JPH05108362A - Fuzzy development back-up device - Google Patents

Abstract

PURPOSE:To provide a fuzzy development back-up device which can perform the rational selection between a look-up table system and a fuzzy inference system according to the limit conditions when a fuzzy inference is applied to a product (device) for execution of the fuzzy control. CONSTITUTION:The fuzzy knowledge is generated based on the output set to the input and the designing limit conditions and stored in a fuzzy knowledge storage part 4. Thus a fuzzy inference system memory capacity/inferring speed calculating part 7 calculates the memory capacity and an inferring speed necessary for execution of a fuzzy inference based on the fuzzy knowledge. Meanwhile a look-up table system memory capacity/inferring speed/approximate accuracy calculation part 9 calculates the memory capacity necessary for production of a look-up table accordant with the output set by the fuzzy inference result obtained based on the input and the produced fuzzy knowledge. Then the memory capacity and the inferring speed that are calculated at both parts 7 end 9 are shown a Pareto optimal solution curve display part 10.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to the construction of optimum fuzzy knowledge for realizing an ideal control response in a fuzzy controller for controlling based on fuzzy rules, and fuzzy inference by the fuzzy controller. The present invention relates to a fuzzy development support device for supporting the determination of the execution form of.

[0002]

2. Description of the Related Art In recent years, industrial applications of fuzzy technology have remarkably spread, and many fuzzy application products have been announced. The biggest bottleneck in developing fuzzy application products is fuzzy knowledge (fuzzy rules and members. Ship function).

In order to support the construction of this fuzzy knowledge, some fuzzy development support devices have been proposed, for example, the membership function shape is graphically displayed to support the design environment of a fuzzy knowledge base designer.

Generally, when fuzzy inference is incorporated into a product (apparatus) to perform fuzzy control, either a look-up table method or a fuzzy inference method in which the microprocessor of the incorporated apparatus (product) operates the fuzzy inference is applied. To be done.

The look-up table method is a method in which fuzzy inference is performed in advance on a combination of input variable values to calculate output values, and a set of output values corresponding to the input values is referred to in a table format. In this method, inference is performed at the same speed as the microprocessor accesses the memory and at a constant speed, but the required memory capacity increases exponentially as the input / output variables and their resolution increase. It also has drawbacks. On the other hand, the fuzzy inference method can realize the fuzzy inference with a relatively small memory capacity even when there are many input / output variables, but has a drawback that the inference speed becomes slow as the number of fuzzy rules increases.

[0006]

The design constraint conditions in the fuzzy knowledge construction such as the number of input / output variables and the approximation accuracy are critical ones that it is not clear which lookup table method or fuzzy inference method should be selected. In many cases, designers have empirically selected either method.

However, empirical selection cannot be said to be rational, and an inexperienced designer may erroneously select a method. If the method is selected incorrectly, the scale (memory capacity) of the fuzzy controller becomes too large,
It may not be possible to obtain sufficient responsiveness for control.

The present invention has been made in view of the above circumstances, and provides a fuzzy development support device capable of rational selection of a look-up table method and a fuzzy inference method that meet design constraints. It is a thing.

[0009]

SUMMARY OF THE INVENTION The present invention is a fuzzy development support device for generating fuzzy knowledge based on an output with respect to an input and a design constraint condition, which is required when performing fuzzy inference based on the generated fuzzy knowledge. Fuzzy inference method evaluation means for obtaining memory capacity and inference speed, and look-up table method memory capacity evaluation means for obtaining memory capacity required to create a look-up table according to an output from a fuzzy inference result based on input and generated fuzzy knowledge And a display means for displaying the memory capacity or the inference speed obtained by the fuzzy inference method evaluation means and the lookup table method memory capacity evaluation means.

[0010]

When the fuzzy knowledge is generated based on the output for the input and the design constraint condition, the memory capacity and the inference speed necessary for performing the fuzzy inference by the generated fuzzy knowledge are obtained by the fuzzy inference method evaluation means, Further, the memory capacity required for creating the look-up table according to the input and the output based on the fuzzy inference result based on the generated fuzzy knowledge is obtained by the look-up table type memory capacity evaluating means. Then, the memory capacity or inference speed obtained by the fuzzy inference method evaluation means and the look-up table method memory capacity evaluation means is displayed on the display means.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 is a schematic block diagram of an embodiment of the device of the present invention.

