JPH0916206A - Controller for fuzzy control - Google Patents

Abstract

PURPOSE: To provide a controller for fuzzy control which is equipped with plural knowledge bases, but can instantaneously switch the knowledge bases. CONSTITUTION: A RAM 12 stores the knowledge bases KB1 -KB3 in different address areas. A ROM 13 includes none of the knowledge bases KB1 -KB3 and stores an inference program IP that the knowledge bases KB1 -KB3 share. When one of the knowledge bases KB1 -KB3 is specified with a switch or by a control program, the inference program IP performs fuzzy inference by using the selected knowledge base.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention is mainly applied to production lines,
The present invention relates to a fuzzy control controller that performs fuzzy control on a system including a plurality of control targets such as a plant and an air conditioning system, and a device including a plurality of control elements such as an electric discharge machine and an NC machining machine.

[0002]

2. Description of the Related Art Generally, a lot of knowledge and experience are required to control the above-mentioned system and equipment, and thus, a control operation has been conventionally performed by a skilled operator. On the other hand, fuzzy control has recently been adopted in various fields so that this kind of knowledge and experience can be utilized for automatic control of systems and devices. In fuzzy control, the experience and knowledge possessed by a skilled person is formalized as a fuzzy conditional proposition based on linguistic control rules, and the manipulated variable is determined from the observed data of the controlled variable by fuzzy reasoning to control the controlled object.

To perform fuzzy control, a knowledge base that stores knowledge such as membership functions and rules used for fuzzy reasoning and a knowledge that is stored in the knowledge base are applied to motion information of a controlled object to perform fuzzy reasoning. A fuzzy control controller having a fuzzy inference unit for performing is used and a support device having a man-machine interface such as a personal computer is used to tune the knowledge stored in the knowledge base according to the controlled object.
The fuzzy control controller and the support device are RS23
Connection is made using a general-purpose interface (here, a serial interface) such as the 2C specification. Further, the fuzzy control controller is provided with an input / output interface for inputting / outputting analog data to / from a control target.

Generally, in a fuzzy control controller, an inference procedure and knowledge are set according to the controlled object.
The knowledge forming the knowledge base and the program for performing the fuzzy inference are stored in the memory such as the ROM. In the processor, the fuzzy inference program is executed so that the knowledge stored in the memory is applied to the analog data from the input / output interface, and the controlled object is controlled based on the execution result.

The knowledge stored in the memory is set by the procedure shown in FIG. That is, basically, in the support device, a knowledge base (membership function and fuzzy rule) for controlling the controlled object is created, and this knowledge base is stored in the memory of the fuzzy control controller together with the fuzzy inference program. To do. Where the memory is RO
If M, the ROM is mounted on the fuzzy control controller after the knowledge and the program are stored in the memory.

This procedure will be described in more detail. First,
A basic design is performed in the support device (S1). The basic design is to determine an input (type of sensor or installation location) for obtaining information from a control target, and to determine an output (actuator, motor, heater, valve) to be controlled by an inference result. Next, as a detailed design, describe the membership functions and rules applied to each input / output (S
2) Create a knowledge base. In most cases, the knowledge base at this stage does not completely match the control target, so the knowledge base that has been set and the fuzzy inference unit prepared in the support device are used to enter the control target. While outputting, the knowledge base is adjusted and tuned to match the control target (S3). The tuning work is repeated until an optimum result is obtained (S4).

Until the tuning is completed, a high-level language dedicated to the assisting device designed to facilitate programming is used, and after the tuning is completed, the finally obtained knowledge base is used for fuzzy inference. The code is converted into a general-purpose language (for example, C language) together with the above program (S5). By converting the code into a general-purpose language in this way, it is possible to create a knowledge base that does not depend on the processor used in the fuzzy control controller. After that, compile so that a machine language corresponding to the processor used in the fuzzy control controller can be obtained (if the general-purpose language is C language, C suitable for the processor is used.
Use a compiler).

As described above, by using a general-purpose language as the intermediate language, it is possible to set a fuzzy inference program and a knowledge base that do not depend on the processor used in the fuzzy control controller. Moreover, in order to compile from a high-level language dedicated to a support device into a machine language suitable for various processors, it is necessary to create a large number of compilers, but a general-purpose language to machine language compiler such as C language is generally used. Since it is commercially available, if a general-purpose language is used as an intermediate language as described above, it is possible to support various processors by creating only a compiler from a high-level language dedicated to a support device to a general-purpose language as an intermediate language. Become.

