JPH05289705A - Fuzzy controller - Google Patents

Abstract

PURPOSE:To surely make a controlling variable match with the target value without spending any time to a fuzzy inference, nor holding many control rules, and nor causing any steady deviation with a simple processing. CONSTITUTION:The deviation value is obtained between the controlled variable of a controlled system 1 and the target value, and a manipulated variable is obtained in accordance with the deviation value with execution of a fuzzy inference and based on a membership function. Under such conditions, the change of the preceding deviation value is detected by a deviation compensating means 10 and a compensation signal is transmitted in response to the detected change of the deviation value. A compensating means 12 receives the compensation signal to compensate the relation between the deviation value and the membership function and then calculates a manipulated variable to eliminate the deviation.

Description

Detailed Description of the Invention

[0001]

FIELD OF THE INVENTION The present invention relates to a fuzzy controller.

[0002]

2. Description of the Related Art FIG. 5 is a block diagram of a fuzzy controller. As the control target 1, as shown in FIG. 6, the object 2 is moved on the base 3 in the arrow (a) direction to reach the target position. The deviation device 4 has a target position Q and an object 2
The current position M is input, and the deviation device 4 outputs a deviation e between the target position Q and the current position X. The deviation e is input to the fuzzy control unit 5, and the fuzzy control unit 5 executes fuzzy inference to obtain the manipulated variable M according to the deviation e.

That is, the fuzzy control section 5 stores a membership function as shown in FIG. This membership function consists of large, medium, and small rules for deviation. If the captured deviation e is e1, the fuzzy control unit 5 obtains a membership function 0.4 from these rules, and obtains a manipulated variable M corresponding to this membership function 0.4. In this way, the operation amount M corresponding to the deviation e is given to the controlled object 1, and the object 2 reaches the target position.

By the way, when the object 2 is brought close to the target position by the fuzzy control, the object 2 comes close to the target position, but the object 2 does not reach the target position due to the frictional force between the object 2 and the base 3. In some cases, the deviation e does not become "0" but becomes constant at a certain value. This is because the frictional force and the driving force acting on the object 2 by the operation amount M are almost the same.

Deviation e between the present position X and the target position Q
As a technique applied to eliminate the problem, for example, Japanese Patent Laid-Open No.
There is a 93904 publication. This technique executes fuzzy inference from the deviation from the current position with respect to the target position and its change amount to obtain a first operation amount, and also executes fuzzy inference from the change amount of the target position to obtain a second operation amount. One manipulated variable is integrated, the second manipulated variable is added to this integrated value, and the final manipulated variable is given to the controlled object.

Another technique is disclosed in Japanese Patent Laid-Open No. 11430/1990.
There is publication No. 1. This technology evaluates the state based on the target position and the deviation between the target position and the current position, selects one control law from the control law group according to this evaluation result, and executes fuzzy inference according to this control law.

However, the former technique requires two fuzzy control units. Therefore, it takes time for each fuzzy control unit to obtain a rule and a membership function, and the amount of state to be handled increases. or,
In the latter technique, many control laws must be retained,
Further, it is difficult to set a standard for selecting the optimum control law.

[0008]

As described above, it takes a long time to obtain rules and membership functions, and it is necessary to maintain many control rules and set a standard for selecting an optimum control rule.

In view of the above, the present invention is capable of reliably matching the control amount to the target amount without causing time consuming fuzzy inference, maintaining a large number of control rules, and making a steady deviation with a simple process. An object is to provide a control device.

[0010]

According to the present invention, a deviation amount between a control amount of a controlled object and a target amount is taken in, and an operation amount corresponding to the deviation amount is obtained by executing fuzzy inference based on a membership function, In a fuzzy control device for giving this manipulated variable to a controlled object, a deviation compensating means for sending a compensation signal according to a change in the deviation amount, and a relationship between the deviation amount and a membership function by receiving the compensation signal from the deviation compensating means. And a compensating means for compensating for the above.

[0011]

By providing such means, when the deviation amount between the controlled variable of the controlled object and the target amount is taken in and the manipulated variable corresponding to the deviation amount is obtained by executing the fuzzy inference based on the membership function, The deviation compensating means detects a change in the deviation amount between the control amount of the controlled object and the target amount, and sends a compensation signal corresponding to this change. When this compensation signal is received by the compensation means, the relationship between the deviation amount and the membership function is compensated, and the operation amount for eliminating the deviation is obtained.

[0012]

DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention will be described below with reference to the drawings. The same parts as those in FIG. 4 are designated by the same reference numerals, and detailed description thereof will be omitted.

FIG. 1 is a block diagram of a fuzzy controller.
The steady deviation compensator 10 has a function of detecting a change in the deviation e and sending a compensation command F corresponding to the change to the fuzzy controller 11. Specifically, the steady deviation compensator 10 has a function of integrating the transfer function (-K / s), and sends the integrated value of the deviation e as a compensation signal F.