Reference numeral 1 is an input / output data storage unit for storing input / output data (experimental data) of the system set by operating an input means (not shown), and 2 is similarly given design constraint conditions (number of input / output variables and approximation). The design constraint storage unit 3 for storing the accuracy) has various fuzzy division numbers (however, 2).
Rule / membership function standard form generating unit 4 for generating fuzzy knowledge (fuzzy rule and membership function) of division or more) is for various fuzzy division numbers generated by the rule / membership function standard form generating unit 3. It is a fuzzy knowledge storage unit that stores fuzzy knowledge.

Reference numeral 5 is a fuzzy knowledge storage for approximating the input / output data of the input / output data storage unit 1 from the fuzzy knowledge stored in the fuzzy knowledge storage unit 4 and the input / output data stored in the input / output data storage unit 1. Fuzzy modeling in which the fuzzy knowledge stored in the unit 4 is modified (modeling), fuzzy inference is performed using each modified fuzzy knowledge, and the approximation accuracy with the input / output data stored in the input / output data storage unit 1 is calculated. An approximation accuracy calculation unit 6 is a fuzzy inference unit used in the fuzzy modeling / approximation accuracy calculation unit 5 and a lookup table type memory capacity / inference speed / approximation accuracy calculation unit 9 described later.

Reference numeral 7 is a fuzzy inference method evaluation for calculating a memory capacity and an inference speed necessary for performing fuzzy inference using fuzzy knowledge of various modeled fuzzy division numbers stored in the fuzzy knowledge storage unit 4. The fuzzy inference method memory capacity / inference speed calculation section 8 as means is a fuzzy inference method memory capacity / inference speed calculation section 7 and a look-up table method memory capacity / inference speed / approximation accuracy calculation section 9 which will be described later. It is an MPU-compatible instruction clock count database that stores the number of operation clocks, the number of memory access clocks, and the clock frequency of various microprocessors (MPUs) that are referred to when calculating.

Reference numeral 9 is necessary when creating a lookup table of various resolutions from the result of fuzzy inference using the modeled fuzzy knowledge of various division numbers stored in the fuzzy knowledge storage unit 4. Memory capacity, inference speed (memory access speed), lookup table method for calculating approximation accuracy Memory capacity as a memory capacity evaluation means Memory capacity / inference speed /
Approximate accuracy calculation unit, 10 is a fuzzy inference method memory capacity
Based on the inference speed, the memory capacity, and the approximation accuracy of both formulas calculated by the inference speed calculation unit 7 and the look-up table method memory capacity / inference speed / approximation accuracy calculation unit 9, a non-inferior (Pareto optimal solution) curve is given to the designer. It is a fuzzy inference method / lookup table method non-inferior solution curve display unit as a display means to be presented.

In this embodiment, for the sake of simplicity, it is assumed that the fuzzy rule uses simple inference such that the consequent part is a real value.

In such an apparatus, the rule / membership function standard form generator 3 is configured to input / output data of the system stored in the input / output data storage 1 and the input / output variables stored in the design constraint storage 2 A standard form of fuzzy knowledge (fuzzy rule, membership function) as shown in FIGS. 2 and 3 is generated from the numbers.

FIG. 2 shows a membership function of the antecedent part when the number of input variables is 2 (A, B) and the number of fuzzy divisions is 3, and the range of input values is equally divided and the shape of the membership function is isosceles. It is a triangle and is set to intersect with the adjacent membership function at a grade value of 0.5. Also,
FIG. 3 shows a fuzzy rule composed of the membership function shown in FIG. 2 (consequent part real number type, output variable number 1). The number of input variables and the number of output variables are arbitrary numbers stored in the design constraint storage unit 2, and the fuzzy division number, the membership function shape, and the intersection condition with an adjacent membership function are also arbitrary values. Is set.