The fuzzy inference program and knowledge base created as described above are written in the ROM,
By incorporating it in the fuzzy controller (S
6) The fuzzy control controller performs a desired operation according to the controlled object. It is also possible to use a general-purpose controller such as a personal computer without using a dedicated controller as the fuzzy control controller. In this case, it is not necessary to incorporate the knowledge base in the ROM and one device is used. Can be used both as a support device and a fuzzy control controller.

[0010]

By the way, the controlled object may be changed to such an extent that it cannot be handled within the range of the rule of the set knowledge base. For example, NC
When a processing machine or robot is used, if there is a change in the product, the knowledge base for controlling the NC processing machine or robot must be changed.

In order to change the knowledge base, generally, the operation of the fuzzy control controller is temporarily stopped,
It is necessary to replace the ROM. However,
If such work was done, the knowledge base could not be changed immediately with the change of the controlled object,
When the equipment is made compatible with a wide variety of products, the operating rate of the equipment will decrease.

Instead of replacing the ROM, it is also possible to transfer the knowledge base or inference program created by the support device to the fuzzy control controller by communication. However, since the knowledge base and the inference program are integrated (that is, knowledge is embedded in the inference program), the amount of data to be transferred is large, and it takes time to replace the knowledge base. This will be the case. Furthermore, when using a general-purpose language as an intermediate language as described above, it is possible to compile the intermediate language into a machine language with a fuzzy control controller, but in this case the amount of data becomes larger and the time required for transfer is increased. Not only becomes long, but also requires a large amount of memory for the fuzzy controller.

Therefore, it is possible to prepare a plurality of knowledge bases in the fuzzy control controller and switch the knowledge bases if the knowledge bases need to be changed. In particular, if one device is also used as a support device and a fuzzy controller, it is necessary to store multiple types of knowledge bases in an external storage device such as a hard disk and change the knowledge bases. In this case, it is possible to read the data from the external storage device into the memory and change the knowledge base. However, when the fuzzy control controller is used by incorporating it into equipment, etc., it is common to use the support device separately and separately, so it is necessary to connect the support device every time it is necessary to change the knowledge base. As a result, the work of changing the knowledge base is troublesome, and in the end it does not lead to an improvement in the operating rate.

The present invention has been made in view of the above circumstances, and an object of the present invention is to immediately switch knowledge bases while having a plurality of knowledge bases and use a memory having a relatively small capacity. Another object of the present invention is to provide a fuzzy control controller that can be used for the above-mentioned purposes and can be incorporated into equipment and used.

[0015]

According to the first aspect of the invention, a knowledge base consisting of knowledge of membership functions and rules used for fuzzy inference and knowledge of the knowledge base is applied to motion information of a controlled object. The interface includes a fuzzy inference unit for performing fuzzy inference, an input / output interface unit for inputting / outputting analog data to / from a controlled object, and an interface unit for exchanging data with an external device different from the controlled object. A fuzzy control controller in which knowledge of a knowledge base is tuned by a support device that is detachably connected via a ROM, a ROM storing an inference program not including the knowledge base, and a ROM
A fuzzy inference unit is configured by the processor executing the inference program in the above, and a plurality of knowledge bases that are tuned by the supporting device and share the fuzzy inference unit are stored in different address areas defined on the rewritable memory. A selection means is provided which enables a desired knowledge base to be used by designating the address area on the memory without depending on the supporting device.

According to another aspect of the present invention, a ROM storing a system program, a RAM including a work area, an input / output interface for inputting / outputting analog data to / from a controlled object, and an input / output operating according to the system program A controller including a processor that receives operation information from a controlled object connected to the interface and gives a work instruction to the controlled object, in which an inference program is stored in a ROM and a knowledge base is stored in a RAM.

According to the third aspect of the invention, the processor and the RA are arranged to control the controlled object according to the sequence program.
A programmable controller including a CPU unit having M and a ROM, and an input / output unit for inputting / outputting data between the CPU unit and a controlled object, wherein an inference program for performing fuzzy inference and a knowledge base are provided to the CPU unit. Set and input / output analog data to / from the controlled object via the input / output unit.

According to a fourth aspect of the present invention, there is provided a programmable controller including a CPU unit for controlling an object to be controlled according to a sequence program and a fuzzy control unit including a reasoning program for performing fuzzy reasoning and a knowledge base, the fuzzy control controller being provided in the fuzzy control controller. The analog data is input to and output from the control target via the input / output interface.

