JPH0293945A - Fuzzy medium system - Google Patents

Abstract

PURPOSE:To prevent a device from dropping its working rate by providing a medium such as a card with personal sense/liking as a membership function and a fuzzy rule expressing fuzzy concept. CONSTITUTION:Respectively different personal sense/liking is stored in an external storage medium as fuzzy system data such as a membership function and a fuzzy rule, and at the time of using a fuzzy inference device, the fuzzy inference device is allowed to read out the contents of the external storage medium to execute inference corresponding to the personal sense/linking. Thereby, the device can be easily used for its user and the accuracy of inference can be improved. Since personal fuzzy system data are stored in the external storage medium and carried by each person, it is unnecessary to expand a memory in the device and personal secrecy can be prevented from being leaked.

Description

DETAILED DESCRIPTION OF THE INVENTION (a) Industrial Application Field The present invention relates to a fuzzy medium system in which fuzzy system data expressing fuzzy inference concepts is stored in an external storage medium such as a card.

(b) Conventional technology By using fuzzy reasoning in the control section of a device operated by a human, the user's ambiguous sensory expression (operation)
The controlled device can be controlled by A fuzzy inference device has a membership function that has a standard sense for stimuli from the outside world, and also has constraints on it. llft is also set up with a standard membership function.

(C) Problems to be solved by the invention By the way, the degree of human sensitivity to stimuli from the outside world;
For example, the sensation of stimuli such as "bright-dark" differs from person to person, but in conventional fuzzy inference devices, standard sensations and control amounts are set as described above, and these can be changed for each person. Since this method was not possible, there was a drawback in that humans had to adjust the settings of the fuzzy inference device to some extent.

On the other hand, when trying to adapt the operation of a fuzzy inference device to the senses and preferences of individual users,
Preferences must be stored in memory, which poses a problem in that the amount of information becomes enormous and the system becomes large-sized, and there is also a risk that personal information may be leaked.

Furthermore, the method in which the user sets his or her own feelings and preferences on the device each time the device is used is time-consuming and has the problem of lowering the operating rate of the device.

This invention was made by focusing on these conventional problems, and solved the above problems by incorporating individual feelings and preferences into a medium such as a card as a membership function or fuzzy rule that expresses a fuzzy concept. It aims to provide a fuzzy media system.

[Means for Solving Problem d1 The first invention according to this application includes a storage medium storing fuzzy system data including membership functions and fuzzy rules, and a system that receives the storage medium and imports the fuzzy system data. A fuzzy inference device that performs fuzzy inference using data.

A second invention according to this application includes a storage medium storing fuzzy system data including membership functions, fuzzy rules, and parameters for recognizing a user's voice pattern; and a fuzzy inference device that takes in the parameters, receives voice input from the user based on the parameters, and performs fuzzy inference on the voice input data using the fuzzy system data.

(e) Function The fuzzy medium system of the present invention stores the feelings and preferences of the individual user who uses the system including the fuzzy inference device in an external storage medium such as a card as membership functions and fuzzy rules that express fuzzy concepts. Then, when this user uses the system, he connects this external storage medium to the fuzzy inference device and operates it according to his own sense.

Here, the fuzzy rule is 1f(x, =A and x, =8=) ther+
It is a proposition expressed in the form (y = Z), and (x , =
A andx2・B...) is called the antecedent part, and (y=Z) is called the consequent part. In the antecedent part, intuitive conditional sentences are set, such as r if it is dark outside (brightness outside - dark), and if the room is a little dark (brightness of the room = a little dark). The subject part says, ``Make the lighting a little brighter (lighting - make it a little brighter)
” etc. are expressed intuitively. Hereinafter, one known method for outputting inference results according to fuzzy rules will be explained with reference to FIG. 7. FIGS. 7A and 7B show membership functions corresponding to two terms of the antecedent part, Same figure (
C) represents the membership function corresponding to the consequent. Here, two membership functions for the antecedent part are shown, but as the number of variables input as conditions for the antecedent part increases, the number of membership functions increases accordingly. In each figure, the horizontal axis represents the value of the variable, and the vertical axis represents the membership value (degree of belonging).

Now, if the value of the variable
See A). Further, if the value of the variable x2 of the second item of the antecedent part is x, I, then the degree of affiliation is 0.3 (see (B) in the same figure). In such a case, the smallest value among the respective degrees of membership is taken in the fuzzy operation.