Fuzzy knowledge of various fuzzy division numbers generated by the rule / membership function standard form generating unit 3 is stored in the fuzzy knowledge storing unit 4. The fuzzy knowledge of various fuzzy division numbers stored in the fuzzy knowledge storage unit 4 uses the input / output data stored in the input / output data storage unit 1 to perform fuzzy modeling / approximation accuracy calculation unit 5 and fuzzy inference unit 6 The input / output data is corrected (modeled) to approximate fuzzy knowledge and stored again in the fuzzy knowledge storage unit 4.

As a method of fuzzy modeling, for example, "automatic tuning of fuzzy reasoning by the steepest descent method and its application to obstacle avoidance" (1990 6th Fuzzy System Symposium Proceedings, pp. 423-4)
26). When the fuzzy modeling is completed, the approximation accuracy of each fuzzy division number (approximation accuracy with the input / output data stored in the input / output data storage unit 1) is also calculated and stored in the fuzzy knowledge storage unit 4. ..

Here, the approximation accuracy is calculated by, for example, the sum of squared errors shown by the following equation. In this equation, E is the error sum of squares, the number of input and output data d is stored in the output data storage unit 1, y _h ^t is h-th output data values,
y _h is an inference value obtained as a result of fuzzy inference using the h-th input data.

[0022]

[Equation 1]

Then, in the fuzzy inference method memory capacity / inference speed calculation unit 7, from the fuzzy knowledge modeled in the fuzzy modeling / approximation accuracy calculation unit 5, the microprocessor (MPU) stored in the MPU compatible instruction clock frequency database 8 is obtained. Inference speed and memory capacity are calculated by referring to the number of instruction clocks for each type.

If the number of input variables, the number of output variables, and the number of fuzzy divisions are determined, the memory capacity of the fuzzy knowledge whose consequent part of the fuzzy rule is a real value is calculated by the following formula.

[0025]

[Equation 2]

For example, a fuzzy knowledge structure having 10 antecedent structure bytes, 8 membership function structure bytes, 10 antecedent structure bytes, and 20 fuzzy rule structure bytes. , The number of input variables is 2,
If the number of output variables is 1 and the number of fuzzy divisions is 3, the fuzzy knowledge memory capacity is 258 bytes from the above equation.

2 × (10 + 3 × 8) + 1 × 10 + 3 ² × 20 = 258 Further, in the case of fuzzy knowledge having a different structure, the memory capacity is obtained by a different formula. For example, if the fuzzy rule consequent part is not a real number but a membership function,

[0028]

[Equation 3]

It is calculated by

Then, the memory capacity required for the fuzzy inference method is obtained by adding the program capacity required for fuzzy inference to the calculated memory capacity for fuzzy knowledge.

Regarding the inference speed, fuzzy inference is performed by the fuzzy inference unit 6 using the input data stored in the input / output data storage unit 1 and the fuzzy knowledge stored in the fuzzy knowledge storage unit. The memory capacity / inference speed calculation unit 7 measures the number of additions, multiplications, and divisions executed by fuzzy inference, and calculates the number of operations (average value or worst value (the maximum number of operations is required for each fuzzy inference. )))). And
The fuzzy reasoning speed in the product to be applied is calculated by multiplying the calculated number of operations with the number of clocks required for each operation registered in the MPU-compatible instruction clock number database 8.

FIG. 4 shows an example of the MPU-compatible instruction clock count database. In this MPU-compatible instruction clock number database, the instruction clock number and clock frequency of the microprocessor (MPU) installed in the product in which fuzzy inference is mounted are registered in advance.

On the other hand, the lookup table type memory capacity / inference speed / approximation accuracy calculation unit 9 calculates the memory capacity required to construct the lookup table from the resolution (the number of bits) required for the input variable and the output variable. To do. For example, if the number of input variables is 2 and the number of output variables is 1 and the resolution is 8 bits each, the value of each input variable is considered from 0 to 255, and the input data pattern is 256 × 256.
= 65536 ways. The resolution of the output variable is 8
Since it is a bit, the memory capacity is 64 Kbytes. This memory capacity is calculated according to the number of input / output data and the resolution.

Further, the lookup table type memory capacity / inference speed / approximation accuracy calculation unit 9 also calculates the approximation accuracy for each resolution.

