WO1994008286A1 - Procede de generation d'une grandeur de reglage de sortie nette a la sortie d'un circuit de reglage flou - Google Patents

Procede de generation d'une grandeur de reglage de sortie nette a la sortie d'un circuit de reglage flou Download PDF

Info

Publication number
WO1994008286A1
WO1994008286A1 PCT/EP1993/002597 EP9302597W WO9408286A1 WO 1994008286 A1 WO1994008286 A1 WO 1994008286A1 EP 9302597 W EP9302597 W EP 9302597W WO 9408286 A1 WO9408286 A1 WO 9408286A1
Authority
WO
WIPO (PCT)
Prior art keywords
rule
fuzzy
degree
output
sharp
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/EP1993/002597
Other languages
German (de)
English (en)
Inventor
Frank Hubert
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
ZETEC GmbH
Original Assignee
ZETEC GmbH
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by ZETEC GmbH filed Critical ZETEC GmbH
Publication of WO1994008286A1 publication Critical patent/WO1994008286A1/fr
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N7/00—Computing arrangements based on specific mathematical models
    • G06N7/02—Computing arrangements based on specific mathematical models using fuzzy logic
    • G06N7/04—Physical realisation
    • G06N7/043—Analogue or partially analogue implementation
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N7/00—Computing arrangements based on specific mathematical models
    • G06N7/02—Computing arrangements based on specific mathematical models using fuzzy logic
    • G06N7/04—Physical realisation