On the other hand, the fuzzy control unit 11 has a function of fetching the deviation e and executing a fuzzy inference to obtain an operation amount M corresponding to the deviation e based on the membership function, and also has a membership function moving unit 12. ing.

The membership function moving unit 12 receives the compensation signal F, detects from the value of the compensation signal F that the deviation e is constant and does not change, and the deviation e and the member are determined according to the deviation e at this time. To compensate for the relationship with the ship function,
That is, it has a function of moving the membership function for the deviation e as shown in FIG. Next, the operation of the device configured as described above will be described.

The target position Q and the current position M of the object 2 are input to the deviation device 4, and the deviation device 4 outputs a deviation e between the target position Q and the current position X. This deviation e is input to the fuzzy control unit 5.

As shown in FIG. 2, if the deviation e fetched as shown in FIG. 2 is e2, the fuzzy control unit 5 obtains the membership function 0.3 from the rule of "small", and calculates the manipulated variable M corresponding to this membership function 0.3. Ask. In this way, the operation amount M according to the deviation e is given to the controlled object 1, and the object 2
Reaches the target position.

However, the case where the distance between the object 2 and the target position Q does not decrease below the deviation e2 will be described. In this case, the frictional force between the object 2 and the base 3 and the operation amount M are as described above. Therefore, the driving force acting on the object 2 is almost the same. In this state, the steady deviation compensator 10 takes in the deviation e and integrates it, and sends the integrated value to the fuzzy controller 11 as a compensation signal F.

The membership function moving unit 12 in the fuzzy control unit 11 receives the compensation signal F and detects from the value of the compensation signal that the deviation e is constant and does not change. In this detection, when the deviation e is constant and does not change, the integrated value increases at a constant degree. Therefore, the constant deviation is detected by the increase of the integrated value. The membership function moving unit 12 moves the membership function to the left in the figure according to the deviation e at this time, as shown in FIG.

As a result, the fuzzy control unit 5 obtains the membership function 0.2 from the rule of "medium" if the deviation e2 is taken in as shown in FIG. 3, and calculates the manipulated variable M corresponding to this membership function 0.2. Ask. Since the operation amount M at this time is obtained from the “medium” rule, it is larger than the operation amount M obtained from the “small” rule. Therefore, the driving force of the object 2 is increased, the object 2 overcomes the frictional force and moves, and reaches the target position.

As described above, in the above embodiment, when the deviation e between the current position X and the target position Q is taken in and the manipulated variable M corresponding to the deviation e is obtained by executing fuzzy inference based on the membership function, The steady-state deviation compensator 10 detects the change in the deviation e, and sends the compensation signal F corresponding to this change to move the membership function for the deviation e to obtain the operation amount M for eliminating the deviation e. The frictional force between the object 2 and the base 3 is not substantially the same as the driving force acting on the object 2, and a driving force equal to or greater than this driving force is applied to the object 2
Can surely reach the target position Q. Further, even if the target position Q moves, the object 2 can be made to follow the target position Q without deviation e.

The present invention is not limited to the above-mentioned embodiment, but may be modified within the scope of the invention. For example, although the membership function moving unit 12 moves the membership function, it may be added to the deviation e by a predetermined amount and the added value may be used as the deviation e ′. Also, the membership function may be compressed as shown in FIG. On the other hand, the control target can be applied to drive amount control of a mechanical system.

[0023]

As described above in detail, according to the present invention, it is possible to reliably perform a fuzzy inference without taking time, maintaining a large number of control rules, and performing a simple process without causing a steady deviation. It is possible to provide a fuzzy control device that can match the control amount with the target amount.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing an embodiment of a fuzzy control device according to the present invention.

FIG. 2 is a diagram showing a membership function in the apparatus.

FIG. 3 is a diagram showing movement of a membership function in the device.

FIG. 4 is a diagram showing a modified example of movement of a membership function in the apparatus.

FIG. 5 is a configuration diagram of a conventional device.

FIG. 6 is a diagram showing a control target.

FIG. 7 is a diagram showing fuzzy inference of a conventional device.

[Explanation of symbols]

1 ... Control object, 4 ... Deviation device, 10 ... Steady-state deviation compensator, 1
1 ... Fuzzy control unit, 12 ... Membership function moving unit.

Claims (1)

[Claims] 1. A deviation amount between a control amount of a controlled object and a target amount is taken in, a fuzzy inference is executed to obtain an operation amount corresponding to the deviation amount based on a membership function, and this operation amount is controlled by the controlled object. In the fuzzy control device, the deviation compensation means for sending a compensation signal according to the change of the deviation amount, and the compensation signal from the deviation compensation means are received to compensate the relationship between the deviation amount and the membership function. A fuzzy control device comprising: a compensation means.