JPH0644390A - Membership value output circuit for fuzzy system - Google Patents

Abstract

PURPOSE:To provide the membership value output circuit for fuzzy system which makes a non-Neuman type of the fuzzy system and can accelerate processing speed. CONSTITUTION:Flash type A/D converters 1a-1d composed of (n) pieces of comparators to compare an input signal with a reference voltage V and decoders to decode the outputs of these comparators are used for detecting the inclination angle of an inverted pendulum 4. Then, the membership values of respective fuzzy sets are instantaneously outputted from the respective A/D converters 1a-1d and processed by following respective CPU 7a-7c, and the Negative-Small and Positive-Small membership values are obtained.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a fuzzy system based on fuzzy inference, and more particularly to a method for outputting membership values of a fuzzy system that requires high speed control
The present invention relates to a technique effectively applied to a membership value output circuit of a fuzzy system capable of shortening the processing time.

[0002]

2. Description of the Related Art Recently, in order to express an ambiguous relationship such as "similar" or "almost the same",
Fuzzy computers and fuzzy controllers that use the concept of fuzzy sets have been developed, and fuzzy systems such as general home appliances that incorporate this technology have appeared.

The fuzzy inference used in these systems, for example, first A / D-converts the input signal to output the membership value of each fuzzy set from the input signal, and then, for each fuzzy set (for example, NB). , NM, N
S, Z, PS, PM, PB, etc.) and the membership value is output.

As a technique similar to this, for example, Iwanami Shoten Co., Ltd., issued May 25, 1990,
The techniques described in "Iwanami Encyclopedia of Information Science" P628 to P629 can be mentioned.

[0005]

However, in the prior art as described above, the fuzzy system itself needs to execute a plurality of instructions at a time in order to solve the limitation of the non-Neumann type description, that is, the high speed processing of the Neumann type. No consideration has been given to the possibility of parallel execution, which causes a problem in terms of processing speed.

That is, in the conventional fuzzy system, it is necessary to sequentially access the memory in order to obtain the membership value of the input signal, which requires a great deal of processing time.

On the other hand, in order to increase the processing speed of the system, it may be considered to increase the speed of the CPU, but it becomes expensive in terms of cost to make a high-speed fuzzy system.

Therefore, the present inventor has paid attention to a non-Neumann type computer capable of parallel execution and found that the merit of this non-Neumann type can be applied to a fuzzy system.

That is, an object of the present invention is to provide a membership value output circuit of a fuzzy system, which can make the fuzzy system non-Neumann type and obtain high processing speed.

The above and other objects and novel features of the present invention will be apparent from the description of this specification and the accompanying drawings.

[0011]

Among the inventions disclosed in the present application, a brief description will be given to the outline of typical ones.
It is as follows.

That is, the membership value output circuit of the fuzzy system of the present invention is a membership value output circuit of the fuzzy system which uses a membership function by fuzzy inference, and a plurality of A / D converters are combined to form a membership value output circuit. The reference voltage value of each A / D converter is set to the value at the end of the fuzzy region.

In this case, the membership value output circuit constructed by combining the plurality of A / D converters is integrated into one chip.

[0014]

According to the membership value output circuit of the fuzzy system described above, the digital value of the membership value for each fuzzy set can be instantaneously output in parallel by combining a plurality of A / D converters.

As a result, the processing speed can be increased by parallel processing and the processing time can be shortened.

[0016]

1 is a schematic configuration diagram showing an A / D converter in a membership value output circuit of a fuzzy system which is an embodiment of the present invention, and FIG. 2 is a membership function in the membership value output circuit of this embodiment. And FIG. 3 is a schematic configuration diagram showing a case where the membership value output circuit of this embodiment is applied to the detection of the tilt angle of the inverted pendulum.

First, the configuration of the A / D converter in the membership value output circuit of the fuzzy system of this embodiment will be described with reference to FIG.

The A / D converter in the membership value output circuit of this embodiment is a flash type A / D converter 1 using, for example, a membership function based on fuzzy inference, and the input signals are reference voltages V ₁ and V _2. N comparators 2 to be compared with, and these comparators 2
And a decoder 3 that decodes the output of the membership number, and the membership value is output in n bits.

The reference voltages V ₁ and V ₂ respectively provide an Upper reference voltage and a Lower reference voltage at the time of A / D conversion. The fuzzy region values of each fuzzy set are given to the reference voltages V ₁ and V ₂ , respectively. For example, as shown in FIG. 2, by giving x ₁ , x ₂ , x ₃ , x ₄ , the membership value for each fuzzy set, Small (x), Me
dim (x) and Big (x) are output respectively.

