JP2532610B2 - Program control target value generator - Google Patents

Description

Description: BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to generation of a target value of a program control device that follows a predetermined time-varying target value.

Conventional Technology A conventional example will be described with reference to FIGS. 8 and 9. The program control is control to follow a target value that is set in a time-varying manner as shown in FIG. 8, and there are two methods for generating the target value. The first method will be described with reference to FIG. FIG. 9 shows the relationship between the target value r _{i and} the time t _i . This relationship is prepared as a ROM table, the ROM is accessed at each time t ₁ , t ₂ , t _3, ... And the stored target value data r ₁ , r ₂ , r ₃ ,.

The second method is to prepare the relationship of the target value with respect to time as a functional expression, and is shown by the following expression, for example.

r (t) = (1 + 3t ² + t ^−1.5 ) ^{2 Based} on the above equation, r (t ₁ ), r for each time t ₁ , t ₂ , t ₃ , ...
(T ₂ ), r (t ₃ ), ... Is obtained and output.

Problems to be Solved by the Invention However, the conventional methods have the following problems. In the conventional first method, target values are set discretely with respect to time, and when the time interval is shortened for more precise control, the number of set target values becomes enormous,
Requires a large ROM capacity.

On the other hand, the second conventional method does not require a memory because the target value is prepared as a functional expression, but when the target value with respect to time has a complicated non-linear relationship, it is not possible to determine the functional expression. It's not easy.

Means for Solving the Problem The present invention provides a program control device that follows a target value that fluctuates with time, wherein time or an input amount other than time is used as an antecedent input variable of a fuzzy inference rule, and the target is A value is used as a consequent part output variable, and the target value is generated by fuzzy inference using a fuzzy inference device in which the consequent part of the fuzzy inference rule is a real value instead of a fuzzy number.

Action The present invention has the above-described configuration to set the target value to be generated to IF
~ THEN… According to the given fuzzy inference rules described in formation, the fuzzy variables of the input variables are converted to the antecedent part (IF
The real value of the output in the consequent part (THEN part) is determined from the degree of conformity (membership value) to the rule. The actual target value to be generated is obtained by taking the barycentric value of the real value of the output by a plurality of rules. This makes it possible to generate a target value for a continuous input amount with a small memory capacity.

First Embodiment A first embodiment of the present invention will be described with reference to the drawings. In FIG. 1, reference numeral 1 is a time input unit for inputting a time measured value by a timer or the like. 2 is an input section other than time, and is used when there is an input other than time. 3 is a fuzzy inference operation unit,
An inference calculation for determining a target value is performed from the input amounts from the time input unit 1 and the input unit 2 other than the time. A fuzzy inference rule storage unit 4 stores inference rules used for fuzzy inference. The inference rules used here are expressed in the form of "IF ~ (antecedent part) THEN ... (consequent part)", the antecedent part is an inference proposition containing fuzzy variables, and the consequent part is not a fuzzy number. It is expressed in normal real numbers. For example:

IF t is Zero and x is Positive Medium Then r is −
A fuzzy variable storage unit 3.5 stores fuzzy variables such as "zero" and "positive and small" used in the antecedent part of the fuzzy inference operation. An inference rule management unit 6 retrieves fuzzy variables from the fuzzy variable storage 5 and inference rules from the fuzzy inference rule storage unit 4. A target value output unit 7 outputs the result of the inference operation.

The operation will be described below. When time and input values other than time are input to the fuzzy inference operation unit 3, the fuzzy inference operation unit 3 retrieves the inference rules or fuzzy variables used in the operation from the fuzzy inference rule management unit 6 and the fuzzy variable storage unit 5. To instruct. The fuzzy inference rule management unit 6 takes out inference rules and fuzzy variables from the fuzzy inference rule storage unit 4 and the fuzzy variable storage unit 5, and outputs them to the fuzzy inference calculation unit 3. The fuzzy inference operation unit 3
Fuzzy inference as shown below is performed.

In the present invention, the fuzzy variables in the antecedent part are defined as triangular symmetric membership functions as shown in FIG. 2, and each fuzzy variable is set to have a constant distance.

Now, time input value t ⁰ and input value other than time X _j ⁰ j = 1,2
When m is measured, the fuzzy number of the conclusion of each rule in fuzzy reasoning is calculated as follows. Where μ
_It (t ⁰ ) and μ _ij (X _j ⁰ ) represent the membership values of the input values t ⁰ and X _j ⁰ for the fuzzy number of the antecedent proposition, and f _i is the real value of the consequent part.

· U = the time of the _{f i μ i (u) =} μ it (t 0) Λμ i1 (X 1 0) Λ ... Λμ im (X m 0) · other when μ _i (u) = 0 where, i = 1,2 ..., n This represents the degree of conformance of the antecedent part. Where Λ is
Represents the min operation.

The comprehensive inference result can be obtained as follows by using the weighted center of gravity.

Hereinafter, the case where the time t and one variable other than the time x exist as inputs will be specifically described. For example, as shown in FIGS. 3 (a) and 3 (b), the fuzzy variable is normalized to the interval of [0,12] for the time t and [−6,6] for the variable x,
The whole is equally divided into 6 and 7 fuzzy variables are set. An example of a fuzzy reasoning rule table is shown in FIG. When the input is t ⁰ = 0.5 and X ⁰ = 3, the following four rules in the rule table of FIG. 4 are applied.