Definitions

  • the invention relates to a method for generating a sharp output manipulated variable at the output of a fuzzy control loop, the fuzzy inference machine of which depends on measured process variables in the entire range
  • 20 of the manipulated variables provides one or more fuzzy sets of possible manipulated variables with different degrees of membership within each fuzzy set, each fuzzy set belonging to a rule processed in the inference machine and the degree of fulfillment of the associated
  • the individual are ne delivered fuzzy quantities with the associated degrees of fulfillment are shown in a diagram, which shows the different degrees of membership of the individual fuZzy quantities cut off by the respective degree of fulfillment as a function of the increasing manipulated variables in the entire manipulated variable range.
  • the manipulated variables form the abscissa and the associated degree of fulfillment and degree of membership the ordinate.
  • the fuzzy sets shown in this way usually have the shape of trapezoidal surfaces or triangular surfaces with more or • less steep flanks below the membership function. Under certain circumstances, the surfaces can also be delimited in a curved manner on the sides.
  • the so-called maximum mean value method is known, in which the mean value forms the output value for rules with the same maximum degree of fulfillment.
  • the so-called accumulation method in which the current value of the manipulated variable is reduced or increased by a value that is fixed in a rule, regardless of the degree of fulfillment of the rule.
  • the invention proposes, based on the method of the type mentioned at the beginning, that the degrees of membership in the various fuzzy quantities are determined step by step for each possible manipulated variable over the entire range of the manipulated variables, starting with the minimum or the maximum manipulated variable and that the manipulated variable is then selected and output as a sharp initial manipulated variable at which the degree of membership in the fuzzy set relevant for the control reaches its maximum for the first time.
  • the left or right upper corner point of the trapezoid is used as the output manipulated variable for determining a sharp manipulated variable.
  • the selected manipulated variable can be as close as desired to the limit of the actuating range.
  • simple size comparisons are sufficient, which can be implemented in the simplest way by means of current mirror circuits (when using analog evaluation technology) or through simple digital comparison registers (when using digital evaluation technology).
  • An advantageous development of the invention provides that the relevance of the different rules is determined in a priority order, which is in the Selection of the rule to be used in each case for the output of the sharp output manipulated variables is additionally taken into account. This makes it possible to determine from the outset a ranking of the rules to be used, with the individual variable being able to determine the manipulated variable to be output very quickly and with little effort. By specifying the order of precedence, a further dimension can easily be added to a very complex set of rules.
  • the rule relevant in the priority order below is used for determining the sharp output manipulated variable if the degree of fulfillment in the rule preceding the priority order is zero or below a certain one , predetermined value.
  • the rule relevant below in the order of priority can be used for determining the sharp output manipulated variable if, in the case of the rule preceding in the order of priority, the degree of fulfillment is below the value of the following rule, another subsequent rule or below the maximum the degree of fulfillment of all the following rules.
  • the method according to the invention makes a selection from the generation of the sharp output manipulated variable the different rules, without being fixed to a predetermined order of priorities. This possibility can also be of great value in various control processes.
  • the step intervals are expediently large over the entire range of the manipulated variables.
  • a clock generator of the simplest type can be used for the ascending or descending passage of the various manipulated variables.
  • FIG. 1 shows schematically a fuzzy control loop with a defuzzifier according to the invention
  • 2b schematically shows the process of defuzzifying using fuzzy set diagrams with priority order
  • 3 shows a block diagram of a defuzzifier according to the invention with digital evaluation technology
  • 4 shows a block diagram of a defuzzifier according to the invention with analog evaluation technology.
  • the fuzzy control circuit shown schematically in FIG. 1 is used to control a technical process, which is denoted overall by reference number 1.
  • Process sensors 2 are assigned to process 1, by means of which certain process variables relevant for the control are measured and processed for input into a fuzzy inference machine 3.
  • the measured and processed process variables are fed to the fuzzy inference machine, in which they are processed on the basis of rules stored there.
  • the fuzzy inference machine delivers fuzzy output variables in the form of fuzzy quantities, depending on the cyclically measured process variables.
  • Each delivered fuzzy set has a specific rule that is processed in the inference machine.
  • the fuzzy sets belonging to the individual rules are characterized by the degree of membership characteristic of the respective rule and are cut off at the top by the degree of fulfillment of the associated rule currently measured in the process.
  • a sharp manipulated variable is then obtained in a defuzzifier 4 which controls an actuator 5, which influences the process to be regulated accordingly.
  • the invention is now concerned with the process of de-fuzzification, ie with the method for obtaining sharp output manipulated variables for the actuation of the actuator 5. The method according to the invention is first explained with reference to the sequences of diagrams shown in FIGS. 2a and 2b .
  • the manipulated variables increasing over the manipulated variable range form the abscissa and the associated degree of fulfillment or degree of membership the ordinate, which can assume values between zero and one.
  • the various fuzzy sets supplied by the reference machine 3 are each represented, for example, as trapezoidal surfaces, the inclined edges of which represent the membership function characteristic of the respective fuzzy set or the respective rule.
  • the flanks can run more or less steeply or can have a shape that deviates from the straight line. For example, they could run in curves, depending on the special properties of the membership function of the rule concerned.
  • the areas belonging to each rule are cut off towards the top by the degree of fulfillment of the rule in question, which appears in the diagrams as a horizontal line.
  • fuzzy sets belonging to the individual rules appear in the diagram as trapezoids or similar figures cut off by a horizontal line at the top, the individual ones depending on the degree of fulfillment determined Figures are cut more or less high.
  • the procedure for determining a sharp output manipulated variable based on the fuzzy quantities supplied by the inference machine is as follows:
  • the manipulated variable determines the degrees of membership or degree of fulfillment in the different fuzzy sets.
  • the manipulated variable is then selected and output as a sharp output manipulated variable at which the degree of membership reaches its maximum for the first time.
  • the first sharp manipulated variable is the abscissa of the upper left corner of the trapezoid belonging to rule 1.
  • the entire range of manipulated variables can be run through from top to bottom, starting with the maximum manipulated variable.
  • the abscissa of the top right corner of the fuzzy set belonging to rule 4 is determined as the second sharp manipulated variable, ie in this case the maximum possible manipulated variable.
  • manipulated variables that lie at the limits of the actuating range can also be generated without further ado.
  • the second diagram of FIG. 2a assumes that the degree of fulfillment of rule 4 has become smaller compared to the first diagram, while the degree of fulfillment of rule 2 has increased and has become as large as the degree of fulfillment of rule 1.
  • the method according to the invention determines the abscissa of the top left corner point of the fuzzy quantity belonging to rule 1 as the first sharp manipulated variable when the manipulated variables run from bottom to top and when the Control variables from top to bottom, the abscissa of the top right corner of the fuzzy set belonging to rule 2 as a second sharp control variable.
  • the degree of fulfillment of rule 1 has decreased.
  • the degree of fulfillment of rule 3 has become as great as the degree of fulfillment of rule 2.
  • the method according to the invention determines the abscissa of the top left corner point when the manipulated variables pass from bottom to top Fuzzy set belonging to Rule 2 as the first sharp manipulated variable and, when the manipulated variable range runs from top to bottom, the abscissa of the top right corner of the fuzzy set belonging to Rule 3 as the second sharp manipulated variable.
  • the degree of compliance with rule 4 has become greater than the degree of compliance with all other rules.
  • the method according to the invention determines the abscissa of the upper left corner of the fuzzy set belonging to rule 4 as the first sharp manipulated variable and the abscissa of the upper right corner as the second sharp manipulated variable.
  • Point of the fuzzy set belonging to rule 4 ie here again similar to the first diagram of FIG. 2a, the maximum possible manipulated variable.
  • This sequence of priorities means that competing rules with the same degree of fulfillment use the rule which has the higher priority rank to determine the first sharp manipulated variable or the second sharp manipulated variable.
  • this has the result that, in deviation from the first diagram of FIG. 2a, the abscissa of the top right corner point of the fuzzy set belonging to rule 1 is output as the second sharp manipulated variable.
  • the top right corner of Rule 4 does not apply despite the same degree of fulfillment because Rule 4 has a lower priority rank than Rule 1.
  • the abscissa of the top left corner of the fuzzy set belonging to rule 2 is output as the first manipulated variable.
  • the upper corner point of the fuzzy set belonging to rule 1 is not considered here because, despite the same degree of fulfillment, rule 2 has a higher priority than rule 1.
  • the abscissa of the upper right corner of the fuzzy set belonging to rule 2 is output as the second manipulated variable.
  • the right upper corner of the fuzzy set belonging to rule 3 is not considered here, despite the same degree of fulfillment, because rule 3 has a lower priority than rule 2.
  • the actuating variables determined correspond to those of FIG. 2a despite the order of priority.
  • Rule 4 used for determining the manipulated variable has the lowest priority here. However, their degree of fulfillment is higher than with all rules with a higher priority. For this reason, only the fuzzy set belonging to rule 4 is used to determine the manipulated variables.
  • This evaluation device has an N-bit binary counter, which specifies the operating cycles and determines whether the manipulated variable range of 13
  • the respective current degrees of fulfillment E1, EI ... EN and the maximum degree of fulfillment EMAX determined in each case, delivered by the inference machine, are entered into digital MF memories with comparators.
  • Each MF memory with comparators belongs to a rule of the fuzzy control algorithm processed in the inference machine.
  • the individual MF memories with comparators are connected to a digital module designated as EVALUATION LOGIC, which compares the values supplied by the various MF memories and comparators and which are explained according to the above with reference to FIGS. 2a or 2b
  • the block EVALUATION LOGIC then outputs corresponding control signals to two output registers N-BIT-LATCH, likewise clocked by the binary counter, for the output of the lower sharp manipulated variable Su and / or the upper sharp manipulated variable So.
  • analog signals VKI, VKi ... VKN are evaluated according to the criteria explained above with reference to FIGS. 2a and 2b and converted into analog signals KT , which are proportional to the abscissas of the selected upper corner points of the relevant fuzzy sets (trapezoids).
  • These signals are clocked alternately for the upper and the lower manipulated variable, sorted by appropriately clocked timers and converted into analog continuous signals VSu (for the lower manipulated variable) and VSo (for the upper manipulated variable).

