Data processing: artificial intelligence – Fuzzy logic hardware
Reexamination Certificate
1997-04-25
2002-05-07
Chaki, Kakali (Department: 2122)
Data processing: artificial intelligence
Fuzzy logic hardware
C318S102000, C318S561000
Reexamination Certificate
active
06385599
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The invention relates to the devising, installation or repairing of an electronic system and in particular to the devising, installation and repairing of a machine controlled by a fuzzy logic program.
2. Discussion of the Related Art
The control of a machine by a fuzzy logic program requires the measurement of measurement data elements by one or more sensors. These measurement data elements are then assigned to brackets of operation, of measurement, that correspond to a dynamic range of measurement of the sensor. These brackets, assessed by membership coefficients, form functions of membership of this measurement. They are also called universes of discourse of the variable constituted by the measurement data element. In general, the number of brackets is small: five brackets, or even two brackets. The measurement performed may be imprecise without leading to any erratic control of the machine. Furthermore, in the memory of the microprocessor, a set of rules of the following type is stored: “If the measurement data belongs to a given bracket, then the control of the machine must be of such a type”.
One problem arises when the machine malfunctions or when the measurement sensor malfunctions. In normal maintenance, this motor or this sensor is replaced by an identical motor or sensor. It may happen, however, that an identical replacement motor or sensor is no longer available. The faulty device is then replaced by a motors or a sensor of a different or merely similar type. The matching of these new elements to the machine is not truly possible just with a conventional software-based approach.
In the field of fuzzy logic, without changing the set of rules and the set of membership functions memorized, the fuzzy model of the system remains reliable. However, the replacement of the measurement or control device leads to a difference in response by the system with respect to the dynamic range of measurement. The modification is then carried out only if there has been a preliminary fuzzy logic modeling of the modification of the machine due to the replacement of the motor, sensor or other device. This modeling calls for the return of the machine and of the microprocessor to the laboratory, which presents a problem in itself and does not make it easy to replace the different units of the machine.
SUMMARY OF THE INVENTION
It is an object of the invention to overcome this and related problems and avert situations where installations are written off when the replacement of certain parts is no longer possible, or when their modification makes them evolve into a range of equipment not planned initially. An aspect of the invention provides the microprocessor that implements the fuzzy logic with an external command by which it will be possible to control the machine manually so as to make it explore its entire dynamic range of development when this command is made.
During a learning stage, measurement data elements pertaining to this exploration may then be acquired, in particular in a central processing unit that has a network type connection with the machine. In this microprocessor, or in this central processing unit, membership functions and/or modifying rules are then prepared. Then, the set of membership functions or the set of rules of the fuzzy logic microprocessor are then modified with these rules or these modifying functions. As the situation dictates, a new set of membership functions may be inserted. The learning procedure is then stopped. It can be seen then that the machine can subsequently work exactly or effectively the same as it worked before, with the same commends applied from the exterior, without the modification resulting in a wrong interpretation of these external commands.
An illustrative embodiment of the invention is directed to a method for the management of an electronic system comprising a central processing unit, a machine, a microprocessor and its memory, and at least one measurement sensor, the machine being controlled by the microprocessor according to a fuzzy logic program. Membership functions are prepared and recorded in the memory of the microprocessor, these membership functions setting up correspondence between measurement data elements and coefficients of membership of these measurement data elements in measurement brackets. Additionally, rules are prepared and recorded, in the memory of the microprocessor, for the control of this machine as a function of the values of these membership coefficients, and measurement data elements delivered by the sensor are read. The result of the control rules for the measurement data elements read may then be computed, and the machine controlled accordingly. As discussed above, the measurement sensor and/or the machine is replaced when it is defective. Then, to adjust this electronic system after this replacing operation, the system is made to develop between two identified points of operation, the measurement data elements corresponding to these two points are read, membership functions and/or modifying rules are prepared, and the membership functions and/or the rules recorded are modified as a function of these modifying rules.
The membership functions and the rules may be modifying rules, local rules or specific rules.
REFERENCES:
patent: 4985824 (1991-01-01), Husseiny et al.
patent: 5285545 (1994-02-01), Payne et al.
patent: 5306995 (1994-04-01), Payne et al.
patent: 5317368 (1994-05-01), Shimomura et al.
patent: 5412291 (1995-05-01), Payne et al.
patent: 5446523 (1995-08-01), Shimomura et al.
patent: 5452438 (1995-09-01), Umeda et al.
patent: 5619614 (1997-04-01), Payne et al.
patent: WO 93 09509 (1993-05-01), None
Kosko, Bart, Neural Networks and Fuzzy Systems: Dynamical Systems Approach to Machine Intelligence, Prentice Hall, Jan. 1992.*
Kosko, Bart, Neural Networks and Fuzzy Systems: A Dynamical Systems Approach to Machine Intelligence, Jan. 1992.*
Kosko, Bart, Neural Networks and Fuzzy Systems: A Dynamical Systems Approach to Machine Intelligence, 1992, pp. 326-327.*
Multi-Sensor Integration System with Fuzzy Inference and Neural Network, Fukuda et al.vol. 2, Jun. 7, 1992, pp. 757-762.
Chaki Kakali
Morris James H.
SGS-Thomson Microelectronics S.A.
Starks, Jr. Wilbert L.
Wolf Greenfield & Sacks P.C.
LandOfFree
Method and apparatus for a fuzzy self-adaptive control system 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 a fuzzy self-adaptive control system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for a fuzzy self-adaptive control system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2905309