In this case, the adaptability of the reference voltages V ₁ and V ₂ of each A / D converter 1 to the Small region is V ₁ =
x ₂ and V ₂ = x ₁ , and V ₁ = x ₂ and V ₂ = x ₁ on the left side of the medium area and V ₁ = x ₄ on the right side,
V ₂ = x _3, and in the Big area, V ₁ = x ₄ , V
₂ = x ₃ .

Next, the operation of this embodiment will be described with reference to FIG. 3 when it is applied to the detection of the tilt angle of the inverted pendulum 4.

First, the inclination of the inverted pendulum 4 is mechanically transmitted to the angle sensor 5. The output of the angle sensor 5 is passed through the amplifier 6 and is, for example, 0.
Normalized to 1V.

In this case, the membership function is made up of three fuzzy sets as in FIG. 2, and each of them is Negative-Small, Zero, and Position.
With ive-Small, in order to output the membership value from these sets, 1, 2,
One A / D converter 1a-1d is required.

Then, U of the A / D converters 1a to 1d
pper-Reference-Voltage and L
power-Reference-Voltage and 1
One of the pressure than the partial reference voltage _{V, V 1, V 2,} V 3, V 4 retrieves the four voltage that is input to the A / D converter 1a~1d as in FIG.

For example, Negative-Small
For the A / D converter 1a in the region, voltages V ₁ and V ₂ ,
The voltage V is applied to one A / D converter 1b in the Zero region.
₁ and V ₂ , the voltage V on the other A / D converter 1c
₃ and V ₄ , A of the Positive-Small region
The voltages V ₃ and V ₄ are input to the / D converter 1d.

As a result, each of the A / D converters 1a-1
From d, the membership value of each fuzzy set is instantly output, processed by each of the next CPUs 7a to 7c, and processed by Negative-Small, Zero, Po.
The output of the passive-Small can be obtained.

Therefore, according to the present embodiment, the membership value for each fuzzy set can be output in parallel by using the four flash type A / D converters 1a to 1d. , For example, one A /
Neumann type processing by D converter 80236 CPU
It is possible to achieve a speedup of about 15,000 times as compared with the case of using, and thus the processing time can be shortened.

When the membership value output circuit by the A / D converters 1a to 1d is integrated into one chip,
As with the conventional LSI, mounting and design are easy, and the fuzzy system can be made compact.

Although the invention made by the present inventor has been specifically described based on the embodiments, the present invention is not limited to the embodiments and various modifications can be made without departing from the scope of the invention. Needless to say.

For example, the membership value output circuit of the present embodiment has been described by way of example when it is applied to the detection of the tilt angle of the inverted pendulum 4, but the present invention is not limited to the above embodiment. Image processing system,
It is widely applicable to voice recognition systems, expert systems, general household appliances, etc., and is particularly suitable for control systems that require high-speed response.

Further, the A / D converter 1 can be applied not only to the flash type but also to a successive approximation type or a dual type type A / D converter.

[0032]

The effects obtained by the typical ones of the inventions disclosed in the present application will be briefly described as follows.
It is as follows.

That is, by combining a plurality of A / D converters and setting the reference voltage value of each of these A / D converters to the value at the end of the fuzzy region, the digitalization of the membership value for each fuzzy set is performed. Since the values can be instantaneously output in parallel, the processing speed can be increased and the processing time can be shortened.

As a result, it is possible to obtain a membership value output circuit of a fuzzy system, which can be favorably applied to a control system that requires particularly high speed and can shorten the processing time.

[Brief description of drawings]

FIG. 1 is a schematic configuration diagram showing an A / D converter in a membership value output circuit of a fuzzy system which is an embodiment of the present invention.

FIG. 2 is an explanatory diagram showing an expression of a membership function in the membership value output circuit of this embodiment.

FIG. 3 is a schematic configuration diagram showing a case where the membership value output circuit of the present embodiment is applied to detection of an inclination angle of an inverted pendulum.

[Explanation of symbols]

 1, 1a-1d A / D converter 2 Comparator 3 Decoder 4 Inverted pendulum 5 Angle sensor 6 Amplifier 7a-7c CPU

Claims (2)

[Claims] 1. A membership value output circuit of a fuzzy system using a membership function by fuzzy inference, wherein a plurality of A / D converters are combined, and each reference voltage value of the A / D converters in a fuzzy region. A fuzzy system membership value output circuit, which is set to the end value and instantaneously outputs the digital value of the membership value for each fuzzy set in parallel. 2. The membership value output circuit of a fuzzy system according to claim 1, wherein the membership value output circuit configured by combining the plurality of A / D converters is integrated into one chip.