That is, in the above example, a degree of affiliation of 0.3 is selected. Next, such inference is performed for one rule that truncates the membership function corresponding to Z at the height of 0.3 above, and the logical sum of the inference results for all rules is shown in Figure 4 (D
) to obtain the inferred value shown in the shaded area. Usually 1
In order to determine the inferred value and control the operation using this value, the center of gravity y'' of this shaded area is determined and output as the determined value of the inference.

In the above inference method, the logical product operation rule of the degree of belonging to the antecedent part (operation to select the smaller degree of membership) and the logical sum operation rule of the trapezoidal part to the consequent part are
This is called the ax rule.

In order to make such an inference, fuzzy rules are needed to determine the membership functions that make up each term and how to combine these membership functions, but the sense of brightness etc. differs from person to person, and , the response to this also differs depending on each person's preferences. By storing these feelings and preferences in an external storage medium, it is now possible to respond appropriately to each person without increasing the memory size of the device.

(r) Embodiment FIG. 1 is an external view of a fuzzy data card used in an embodiment of the present invention. This card is a laser card, and a laser storage section 2 is formed on the surface of the card 1.

Of the storage units 2, 2a is a fuzzy system data storage unit consisting of membership functions and fuzzy rules;
b is a voice parameter storage section in which parameters for recognizing the owner's voice are stored.

FIG. 2 is a block diagram of an online terminal device of an online stock trading system which is an embodiment corresponding to the first invention of this application. The microcomputer 10, which is a control device, includes an inference section 12. Fuzzy card interface 11. A CRTI3 keyboard 141 and communication interface 15 are connected. A communication interface is connected to an online stock trading host computer. The inference unit 12 is a device that performs fuzzy inference regarding the buying and selling of stocks based on system data read from the fuzzy card and general information (stock prices, economic trends, etc.) received from the host computer. The configuration is shown below. In FIG. 1A, the inference section 12 consists of a plurality of fuzzy calculation sections 30 and a defuzzifying section 31. In addition, in the same figure (B), the fuzzy operation unit 30
is the same number of MFCs (membership function generation circuits) 322 as the number of terms in the antecedent part, and the conditional AND part 33. It has an MFC 34 corresponding to the consequent part and a conclusion logical product part 35. Variables and membership function selection signals are input to the MFC 32 of the antecedent part in order to calculate each term of the antecedent part. The output of this calculation is input to the conditional AND section 23, where it is ANDed (the minimum value is selected) and the result is inputted to the conclusion AND section 35. The MFC 34 generates a membership function of the consequent part for inferring the probability (an index indicating the extent to which stocks should be bought and sold), and manually inputs it to the conclusion logical product part 35. The conclusion logical product unit cuts the membership function value input from the MFC 34 with the value input from the conditional logical product unit 33, and outputs it to the defuzzify unit 31 as the calculated value of this fuzzy calculation unit. This fuzzy operation part is based on the rules (if...,...then) set for inference.
~) are provided.

In the same figure (C), the defuzzify section 31 consists of a logical sum section 37 and a definite value calculation section 38, and the calculated values manually inputted from the fuzzy calculation section 30 are synthesized in the logical sum section 37 and ambiguous (width) (possibility) is determined. This ambiguous possibility is input to the determined value calculating section 38, and its center of gravity is determined.
Output to 3.

Fuzzy Card 1 stores feelings about the stock price (P) such as "high (PL), slightly high (ps), normal (ZR), slightly low (NS), low (NL)" and the rise and fall of the stock price (P').
Feelings about “increased (PL), slightly increased (ps)”
), flat (ZR), slightly down (NS), and down (NL) J are stored as membership functions, and the probability of buying or selling the stock (
Recommendation level (R)) is “Must buy (PL), Must buy (PL)”
PM), I'd rather buy it (PS), I can't say either way (ZR), I'd rather sell it (NS), I should sell it (NM), I think it should shine (NL)J
It is stored as a membership function. This membership function is different for each card (for each user), and is designed to reflect each person's way of thinking about stock buying and selling. Examples of expansion and flexibility of this membership function are shown in FIGS. 4(A) to 4(C). Methods for storing member functions include a method for storing function formulas and a method for storing plotting data.