[0019]

According to the configuration of the present invention, since a plurality of knowledge bases are prepared and the required one of them is designated by the selecting means, the designated knowledge base can be used. Therefore, the knowledge base is changed. Can be done quickly. That is, in the manufacturing process and the like, complicated control including the change of the knowledge base becomes possible, the controllability is improved, and the change of the knowledge base can be dealt with immediately without stopping the operation of the equipment. This has the advantage that the operating rate of the equipment can be increased. Further, since the fuzzy inference unit is fixedly set in the ROM independently of the knowledge base and the fuzzy inference unit is shared by a plurality of knowledge bases, the fuzzy inference unit (that is, inference program) and the knowledge Compared with the case where the base is integrated, only the knowledge base can be changed without changing the inference program. As a result, even if a plurality of knowledge bases are provided, the occupied capacity of the memory becomes smaller than that of the conventional one, and the knowledge bases can be transferred in a short time.

[0020]

【Example】

(Embodiment 1) The fuzzy control controller in this embodiment is realized by a fuzzy control unit that adds a fuzzy control function to a programmable controller that performs sequence control. That is, a programmable controller is provided on a mother board provided with a power supply unit, data (digital data and analog data) between a CPU unit provided with a processor that executes a sequence program and a program memory that stores the sequence program, and a control target. ), And an fuzzy control unit as a fuzzy control controller.

As shown in FIG. 2, the fuzzy control controller includes a dedicated processor 11 made up of a microprocessor, a RAM 12 for storing knowledge bases KB _{1 to} KB ₃ and a work area WM, and a system program SP (see FIG. 1). (See (a)) and a ROM 13 storing an inference program IP for performing fuzzy inference, a serial interface 14 such as RS232C capable of connecting a supporting device composed of a personal computer, and an analog between a controlled object. An input / output interface 15 capable of exchanging data is provided. In addition, a parallel interface 18 that receives input information from the switch 16 described below and outputs operation information of the switch 16 to display the operation information on the display 17 is provided. Here, a fuzzy inference unit is configured by the ROM 13 storing the inference program IP and the dedicated processor 11, and a plurality of knowledge bases KB _{1 to} KB ₃ (here, three) are provided.
It is stored for each required area set on the AM 12. R
The storage area of the knowledge bases KB _{1 to} KB _{3 in} the AM 12 may be replaced with a flash ROM. Flash RO
If M is used, the knowledge bases KB _{1 to} KB ₃ can be rewritten, and the backup by the battery is not necessary at the time of power failure, etc., and the maintenance becomes easy.

The knowledge bases KB _{1 to} KB ₃ are created in the same manner as in the conventional example by an assisting device (personal computer or the like) connected via the serial interface 14,
After being compiled into a machine language, it is downloaded via the serial interface 14. Here, Fig. 1 (b)
, The inference program IP has each knowledge base KB ₁ ...
It is set to be shared by KB ₃ , and ROM 1 is set in advance.
It is written in 3. Therefore, if the knowledge bases KB _{1 to} KB ₃ are selected by designating the address area on the memory, the fuzzy inference unit (the inference program IP and the dedicated processor 11) causes the selected knowledge bases KB _{1 to} KB ₁ to be selected.
Fuzzy control is performed using information from the control target obtained through B ₃ and the input / output interface 15. That is, when the knowledge bases KB _{1 to} KB ₃ are selected, they are integrated with the inference program IP as shown in FIG. Here, the selection of the knowledge bases KB _{1 to} KB ₃ is instructed by the switch 16.

Now, the state of the process to be controlled is 3
There are different types of knowledge base KB that can be used as the process progresses.
The _{1 ~KB 3,} and what needs to be changed in the order of _{KB 1, KB 2, KB 3} . At this time, in the initial step, the knowledge base KB ₁ is selected, in the next step the operator changes the knowledge base KB ₂ by operating the switch 16 while operating the equipment, and in the final step, the operator selects the switch 16 The knowledge base KB ₃ may be selected by the operation of. Here, the knowledge base KB by the switch 16
The correspondence relation of selecting _{1 to} KB ₃ is defined by the system program SP. That is, if a numerical value is designated by the switch (here, a thumbwheel switch) 16, the knowledge bases KB ₁ to KB ₃ are converted into the top address of the address area in which the knowledge bases KB ₁ to KB ₃ are stored,
Knowledge base KB ₁ ~ corresponding to the value indicated by the switch 16
KB ₃ is selected. Also, the selected knowledge base KB
Numerical value 7 corresponding to _{1 to} KB ₃
Will be displayed. That is, it is possible to confirm that the knowledge bases KB _{1 to} KB ₃ selected by the switch 16 have been selected by the display on the display unit 17.

