it was confirmed that by applying a compensator that employs the control input generation method of the present invention, it is possible to control the operation of a pneumatic cylinder so that it exhibits good tracking to the target value even when there is a load fluctuation. In particular, it was confirmed that by designing a control system to be a type 2 control system that has a disturbance observer, it is possible to control the operation of a pneumatic cylinder so that it moves in nearly agreement with the target value even when there is a load fluctuation.

[0138] The control input generating device of the present invention is suitable as a device for generating a control input for controlling the operation of an actuator operated by pneumatic pressure.

[0139] REFERENCE SIGNS LIST 1 Control system 2 Pneumatic cylinder 3 Piping 4 Pneumatic pressure source 5 Servo valve 10 Control device 11 Main control input generation unit 15 Compensation control input generation unit 16 First control input generation function 17 Second control input generation function 17a Arrival time calculation function 17a 17b Estimated step number calculation function 18 Output generation function S1 First control input S2 Second control input Yc Compensation input L Control amount Lr Target value e Deviation

Claims

1. A device that generates a control input for controlling a pneumatic actuator, comprising: a first control input generation function that generates a first control input when there is a deviation of a controlled variable from a target value; a second control input generation function that generates a second control input that increases as the deviation of the controlled variable from the target value decreases when there is a time change in the controlled variable; and an output generation function that generates the control input based on the first control input and the second control input.

2. A control input generating device as claimed in claim 1, characterized in that the first control input generating function has a function of generating a steady-state control input as the first control input, or when there is a time change in the controlled variable, has a function of generating a control input as the first control input that is a combination of a steady-state control input and a fluctuating control input that becomes smaller as the deviation of the controlled variable from the target value becomes smaller.

3. The control input generating device according to claim 1, characterized in that the second control input generating function calculates a predicted number of steps N until the controlled variable coincides with the target value based on the deviation of the controlled variable from the target value and the operating state of the pneumatic actuator, and generates the second control input based on the predicted number of steps N so that the second control input becomes larger as the predicted number of steps N becomes smaller.

4. The control input generating device according to claim 2 or 3, characterized in that the first control input generating function generates the first control input S1 based on formula 1, the second control input generating function generates the second control input S2 based on formula 2, and the output generating function generates the control input based on formula 3. Formula 1 S1 = K １ ・sgn(e) Formula 2 S2=K2 / N α ・sgn(v) Formula 3 Yc=K １ ・sgn(e)-K2 / N α sgn(v) e: deviation of the controlled variable from the target value v: displacement speed of the pneumatic actuator α: index for adjusting the timing at which the suppression force that suppresses the operation of the pneumatic actuator starts to act K １ = Gain that generates thrust K ２= Gain that generates inhibitory force N = Number of expected steps 5. The first control input generation function is a gain K that compensates for the nonlinear friction of the pneumatic actuator based on Equation 2-1 and generates a thrust force. １ The second control input generation function is configured to set a gain K that generates a suppression force for reducing an overshoot with respect to a target value based on Equation 2-2. ２ The control input generating device according to claim 4, further comprising: １ = Ka | e | > Eth K １ = Ka / (Eth |e|) When |e|≦Eth, Ka: Specified control input, Eth: Specified displacement, |e|: Absolute value of deviation e, Equation 2-2 K ２ =Kb |v|>Vth K ２ = 0 |v| ≦ Vth Kb: defined control input Vth: defined displacement velocity |v|: absolute value of velocity v 6. The control input generating device according to claim 3, characterized in that the second control input generating function has an arrival time calculating function which calculates the time Ta required for the controlled variable to match the target value based on the deviation of the controlled variable from the target value, and an expected step number calculating function which calculates an expected step number N = Ta / Δt, which is the time Ta divided by a sampling period Δt which detects the deviation of the controlled variable from the target value, and when the expected step number N is smaller than a predetermined threshold value, the expected step number calculating function sets the expected step number N as a threshold value.

7. The control input generating device according to claim 6, wherein the arrival time calculation function calculates the arrival time Ta based on the following formula 4. Formula 4 (1 / 2) a Ta ２ +v Ta - e = 0 a: Displacement acceleration of the pneumatic actuator v: Displacement speed of the pneumatic actuator Ta: Arrival time (arrival time - current time) e: Deviation (target position - current position) 8. The control input generating device according to claim 3, characterized in that the second control input generating function calculates the predicted number of steps N based on the following formula 5. Formula 5 N = e / ΔL = e / (v·Δt) e: deviation of the controlled variable from the target value v: displacement speed of the pneumatic actuator ΔL: amount of change in displacement from one sampling time before 9. The control input generating device according to claim 1, characterized in that the target value is a target position of an object operated by the pneumatic actuator, a driving speed of the pneumatic actuator, or a pressure generated by the pneumatic actuator.

10. A control device for controlling a pneumatic actuator, comprising the control input generating device according to claim 1.

11. A control device for controlling a pneumatic actuator, comprising: a main control input generating section for generating a main control input for controlling the pneumatic actuator; and a compensation control input generating section for generating a compensation control input for correcting the main control input, wherein the compensation control input generating section is the control input generating device according to claim 1.

12. The control device according to claim 10 or 11, characterized in that the control device is a type 2 control system having a disturbance observer.

13. The control device according to claim 12, characterized in that the control device performs feedforward control.

14. An actuator comprising: an operating section that is operated by air pressure; and a control section that controls the operation of said operating section, said control section being the control device according to claim 10 or 11.