For example, data in the range of -1 to +1 may be input to a certain input variable, and when the resolution is 4 bits, the correspondence between the input data value and the bit pattern shown in FIG. 5 is obtained. Is attached. In FIG. 5, the bit pattern is 0000 if the input data value is in the range of -1 to 0.8667. Then, using this input data-bit pattern correspondence relationship, fuzzy inference is executed by the fuzzy inference unit 6, and the approximation accuracy is obtained by the above sum of squared errors. Incidentally, the input data-bit pattern correspondence relationship as shown in FIG. 5 is not limited to one input variable, and similarly for other input variables and output variables, the data range and the resolution thereof are determined, and each data value The bit pattern correspondence is determined.

Since the inference speed is the same as the memory access speed, the MPU corresponding to the MPU-compatible instruction clock count database 8 is installed in the product to be incorporated.
The memory access speed corresponding to is the call inference speed.

In this way, the fuzzy reasoning method memory capacity
When the memory capacity and the approximation accuracy are calculated by the inference speed calculation unit 7 and the fuzzy inference method memory capacity / inference speed calculation unit 9, the respective memories calculated by the fuzzy inference method / lookup table method non-inferior solution curve display unit 10 are obtained. The non-inferior solution (Pareto optimal solution) curve of each method is displayed based on the capacity and the approximation accuracy. A display example of the non-inferior solution curve is shown in FIG. Thus, both displayed non-inferior solutions (Pareto
By looking at the (optimal solution) curve, it becomes possible for the designer to make a rational choice between the look-up table method and the fuzzy reasoning method.

[0038]

As is apparent from the above description, the present invention can be applied to design constraint conditions (the number of input / output variables, accuracy, etc.) and system input / output data (for example, experimental data) when incorporating fuzzy inference into an applied product. Accordingly, fuzzy knowledge that satisfies the design constraint is generated, and the non-inferior solution curve regarding the memory capacity and the inference speed when the lookup table method and the fuzzy inference method are adopted is displayed. This allows the designer to rationally select the look-up table method or the fuzzy inference method.

Further, by giving different design constraint conditions (the number of input / output variables, approximation accuracy, etc.), it is possible to estimate the inference speed of each system and the required memory capacity. When there is no feasible solution, it is possible to support interactive design work such as obtaining a feasible solution by relaxing constraints.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram of an embodiment of a device of the present invention.

FIG. 2 is a diagram showing a shape of a standard membership function in the case of a fuzzy division number of 3;

FIG. 3 is a diagram showing an example of a simplified fuzzy rule.

FIG. 4 is a diagram showing an example of an MPU-compatible instruction clock count database according to an embodiment of the device of the present invention.

FIG. 5 is a diagram showing an example of an input data-bit pattern correspondence relationship according to an embodiment of the device of the present invention.

FIG. 6 is a diagram showing a display example of a non-inferior solution curve according to an embodiment of the device of the present invention.

[Explanation of symbols]

1 input / output data storage unit 2 design constraint condition storage unit 3 rule membership function standard form generation unit 4 fuzzy knowledge storage unit 5 fuzzy modeling / approximation accuracy calculation unit 6 fuzzy inference unit 7 fuzzy inference method memory capacity / inference speed calculation unit (Fuzzy inference method evaluation means) 8 MPU-compatible instruction clock frequency database 9 Look-up table method memory capacity / inference speed /
Approximation accuracy calculation unit (lookup table type memory capacity evaluation means) 10 Fuzzy inference method / lookup table type non-inferior solution curve display unit (display means)

Claims (1)

[Claims] 1. A fuzzy inference method for obtaining a memory capacity and an inference speed required when performing fuzzy inference based on the generated fuzzy knowledge in a fuzzy development support device that generates fuzzy knowledge based on an output with respect to an input and a design constraint condition. Evaluating means, look-up table type memory capacity evaluating means for obtaining a memory capacity required for creating a look-up table according to an output based on a fuzzy inference result based on input and generated fuzzy knowledge, the fuzzy inference method evaluating means and look A fuzzy development support device, comprising: a display means for displaying the memory capacity or the inference speed obtained by the up-table type memory capacity evaluation means.