In the configuration of this embodiment, a programmable controller having a CPU unit and a fuzzy control unit is assumed, but a general-purpose controller (which may be a personal computer) having a CPU, a ROM and a RAM is used for fuzzy control. The inference program IP and the knowledge bases KB _{1 to} KB ₃ may be provided to perform the above processing.

(Second Embodiment) In the first embodiment, the switch 16
Although the knowledge bases KB _{1 to} KB ₃ were selected by the operation of, the present embodiment shows an example controlled by a program. That is, an example is shown in which the knowledge bases KB _{1 to} KB ₃ used in the fuzzy control unit are selected by the sequence program executed by the CPU unit. A plurality of knowledge bases KB _{1 to} KB ₃ are stored in the memory as in the first embodiment, as shown in FIG. 3B. Here, the start addresses of the address areas on the memory storing the knowledge bases KB _{1 to} KB ₃ are H1000 and H, respectively.
2000 and H3000. The H at the beginning indicates that it is a hexadecimal number.

The input / output interface 15 is used not only for inputting / outputting analog data to / from the controlled object, but also for exchanging data with the CPU unit. A shared memory is provided in a portion of the input / output interface 15 for exchanging data with the CPU unit, and the CPU unit and the fuzzy control controller asynchronously exchange data via the shared memory.

When it is necessary to change the knowledge bases KB ₁ to KB ₃ during the execution of the sequence program, FIG.
(A), in inserting the contact X ₁ to X ₃ where necessary to change the knowledge base KB ₁ ~KB ₃ on ladder diagram describing the sequence program, series each contact X ₁ to X ₃ The shared memory setting command CM may be connected to the. If you create such a sequence program, each contact X ₁
When to X ₃ is turned on, in that the contact X ₁ to X ₃ in the shared memory setting instruction CM is executed which are connected in series, the fuzzy control unit using the address that was entered as a parameter in the shared memory setting instruction CM is Knowledge Base KB ₁
To select the ~KB _3. For example, in “F *** H1000” in FIG. 3A, the start address is H10.
This means selecting a knowledge base KB ₁ that is 00.

In this embodiment, an example in which the fuzzy control controller is controlled by the CPU unit is shown.
The same operation can be performed by storing a fuzzy control program and a knowledge base in a general-purpose controller and using a general-purpose language (C, BASIC, etc.) instead of the ladder diagram. (Embodiment 3) The fuzzy control controller in this embodiment includes a CPU unit in addition to the fuzzy control unit. That is, in general CPU unit, taking into account that it includes memory resources of a large capacity as compared with the fuzzy control unit, a plurality of knowledge bases KB ₁ ~K
Store B ₃ in the CPU unit, and if necessary, C
From the PU unit is to transfer one knowledge base KB ₁ ~KB ₃ to the fuzzy control unit.

That is, only one knowledge base KB _{1 to} KB ₃ exists in the memory serving as the knowledge bases KB _{1 to} KB ₃ in the fuzzy control unit, which is necessary as the sequence program in the CPU unit progresses. knowledge base KB ₁ ~KB ₃ is transferred to the fuzzy control unit, is transferred knowledge base KB ₁ ~KB ₃ are to be used. Which knowledge bases KB _{1 to} KB ₃ stored in the CPU unit are transferred to the fuzzy control unit is instructed by the sequence program. For example, assuming that the sequence program is described by the ladder diagram as shown in FIG. 4, the contacts X _{1 to} X
_A knowledge base transfer instruction KT may be connected in series to _3, and a numerical value corresponding to the knowledge bases KB _{1 to} KB ₃ to be selected may be given as a parameter of the knowledge base transfer instruction KT. For example, “F ++ H1” in FIG. 4 means to transfer the knowledge base KB ₁ corresponding to H1.

In this embodiment, since the fuzzy control unit holds only _one knowledge base KB _{1 to} KB ₃ , there is an advantage that the fuzzy control unit does not need a large capacity memory. That is, the storage of the knowledge base KB ₁ ~KB ₃ in fuzzy control unit, in the case of using a relatively expensive flash ROM, by adopting the configuration of this embodiment, there is provided a fuzzy control unit at relatively low cost The advantage is that this can be done. here,
Since the knowledge bases KB _{1 to} KB ₃ are compiled in machine language, the transfer time is relatively short and the memory capacity occupied by the CPU unit is also relatively small.
Moreover, since the transfer of the inference program is unnecessary and only the knowledge bases KB _{1 to} KB ₃ are transferred, this also makes the transfer time relatively short. In other words, the transfer time can be set to a level that does not pose a problem in actual use.