15. A calculation circuit comprising: a circuit to which a target value and a control quantity are input as input signals, the calculation circuit comprising: a first control input generating unit that generates a first control input when there is a deviation of the control quantity from the target value; a second control input generating unit that generates a second control input that increases as the deviation of the control quantity from the target value decreases when there is a change in the control quantity over time; and an output generating unit that generates and outputs a control input based on the first control input and the second control input.

16. A control board having an input terminal to which a target value and a control amount are input as input signals, and an output terminal for generating and outputting a control input, the control board being provided with: a main controller for generating a main control input based on the signal input from the input terminal; a compensator for generating a compensation control input based on the signal input from the input terminal; and an output generating unit which is an adder circuit that combines the main control input and the compensation control input to generate a control input and supplies it to the output terminal, the compensator having: a first control input generating function for generating a first control input when there is a deviation of the control amount from the target value; a second control input generating function for generating a second control input which increases as the deviation of the control amount from the target value decreases when there is a time change in the control amount; and an output generating function for generating and outputting a control input based on the first control input and the second control input.

17. A control board having an input terminal to which a target value and a control amount are input as input signals, and an output terminal to generate and output a control input, the control board being provided with: a compensator that generates a compensation control input based on the signal input from the input terminal; and an output generating unit which is an adder circuit that generates a control input by combining the compensation control input with a control input supplied from the outside and supplies the control input to the output terminal, the compensator having: a first control input generating function that generates a first control input when there is a deviation of the control amount from the target value; a second control input generating function that generates a second control input which becomes larger as the deviation of the control amount from the target value becomes smaller when there is a time change in the control amount; and an output generating function that generates and outputs a control input based on the first control input and the second control input.

18. A method for generating a control input for controlling a pneumatic actuator, comprising: generating a first control input when there is a deviation of a controlled variable from a target value; generating a second control input that increases as the deviation of the controlled variable from the target value decreases when there is a time change in the controlled variable; and generating the control input based on the first control input and the second control input.

19. A control input generating method as described in claim 18, characterized in that the first control input is a steady-state control input, or, when there is a time change in the controlled variable, the first control input is a control input that is a combination of a steady-state control input and a fluctuating control input that becomes smaller as the deviation of the controlled variable from the target value becomes smaller.

20. A control input generating method as claimed in claim 18, characterized in that a predicted number of steps N until the controlled variable coincides with the target value is calculated based on the deviation of the controlled variable from the target value and the operating state of the pneumatic actuator, and the second control input is generated based on said predicted number of steps N so that the smaller the predicted number of steps N, the larger the second control input becomes.

21. The control input generating method according to claim 18 or 19, further comprising: generating the first control input S1 based on an equation 1; generating the second control input S2 based on an equation 2; and generating the control input based on an equation 3. １ ・sgn(e) Formula 2 S2=K2 / N α ・sgn(v) Formula 3 Yc=K １ ・sgn(e)-K2 / N α sgn(v) e: deviation of the controlled variable from the target value v: displacement speed of the pneumatic actuator α: index for adjusting the timing at which the suppression force that suppresses the operation of the pneumatic actuator starts to act K １ :Gain that generates thrust K ２ : Gain that generates the suppression force N: Number of predicted steps 22. Gain K, which compensates for the nonlinear friction of the pneumatic actuator based on Equation 2-1 and generates thrust. １ and a gain K that generates a suppression force that reduces an overshoot with respect to the target value based on Equation 2-2. ２ The control input generating method according to claim 21, further comprising: setting the following equation 2-1 K １ = Ka | e | > Eth K １ = Ka / (Eth |e|) When |e|≦Eth, Ka: Specified control input, Eth: Specified displacement, |e|: Absolute value of deviation e, Equation 2-2 K ２=Kb |v|>Vth K ２ = 0 |v| ≦ Vth Kb: defined control input Vth: defined displacement velocity |v|: absolute value of velocity v 23. A method for generating control input as described in claim 20, characterized in that it calculates a time Ta required for the controlled variable to match the target value based on the deviation of the controlled variable from the target value, calculates an expected number of steps N = Ta / Δt, which is the time Ta divided by a sampling period Δt for detecting the deviation of the controlled variable from the target value, and if the expected number of steps N is smaller than a predetermined threshold value, sets the expected number of steps N as the threshold value.

24. The control input generating method according to claim 23, further comprising the step of calculating the arrival time Ta based on the formula 4. Formula 4 (1 / 2) a Ta ２ +v Ta - e = 0 a: Displacement acceleration of the pneumatic actuator v: Displacement speed of the pneumatic actuator Ta: Arrival time (arrival time - current time) e: Deviation (target position - current position) 25. The control input generating device according to claim 20, characterized in that the second control input generating function calculates the predicted number of steps N based on the following formula 5. Formula 5 N = e / ΔL = e / (v·Δt) e: deviation of the controlled variable from the target value v: displacement speed of the pneumatic actuator ΔL: amount of change in displacement from one sampling time before 26. The control input generating method according to claim 18, wherein the target value is a target position of an object operated by the pneumatic actuator, a driving speed of the pneumatic actuator, or a pressure generated by the pneumatic actuator.

27. A control input generating program, which is a program for causing a control device that controls the operation of a pneumatic actuator to execute the control input generating method according to any one of claims 18 to 25.

28. A storage medium having a control input generation program stored thereon, characterized in that the storage medium has the control input generation program of claim 27 stored thereon.