Data processing: artificial intelligence – Fuzzy logic hardware – Having function calculator
Reexamination Certificate
1997-02-26
2001-02-13
Hafiz, Tariq R. (Department: 2762)
Data processing: artificial intelligence
Fuzzy logic hardware
Having function calculator
C706S008000, C706S001000, C706S003000
Reexamination Certificate
active
06188998
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
This invention relates to fuzzy logic, and more particularly to methods for storing membership functions, as well as processors of the fuzzy type for implementing the method.
2. Discussion of the Related Art
A survey of both theoretical and practical aspects of fuzzy logic processing is to be found in M. M. Gupta and T. Yamakawa, “FUZZY COMPUTING”, North-Holland, 1988.
One of the most critical aspects in the design of architectures for fuzzy logic processing is he representation of the linguistic terms that collectively form the so-called dictionary of terms, which is in practice the equivalent of representing the functions of membership to such terms.
The techniques for representing membership functions can be divided into two general classes: those which take no account of the function shape, and usually provide for function sampling, and those which only apply to particular shapes, such as triangular, trapezoidal, and other shapes of the functions.
Some methods of storing membership functions from the former class are described in Patent Applications EP 628 902 and EP 628 903, wherein only the respective non-null samples of each function are stored.
A method of storing membership functions from the latter class is described in Patent Application EP 675 431, wherein an apex and the slopes on the left-hand and right-hand sides of the apex are stored for each function. A circuit for computing such functions, starting from their storing in the manner outlined above, is described in Patent Application EP 675 430.
It will be appreciated from the foregoing concise review of prior methods that considerable work has been devoted to minimizing the storage capacity requirements for each individual membership function.
SUMMARY OF THE INVENTION
An object of the present invention is to minimize the amount of storage capacity required to store a whole set of membership functions, namely all those required to implement a fuzzy type of control.
An embodiment of the invention is directed to a method for storing, in a fuzzy type of processor, one or more natural membership functions that respectively correspond to one or more natural variables, each natural variable being defined within a respective natural universe of discourse having a respective lowest natural value and a respective highest natural value. The method includes the steps of (a) obtaining one or more absolute membership functions by normalizing the one or more natural membership functions through respective normalization coefficients such that the one or more natural membership functions are defined within a same predetermined absolute universe of discourse having a lowest absolute value and a highest absolute value, and (b) storing the one or more absolute membership functions and the normalization coefficients in the processor such that identical absolute membership functions are stored only once.
Another embodiment of the invention is directed to an apparatus for storing membership functions. The apparatus includes a storage circuit for storing membership functions defined within a same predetermined universe of discourse, the storage circuit being arranged not to store information about universes of discourse.
Another embodiment of the invention is directed to a fuzzy type of processor, including a processor input for inputting a plurality of natural variables; a processor output for outputting at least one control parameter; a fuzzyfier unit having an input and an output; a rule computation unit having an input coupled to the output of the fuzzyfier unit, and an output; a defuzzyfier unit having an input coupled to the output of the rule computation unit, and an output coupled to the processor output; and a normalization unit having an input coupled to the processor input, and an output coupled to the input of the fuzzyfier unit.
Another embodiment of the invention is directed to a fuzzy type of processor, including a processor input for inputting a plurality of natural variables; a processor output for outputting at least one control parameter; a fuzzyfier unit including a storage circuit for storing membership functions defined within a same predetermined universe of discourse, an input coupled to the processor input, and an output, wherein the storage circuit is arranged not to store information about universes of discourse; a rule computation unit having an input coupled to the output of the fuzzyfier unit, and an output; and a defuzzyfier unit having an input coupled to the output of the rule computation unit, and an output coupled to the processor output.
An embodiment of the present invention utilizes, for the purpose of fuzzy logic processing, a “dictionary of absolute terms”, i.e. of terms having the same universe of discourse and on this account referred to as being “absolute”. In this way, redundancies, that appear each time that complex fuzzy controls are implemented, can be eliminated.
Of course, this requires normalization of the natural variables to be processed. The last-named operation can be carried out either within the fuzzy processor itself, or upstream of its inputs for the natural variables by means of specialized hardware.
REFERENCES:
patent: 5295226 (1994-03-01), Yamakawa
patent: 5343553 (1994-08-01), Miyazawa et al.
patent: 5710867 (1998-01-01), Giacalone et al.
patent: 0 471 843 (1992-02-01), None
patent: 0 488 883 (1992-06-01), None
patent: 0 636 967 (1995-02-01), None
European Search Report from European Patent Application 96830090.5. filed Feb. 29, 1996.
Microprocessors and Microsystems, vol. 18, No. 7, Sep. 1, 1994, pp. 409-420, “Fuzzy Logic and the Neuron Chip”.
Kaufmann, A., and Gupta, M.M.: “Introduction to Fuzzy Arithmetic,” Van Nostrand Reinhold, New York, pp. 80-81, 10-13, 68-70, Jan. 1991.
Zadeh, L.: “Fuzzy Logic,” Computer Magazine, vol. 21, pp. 83-93, Apr. 1988.
Dubois, D., Prade H.: “What are Fuzzy Rules and How to Use Them,” Fuzzy Sets and Systems, vol. 84, pp. 169-185, Feb. 1996.
Cuce' Antonino
Lo Presti Matteo
Galanthay Theordore E.
Hafiz Tariq R.
Morris James H.
SGS--Thomson Microelectronics S.r.l.
Starks, Jr. Wilbert L.
LandOfFree
Method and apparatus for storing one or more natural... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for storing one or more natural..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for storing one or more natural... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2615142