In each of the above-described embodiments, the example in which the input / output interface 15 capable of inputting / outputting analog data is provided in the fuzzy control unit is shown, but the input / output unit may be provided separately from the fuzzy control unit.

[0032]

According to the present invention, since a plurality of knowledge bases are prepared and the required one of them is designated by the selecting means, the designated knowledge base can be used, the knowledge base can be changed quickly. The advantage is that In other words, in the manufacturing process, it is possible to perform complicated control including the change of the knowledge base to improve the controllability, and to immediately respond to the change of the knowledge base without stopping the operation of the equipment. And has the advantage that the operating rate of the equipment is increased. Further, since the fuzzy inference unit is fixedly set in the ROM independently of the knowledge base and the fuzzy inference unit is shared by a plurality of knowledge bases, the fuzzy inference unit (that is, inference program) and the knowledge Compared with the case where the base is integrated, there is an advantage that only the knowledge base can be changed without changing the inference program. as a result,
Even when a plurality of knowledge bases are provided, the occupied capacity of the memory becomes smaller than that of the conventional one, and the knowledge bases can be transferred in a short time.

[Brief description of the drawings]

1A and 1B show a first embodiment, FIG. 1A is a diagram showing a memory usage example, FIG. 1B is a diagram showing a relationship between an inference program and a knowledge base, and FIG. 1C is a knowledge base selected. It is a figure which shows the relationship with.

FIG. 2 is a block diagram of a first embodiment.

3A and 3B show a second embodiment, FIG. 3A is a diagram showing a program example, and FIG. 3B is a diagram showing a knowledge base setting example.

FIG. 4 is a diagram showing an example of a program in the third embodiment.

FIG. 5 is a diagram showing a procedure for creating a knowledge base in a conventional example.

[Explanation of symbols]

11 dedicated processor 12 RAM 13 ROM 14 serial interface 15 input / output interface 16 switch 17 indicator 18 parallel interface IP inference program KB _{1 to} KB ₃ knowledge base SP system program WM work area

─────────────────────────────────────────────────── ─── Continuation of the front page (51) Int.Cl. ⁶ Identification code Internal reference number FI Technical display location G05B 19/05 F

Claims (4)

[Claims] 1. A knowledge base composed of knowledge of membership functions and rules used for fuzzy inference, a fuzzy inference unit for applying fuzzy inference by applying knowledge of knowledge base to motion information of a controlled object, and a controlled object. An input / output interface unit for inputting and outputting analog data to and from,
A fuzzy control controller that has an interface unit that sends and receives data to and from an external device that is different from the control target, and whose knowledge base knowledge is tuned by a support device that is detachably connected through the interface unit. A fuzzy inference unit is composed of a ROM that stores an inference program that does not include a base and a processor that executes the inference program in the ROM, and multiple knowledge bases that are tuned by a support device and that share the fuzzy inference unit can be rewritten. Storage means are provided in different address areas defined on different memories, and the selection means is provided for enabling a desired knowledge base by designating the address areas on the memory without depending on the support device. Fuzzy control controller. 2. A ROM storing a system program
And a RAM including a work area, an input / output interface for inputting / outputting analog data to / from a control target, and operating information according to a system program and receiving operation information from the control target connected to the input / output interface, and to the control target. 2. The fuzzy control controller according to claim 1, wherein the inference program is stored in the ROM and the knowledge base is stored in the RAM. 3. A programmable controller comprising a CPU unit having a processor, a RAM and a ROM so as to control a controlled object according to a sequence program, and an input / output unit for inputting / outputting data between the CPU unit and the controlled object. 2. The fuzzy control according to claim 1, wherein an inference program for performing fuzzy inference and a knowledge base are set in the CPU unit, and analog data is input / output to / from the controlled object via the input / output unit. Controller. 4. A programmable controller comprising a CPU unit for controlling an object to be controlled according to a sequence program and a fuzzy control unit provided with an inference program for performing fuzzy inference and a knowledge base, wherein an input / output interface provided in the fuzzy controller is provided. 2. The fuzzy control controller according to claim 1, wherein analog data is input to and output from the control target via the control target.