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Physics & Mathematics (AREA)
  • Computational Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Molecular Biology (AREA)
  • Fuzzy Systems (AREA)
  • Biomedical Technology (AREA)
  • Algebra (AREA)
  • Artificial Intelligence (AREA)
  • Health & Medical Sciences (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Automation & Control Theory (AREA)
  • Mathematical Analysis (AREA)
  • Mathematical Optimization (AREA)
  • Pure & Applied Mathematics (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Feedback Control In General (AREA)

Abstract

Un procédé permet de générer une grandeur de réglage de sortie nette à la sortie d'un circuit de réglage flou constitué d'un processeur (1), de capteurs (2) de valeurs de mesure, d'une machine d'inférence (3), d'un éliminateur de flou (4) pour générer des grandeurs de réglage nettes et d'un organe de réglage (5). Selon ce procédé, la machine d'inférence floue (3) fournit pour chaque règlage une quantité floue de grandeurs de réglage possibles délimitées vers le haut par le dégré de conformité avec le réglage correspondant. Afin de générer avec l'éliminateur de flou une grandeur de réglage de sortie nette avec aussi peu que possible de calculs et de matériel, le degré de conformité de chaque grandeur de régglage est déterminé graduellement dans les différentes quantités floues de toute la fourchette de grandeurs de réglage, à commencer par la grandeur de réglage minimale ou maximale, puis la grandeur de réglage dont le degré de conformité atteint pour la première fois son maximum dans la quantité floue pertinente pour le réglage et sélectionnée et transmise comme grandeur de réglage nette.
PCT/EP1993/002597 1992-09-30 1993-09-24 Procede de generation d'une grandeur de reglage de sortie nette a la sortie d'un circuit de reglage flou Ceased WO1994008286A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
DE4232752A DE4232752C1 (de) 1992-09-30 1992-09-30 Verfahren zur Erzeugung einer scharfen Ausgangsstellgröße am Ausgang eines Fuzzy-Regelkreises
DEP4232752.0 1992-09-30

Publications (1)

Publication Number Publication Date
WO1994008286A1 true WO1994008286A1 (fr) 1994-04-14

Family

ID=6469213

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/EP1993/002597 Ceased WO1994008286A1 (fr) 1992-09-30 1993-09-24 Procede de generation d'une grandeur de reglage de sortie nette a la sortie d'un circuit de reglage flou

Country Status (2)

Country Link
DE (1) DE4232752C1 (fr)
WO (1) WO1994008286A1 (fr)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
DE19516627A1 (de) * 1995-05-05 1996-11-07 Ranco Inc Verfahren und Vorrichtung zur Regelung eines Prozesses

Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH04223578A (ja) * 1990-12-25 1992-08-13 Res Dev Corp Of Japan デファジネーション回路

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5167005A (en) * 1988-08-19 1992-11-24 Research Development Corporation Of Japan Fuzzy computer

Patent Citations (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH04223578A (ja) * 1990-12-25 1992-08-13 Res Dev Corp Of Japan デファジネーション回路

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
BRUBAKER: "Fuzzy-logic basics: intiutive rules replace complex math", EDN ELECTRICAL DESIGN NEWS., vol. 37, no. 13, 18 June 1992 (1992-06-18), NEWTON, MASSACHUSETTS US, pages 111 - 116, XP000301373 *
LEE: "Fuzzy Logic in Control Systems: Fuzzy Logic Controller, Part II", IEEE TRANSACTIONS ON SYSTEMS, MAN AND CYBERNETICS, vol. 20, no. 2, April 1990 (1990-04-01), NEW YORK US, pages 419 - 435, XP000164959 *
PATENT ABSTRACTS OF JAPAN vol. 16, no. 577 (P - 1460) *
Proceedings 'Clear Applications of Fuzzy Logic',17 October 1991, Delft, Netherlands, pp. 57-82; Hellendoorn: 'Fuzzy Logic and Fuzzy Control' *

Also Published As

Publication number Publication date
DE4232752C1 (de) 1994-06-09

Similar Documents

Publication Publication Date Title
DE69420734T2 (de) Mehrbereichs-fuzzy-logik regelungssystem mit hilfsvariablen
EP2185980B1 (fr) Procédé de commande et/ou de réglage assisté par ordinateur à l'aide de réseaux neuronaux
EP0663632B1 (fr) Méthode et dispositif pour la comande d'un processus
DE69315423T2 (de) Neuro-Pid Regler
DE4008510C2 (de) Regeleinrichtung mit Optimal-Entscheidungseinheit
DE69434487T2 (de) Methode und vorrichtung zur fuzzy logicsteuerung mit automatischem abstimmverfahren
DE19606480C2 (de) Verfahren und Anordnung zur Adaption eines Fuzzy-Reglers
EP2112568A2 (fr) Procédé de commande et/ou réglage assistées par ordinateur d'un système technique
WO2009056416A1 (fr) Procédé pour l'exploration assistée par ordinateur des états d'un système technique
DE10261727A1 (de) Steuersystem in Fuzzy-Logik für ein Rad eines Kraftfahrzeugs und Verfahren zum Implementieren einer Fuzzy-Logikeinheit für derartiges Rad eines Kraftfahrzeuges
EP0706680B1 (fr) Systeme regulateur, en particulier pour un processus non lineaire variable dans le temps
DE4439225B4 (de) Defuzzifizierungsverfahren für Fuzzy Logic-Ausgangssignale und Reglervorrichtung hierzu
EP0998700A2 (fr) Procede pour la generation de trajectoires d'accostage, utilisable pour le guidage d'un vehicule vers une trajectoire cible predefinie
WO2000022487A1 (fr) Dispositif de regulation pour la regulation d'un systeme asservi a plusieurs grandeurs reglees couplees
DE4409169C2 (de) Verfahren und Anordnung zur Fuzzy-Regelung
DE69126052T2 (de) Neuronales Netzwerk
DE69808779T2 (de) Verfahren zum vermeiden von windup in pid-reglern durch nichtlineare verstärkung
EP0752630A1 (fr) ContrÔleur auto-ajustable et méthode pour l'auto-ajustement du contrÔleur
WO1994008286A1 (fr) Procede de generation d'une grandeur de reglage de sortie nette a la sortie d'un circuit de reglage flou
DE2836308A1 (de) Verfahren zur regelung der gleichstromuebertragung bei gleichstrom-hochspannungsfernleitungen und regeleinrichtung zur ausfuehrung des verfahrens
WO1994028470A1 (fr) Dispositif de regulation
DE102021116388A1 (de) Verfahren zum Optimieren eines Reglers und Regelsystem mit einem derart optimierten Regler
DE4202699A1 (de) Vorrichtung zur steuerung des lenkwinkels
DE69426858T2 (de) Fuzzy-Regler
DE102009009328B4 (de) Verfahren zur Reduzierung von Strukturschwingungen und Regelungseinrichtung hierzu

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A1

Designated state(s): JP KR US

AL Designated countries for regional patents

Kind code of ref document: A1

Designated state(s): AT BE CH DE DK ES FR GB GR IE IT LU MC NL PT SE

DFPE Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101)
121 Ep: the epo has been informed by wipo that ep was designated in this application
122 Ep: pct application non-entry in european phase