Rule 1 lf t is ZO and x is PS Then r is −2. Rule 2 lf t is S and x is PS Then r is −1. Rule 3 lf t is ZO and x is PM Then r is −1. Rule 4 lf t is S and x is PM Then r is −0. Figure 5 shows the inference process. From Fig. 5, the conformity of each rule from Rule 1 to Rule 4 is μ ₁ (u) = 0.5, μ ₂ (u)
= 0.25, μ ₃ (u) = 0.5, μ ₄ (u) = 0.25.

Therefore, the target value r ^* is By preparing the rule table of FIG. 4 in this way, it is possible to generate target values for all planes defined by the variables t and x.

A second embodiment of the present invention will be described. The second embodiment relates to a device having a function of adjusting the real value of the consequent part of a fuzzy inference rule for obtaining a desired target value in a fuzzy program control device. In FIG.
The same parts as those in FIG. 1 are designated by the same reference numerals and the description thereof will be omitted. An input generation unit 8 generates an input that covers the entire range of the input. Reference numeral 9 is a consequent part real value input unit for inputting a real value of the consequent part of the inference rule to be changed and its inference rule. 10 is a consequent part management part, which is based on the inference rule to be changed specified from the consequent part real value input part 9 and the consequent part real value,
The contents of the fuzzy inference rule storage unit 4 are changed. An inference result storage unit 11 stores the result of inference by the fuzzy inference operation unit 3 with respect to the input values of all the domains generated by the input generation unit 8. A display unit 12 displays the inference results of all the input domains stored in the inference result storage unit 11.

Hereinafter, the operation will be described by taking as an example the case where the time t and one variable x other than the time exist as input variables. Input generator 8
Then, an input is generated that covers the entire inside of each domain of the tx plane spanned by two variables t, x. Based on the input value generated by the input generation unit 8, as described in the first embodiment, the fuzzy inference operation unit 3, the fuzzy inference rule storage unit 4, the fuzzy variable storage unit 5, and the inference rule management unit 6 perform fuzzy logic. Inference is performed, an inference result is obtained, and the result is stored in the inference result storage unit 11. The display unit 12 displays the inference result for the entire input domain stored in the inference result storage unit 11. Seventh example of display
Shown in the figure. This display result is the target value generated by the fuzzy program controller. If this displayed target value is different from the desired target value, adjust the consequent real value of the consequent part of the rule that is involved in the generation of the different target value. At the time of adjustment, the consequent part real value of the rule to be changed is input to the consequent part real value input unit 9, inference is performed again, and the change result is confirmed on the display unit. By repeating the input of real values and the confirmation on the display, the target value is approached.

Effects of the Invention As described above, the present invention relates to generation of a target value of a program control device, and generates a target value by fuzzy inference. As a result, the target value for the entire input domain can be set by preparing a small amount of fuzzy inference rules, and the memory capacity can be saved.
In addition, when determining the real value of the consequent part, it is possible to visually check the inference result on the display unit, which facilitates adjustment.

[Brief description of drawings]

FIG. 1 is a block diagram of a program control target value generator according to the first embodiment of the present invention, FIG. 2 is a diagram showing fuzzy variables, and FIGS. 3 (a) and 3 (b) are the first embodiment of the present invention. Showing fuzzy variables of FIG. 4, FIG. 4 is a fuzzy inference rule table of the first embodiment of the present invention, FIG. 5 is an explanatory view of the fuzzy inference process of the first embodiment of the present invention, and FIG. 6 is the present invention. FIG. 7 is a block diagram of a program control target value generation device of the second embodiment, FIG. 7 is a target value display example diagram of the second embodiment of the present invention, FIG. 8 is a block diagram of the program control device, and FIG. It is a figure explaining the target value generation method of a prior art example. 1 ... time input unit, 2 ... input unit, 3 ... fuzzy inference operation unit, 4 ... fuzzy inference rule storage unit, 5 ... fuzzy variable storage unit, 6 ... inference rule management unit, 7 ... goal Value output part, 8 ... Input generation part, 9 ... Consequent part Real value input part,
10 …… Consequent part management part, 11 …… Inference result storage part, 12 …… Display part.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (56) References Maeda and Murakami “Self-adjusting fuzzy controller”, Transactions of the Society of Instrument and Control Engineers Vol. 24, No. 2 (Shown 63.2) PP. 191-197

Claims (3)

(57) [Claims] 1. In a program control device that follows a target value that varies with time, time or an input amount other than time is used as an antecedent part input variable of a fuzzy inference rule, and the target value is output as a consequent part. A target value generation device for program control, characterized in that the target value is generated by fuzzy inference using a fuzzy inference device in which the consequent part of the fuzzy inference rule is a real value instead of a fuzzy number as a variable. 2. The program control target value generation device according to claim 1, further comprising an inference result storage unit for storing a result of the fuzzy inference operation as a real value instead of a fuzzy number. 3. The target value generation device for program control according to claim 2, further comprising a display unit for displaying the inference result stored in the inference result storage unit.