KR960014174B1 - Real time controlling device - Google Patents

Abstract

a belong function generating part converting one of the belong functions to 25 element fuzzy word; a fuzzy inference part performing the inference by performing MUX operation of the MIN operation result of the belong function and the MAX operation result of the fuzzy word; a non-fuzzy part outputting a control input by converting 25 parallel lines to one signal; a processor generating a control signal to an object system; and a look-up table between the non-fuzzy part and the processor.

Description

Real time control device using lookup table of fuzzy controller

1 is a block diagram of a real-time control device using a look-up table of the fuzzy controller according to an embodiment of the present invention.

2 is a block diagram of a conventional fuzzy controller.

3 is a block diagram of an overall fuzzy data parallel inference engine.

4 is an explanatory diagram showing in detail a process of converting a belonging function from a decoder into a fuzzy word of 25 elements in the belonging function generator.

* Explanation of symbols for main parts of the drawings

8: Decoder 10: Member function generator

20: fuzzy reasoning unit 21,22: membership function circuit

23,24: MIN calculator 25: MAX calculator

30: non-fuzzy section 40: look-up table

50: processor

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fuzzy controller, and more particularly, to a real-time control apparatus using a lookup table of a fuzzy controller having reduced memory capacity by using a look-up table.

Conventionally, fuzzy technology has been mainly used in industrial fields such as plant control, expert system information processing, subway, etc., but recently, it has been widely applied in various fields such as home appliances, etc. due to low price of hardware. .

The conventional technology of such a purge controller is one having the configuration shown in FIG. In the figure, reference numeral 10 denotes a plurality of membership functions input from the decoder 8 via a selection bus, that is, a membership function generator that converts one of the seven membership functions into a 25 element fuzzy wand. It consists of two membership function circuits (21, 22), two MIN operators (23, 24), and a MAX operator (25), and generates a membership function for any two input variables. It represents a fuzzy inference unit that performs inference by performing the MAX operation with the fuzzy Wonder from the belonging function generator 10, 30 is 25 parallel lines by analyzing the fuzzy words inferred by the fuzzy inference unit 20 by the center of gravity method Denotes a non-fuge unit for outputting a control input by converting the signal into a single signal, and 50 denotes a fuzzy inference unit while generating a control signal to a target system to be controlled based on the control input from the non-fuge unit 30. 2) as It shows a processor which feeds back as one input variable. 3 is a block diagram of the entire fuzzy data parallel inference engine, and FIG. 4 is a membership function from the decoder 8, that is, NL (abbreviated as NL) and NM (Negative Medium: NM). Abbreviated), ZR (Zero: abbreviated as ZR), PS (Positive Small: abbreviated to PS), PM (Positive Medium: abbreviated to PM), PL (positive large: abbreviated to PL) 7 shows a detailed process of changing the belonging function of the belonging function generator 10 into 25 fuzzy words.

In the conventional fuzzy controller having the above configuration, each belonging function circuit section 21 and 22 has one input and an output, but the belonging function generating section 10 has only 25 output signals and no input is required. Therefore, the conventional fuzzy controller outputs the membership function for any input, as shown in FIG. 2, and then performs the MIN operation through the two MIN operators 23 and 24, and performs the MAX operator 25. Performs the MAX operation. Then, the fuzzy wandering unit 30 analyzes the purge wand by the center of gravity method, converts 25 parallel lines into a single signal and outputs a control input, thereby allowing the processor 5 to control based on this control input. The control signal y is provided to the target system. At this time, the output control signal y of the processor 50 is also input to the belonging function circuit section 22 in the fuzzy inference section 20 as a state input variable of the system.

However, when the prior art as described above is applied to the control of a high dimensional and complex system, efficient control of the system becomes impossible. In other words, the real-time control of the system is difficult.

Accordingly, an object of the present invention is to provide a fuzzy controller capable of real-time control of a high-dimensional and complex system by using a look-up table.

In order to achieve the above object, the present invention provides a membership function generator for converting one of a plurality of membership functions inputted from a decoder through a selection bus into a fuzzy wand of 25 elements, and a membership function for any two input variables. A fuzzy inference unit that generates the inference by performing the MAX operation between the MIN operation result and the fuzzy word of the belonging function, and analyzes the inferred fuzzy word by the center of gravity method and converts 25 parallel lines into one signal. In the control device of the fuzzy controller comprising a non-fuzzy unit for outputting a control input and a processor for generating a control signal to the target system to be controlled based on the control input and returning it as one input variable of the fuzzy inference unit. It further includes a look-up table, which is created from a fuzzy value between the fuzzy part and the proset, and controlled through this look-up table. Look of a fuzzy controller to quantify the force, characterized in that provided in the processor provides real-time control device with up table.