Further, the fuzzy rules stored in the fuzzy card 1 include, for example, r■ stock price is low, ■ stock price rises slightly,
■ If the outlook for the industry is good, ■ Must buy, defined as 1, and the card has rP (NL).
It is stored as nP'(PS)nF(PL)→R(PL)J. Alternatively, a matrix representing all the states may be stored.

FIG. 5 is a flow chart showing the operation of the microcomputer 10. At nl, the inserted fuzzy card is read and the stored membership functions and fuzzy rules are set in the inference section 12 (n2). Next, data from the host computer (data that can be shared, such as stock prices).

- social situation of boat fishing, etc.) and performs fuzzy inference based on these (n4). Based on this inference, if there is a stock that is a candidate for the user to buy or sell, it is displayed on the CRT (n5). When a user buys or sells this stock, n
Press l to confirm.

Then, the microcomputer 10 notifies the host computer of this fact and makes a stock purchase/sale request. If there are no trading candidates (n5) or if you cancel the trading of the displayed stock, use n3 to perform the next fuzzy inference.
Return to

FIG. 6 is a block diagram of a voice operation control device which is an embodiment of the second invention of this application. This device has a fuzzy card interface 20, and data read from the fuzzy card I is input to an inference section 21 and a speech recognition device 23.

Fuzzy system data is manually input to the fittt1 section 21, and voice parameters are input to the voice recognition device 23. A microphone 24 is connected to the voice recognition device 23 and listens to voice commands from the operator (the person who inserted the fuzzy card). An operation control section 22 is connected to the inference section 21 and controls the operation of the lighting 25 and the trolley 26 that are the objects of operation based on the operation amount input from the inference section 21 . If the operator says in voice that 1 light is bright and the trolley is far away from the light, move the trolley a little closer to the light and give a command such as j to the microphone 24, the voice recognition device 23 will recognize this and change the content. is encoded and sent to the inference section 21. The inference section 21 performs fuzzy inference on this input to determine the brightness of the illumination and the amount of movement of the cart, and sends it to the operation control section 22. The operation control unit 22 controls the lighting 2 based on the input control 4TOII.
51 controls the operation of the truck 26.

In this embodiment, the fuzzy card 1 is configured with a laser card, but the same function can be achieved with an IC card or a magnetic card. Further, in the first embodiment, the inference section 12 is provided separately from the microcomputer 10 which is the control device, but if the fuzzy inference is executed by a program, the microcomputer 10 can also serve as the fuzzy inference section. .

fg) Effects of the invention According to the invention, the different feelings and preferences of each person are stored in an external storage medium as fuzzy system data such as membership functions and fuzzy rules, and a fuzzy inference device is used. By having the fuzzy reasoning device read an external storage medium, one can understand one's sense of self.
It is possible to make inferences tailored to your preferences. This makes it possible to make the device easier for the user to use and to improve the accuracy of inference.

In addition, by storing each person's fuzzy system data in an external storage medium that each person owns, there is no need to enlarge the memory of the device, and there is no fear of leaking personal secrets.

Furthermore, by simultaneously storing parameters for user's voice recognition in this external storage medium, voice input can be accepted and operations can be made simpler.

[Brief explanation of drawings]

Fig. 1 is an external view of a fuzzy card used in an embodiment of this invention, Fig. 2 is a diagram showing an online terminal device of an online stock trading system which is an embodiment of invention 10 of this application, and Fig. 3 (A) to (C) are diagrams showing the configuration of the inference section of the online terminal device, and Figures 4 (A) to (C) are diagrams showing examples of membership functions used in the online stock trading system. Figure 5 is a flowchart showing the operation of the microcomputer of the online terminal device.
The figure shows an embodiment of the second invention of this application; 1-fuzzy card, 2-laser storage section, 2a-fuzzy system data storage section, 2b-speech parameter storage section.

Claims (2)

[Claims] (1) A fuzzy medium consisting of a storage medium that stores fuzzy system data including membership functions and fuzzy rules, and a fuzzy inference device that receives this storage medium, takes in the fuzzy system data, and executes fuzzy inference using this data. system. (2) A storage medium that stores fuzzy system data including membership functions and fuzzy rules, and parameters for recognizing the user's voice pattern; a fuzzy inference device that accepts voice input from a user based on the above and performs fuzzy inference on the voice input data using the fuzzy system data.