Data processing: artificial intelligence – Knowledge processing system – Knowledge representation and reasoning technique
Patent
1995-11-02
1998-10-20
Hafiz, Tariq R.
Data processing: artificial intelligence
Knowledge processing system
Knowledge representation and reasoning technique
706 8, 706 9, G06F 1518
Patent
active
058262512
DESCRIPTION:
BRIEF SUMMARY
CROSS REFERENCES TO RELATED APPLICATIONS
German Patent 43 08 083 Kiendi 1994 (Verfahren zur Erzeugung von Stellgro.beta.en am Ausgang eines Fuzzy-Reglers und Fuzzy-Regler hierfur)
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to a system of regulating or monitoring processes or industrial plants by generating an unambiguous scalar value of an output variable u or an unambiguous vector u of output variables at the output of a fuzzy unit as a function of the value of an input variable e or the values of several input variables combined to form a vector e, a fuzzy unit for implementing the said system and beneficial applications of the said system. The description that follows is based on fuzzy units revealing only one input variable e and one output variable u. The subject of the invention can, however, be applied analogously to fuzzy units revealing several input and several output variables.
2. Description of the Related Art
Signal-processing fuzzy units are familiar for instance in the for of fuzzy controllers. Their method of functioning is established, for exaple, in the article "Fuzzy Control" by H. Kiendl and M. Fritsch, in at Autoatisierungstechnik 41 (1993) 2, pp. A 5-A8 and can be described with the aid of FIG. 3 as follows: The controller input variable e is conducted to the fuzzification module 1. By means of the membership functions 2 on the input side, this module establishes for each of the linguistic values a.sub.i, e.g. "vanishing", "positively small" and "positively large" as applied to input variable e, the degree to which it is allocated to a currently applied value of input variable e. These values, also referred to as truth values w(e=a.sub.i) of the linguistic statements e=a.sub.i are conducted to a fuzzy logic module 3 containing linguistic rules, combined to form a rule basis fuzzy operators and an inference machine. Proceeding from these truth values, the logic operators assist in establishing to what degree the premises of the rules are fulfilled. The outcome of this determines the conclusions of the inference machine and--by means of the membership functions 4 on the output side--a resulting membership function .mu..sub.c (u). This function indicates for each value of the output variable u to what degree this is "favourable", i.e. recommended to serve as an output variable value as based on the conclusions of all rules. This membership function .mu..sub.c (u) is conducted to a defuzzification module 5. This then determines--for example employing the familiar center of gravity method--a resulting unambiguous value of the output variable u.
The method of functioning of conventional fuzzy controllers is also described in H. -P. Preu.beta.'s article "Fuzzy Control--heuristische Regelung mittels unscharfer Logik", atp Automatisierungstechnische Praxis 1992, 4 pp. 176-184 and 5, pp. 239-246. The internal makeup of a fuzzy controller described in this article on the basis of FIG. 12 and 13 (q.v.) corresponds fully to the structure depicted in FIG. 3 of this patent application. Differences occur only as a result of the terminology and graphic layout. Thus, in FIG. 3, the output variables of the fuzzification unit, which are truth values in conformity with Preu.beta.'s article (q.v., p. 240, left-hand column, sentence 1) are designated w(e=a.sub.i) for the purposes of illustration. Furthermore, in FIG. 3, the function of processing the various rules (Preu.beta.=inference) as well as that of combining the conclusion of all rules (Preu.beta.=composition) are incorporated in the fuzzy logic function module. Likewise in FIG. 3, the membership function to be defuzzified, which results on the output side from the combined interaction of all rules, is indicated by use of the symbol .mu..sub.c (u). This designation is absent in FIGS. 12 and 13 of Preu.beta.'s article. Nevertheless, this function is highlighted in the right-hand illustration section of FIG. 12 of the article by being underscored in black. Finally, in FIG. 3, the "module for membership functions on
REFERENCES:
patent: 5131071 (1992-07-01), Tsutsumi et al.
patent: 5179625 (1993-01-01), Hisano
patent: 5255344 (1993-10-01), Takagi et al.
patent: 5285376 (1994-02-01), Struger et al.
patent: 5295226 (1994-03-01), Yamakawa
patent: 5303331 (1994-04-01), Namba
patent: 5376611 (1994-12-01), Iwashita et al.
patent: 5408584 (1995-04-01), Hirakura et al.
patent: 5425131 (1995-06-01), Basehore
patent: 5495574 (1996-02-01), Miyazawa et al.
patent: 5600757 (1997-02-01), Yamamoto et al.
Zhang et al; IEEE Int. Conf. on Neural Networks, ICNN 1988, pp. I-231-I-238.
Patent Abstracts of Japan, vol. 16, No. 186 (p-1347) dated May 17, 1992.
Kacprzyk et al; 3rd Int. Conf. on Information Processing and Management of Uncertainty in Knowledge-Based Systems, IPMU 1990, pp. 424-430.
Pfluger et al; IEEE Int. Conf. on Fuzzy Systems, ICFS 1992, pp. 717-723.
Hafiz Tariq R.
Shah Sanjiv
LandOfFree
System of controlling or monitoring processes or industrial plan does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System of controlling or monitoring processes or industrial plan, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System of controlling or monitoring processes or industrial plan will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-259958