Other objects and various advantages of the present invention will become more apparent from the preferred embodiments of the present invention described below with reference to the accompanying drawings by those skilled in the art. Hereinafter, preferred embodiments of the purge controller according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram of a real-time control apparatus using a lookup table of a fuzzy controller according to the present invention. As shown in the figure, the fuzzy controller according to the present invention uses one of the seven membership functions, PL, PM, PS, ZR, NS, NM, and NL, which are input from the decoder 8 through the selection bus. Affiliation function generating unit 10 for converting a 25 element fuzzy word, two affiliated function circuits 21 and 22, two MIN operators 23 and 24 and a MAX operator 25 Fuzzy inference unit 20 for generating inference function for input variable (x, y) and performing inference by performing MIN operation result of this belonging function and MAX operation with the fuzzy wonder from the above mentioned function function generator 10 And a non-fuzzy unit 30 for analyzing the deduced fuzzy words outputted from the fuzzy inference unit 20 by the center of gravity method and converting 25 parallel lines into one signal to output a control input Z. Quantify control inputs through tables created from non-fuzzy values Generating a control signal (y) to the target system to be controlled based on a quantized control input from the look-up table (40) which makes the output 24 quantization levels for the entire set; At the same time, it consists of a processor 50 which feeds back as an input variable of the belonging function generator 22 in the fuzzy inference unit 20. In the same figure, the member shown as the same reference numeral as that shown in FIG. Represents the same structural member performing the same function.

The fuzzy controller according to the present invention, as clearly shown in FIG. 1 and described above, has a look-up table 40 in the prior art described above, which makes the unfuzzy outputs 24 quantization levels for the entire set. The main feature is that it is provided between the non-fuzzy section 30 and the processor 50, by making the output values obtained through the inference process in the form of a table can save a lot of time during inference This enables real-time control of the system.

On the other hand, in the fuzzy controller of the present invention, in order to operate a hardware system, a control composed of language control rules of a person skilled in the art should be designated as a terminal terminal of a board. It must be programmed to be used. Therefore, by the programming is assigned a function belonging mainly to the language control rules, the label of this function and the implementation of the hardware system is as shown below and shown in Table 1 and Table 2, respectively.

Next, the synthesis rule of the inference according to the present invention is subjected to the MAX-MIN center method, where the non-fuzzy output (Z) of the non-fuzzy section 30 uses the center of gravity method as shown in the following equation. Will be used.

In the above formula, μ represents the membership function value.

In addition, as described above, the look-up table generated by quantizing 24 levels of the entire set from the non-fuzzy values according to the present invention is shown in Table 3 below.

As described above, in the fuzzy controller of the present invention, as shown in Table 2, the look-up table value is a control input of the processor 50, so that not only the storage capacity is reduced but also the real-time control of the system is possible. Also, the quantified output of the processor is inputted back into the belonging function circuit 22 in the fuzzy inference unit 20 and fed back to infer. As described above, the present invention enables the real-time control of a high-dimensional and complex system by using a look-up table made from non-fuzzy values.

Therefore, in the future, it is expected that the implementation of a fuzzy computer system is possible due to the hardware configuration of the fuzzy control that can be realized in full-scale real time control as in the present invention. It is anticipated that the application of such a system to the system control field, which was possible only by humans, will be visualized.

Claims (1)

A membership function generator for converting one of a plurality of membership functions inputted from the decoder through the selection bus into a fuzzy word of 25 elements, and a membership function for any two input transformations are generated. A fuzzy inference unit for performing inference by performing MAX operation with the fuzzy word, a non-fuzzy unit for outputting a control input by converting 25 parallel lines into a single signal by analyzing the deduced fuzzy wand by a center of gravity method; A fuzzy controller comprising a processor which generates a control signal to a target system to be controlled based on the control input and feeds back a single input variable to the fuzzy inference unit, wherein the non-fuge unit is disposed between the non-fuge unit and the processor. It further comprises a look-up table made from the values, through the look-up table to quantify the control input Fuzzy controller, characterized in that provided in the processor group.