Data processing: artificial intelligence – Fuzzy logic hardware
Reexamination Certificate
1998-08-11
2002-08-06
Davis, George B. (Department: 2122)
Data processing: artificial intelligence
Fuzzy logic hardware
C706S004000, C706S008000, C706S900000, C706S903000
Reexamination Certificate
active
06430544
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates generally to logic control systems, and more particularly, to a priority based constraint controller system comprising a plurality of cascaded single variable priority constraint controllers for controlling a system having multiple variables and constraints.
BACKGROUND OF THE INVENTION
Control systems and computer-controlled electronic devices have historically been controlled by digital control systems. Such control systems use bi-state digital logic which requires a value of either “TRUE” or “FALSE, so that approximations are often required of real-world control problems. For example, an input/output relationship y=f(x) would be specified either as a mathematical function or as a series of points using, for example, a look-up table. The former use of a mathematical function may require complex mathematics to accurately represent real-world control problems. Further, the latter use of a look-up table, such as a ROM, introduces problems such as large memory requirements for adequate approximation, the concomitant addressing function associated with large memories, as well as interpolation problems.
For example,
FIG. 1A
shows an exemplary nonlinear sigmoidal function y=f(x). If digital logic was used to characterize the function y=f(x), it would be necessary to approximate the function shown in
FIG. 1A
by using discrete values, y
i
=a
i
·x
i
(I=1,2, . . . , n), as shown in FIG.
1
B. Since the number n of “crisp” values is limited, there inherently exists an interpolation error for values of x between x
i
and x
i+1
. The term “crisp” refers to an input having a single discrete value. In addition, it becomes impractical to write a rule for every input combination where there exists a large number of inputs whose values can cover a wide input range.
An alternative approach to control theory, known as “fuzzy logic”, was developed by L. Zadeh in 1963. Rather than evaluating the two values “TRUE” and “FALSE” as in digital logic, fuzzy terms admit to degrees of membership in multiple sets so that fuzzy rules may have a continuous, rather than stepwise, range of truth of possibility. For example, in applying fuzzy logic, a person need not strictly be included or excluded as a member from a set of “tall persons”; rather, to the extent a person may be “tall” to a greater or lesser degree, the member is assigned to the set with a degree of membership between the values of “1” and “0”.
FIG. 1C
illustrates the principle of fuzzy logic in evaluating the function illustrated in FIG.
1
A. The function f(x) is approximated by a plurality of fuzzy sets
10
which overlap. Rather than approximating a continuous value x by a discrete value x
i
, logic determines for a given value x whether the value x is nearest to the center of a fuzzy set
10
. If an x value is equidistant from two or more fuzzy sets, the resultant y value can be made proportional to the output values suggested by all the fuzzy sets of which the value x is a member. Thus, a fuzzy number may be two dimensional, having assigned fuzzy sets and corresponding membership values.
Since fuzzy logic can operate within the relative imprecision of the real-world environment, the advantages of fuzzy logic and fuzzy set theory have become apparent in numerous areas, such as robotics, natural language recognition, the automobile and aircraft industry, artificial intelligence, etc.
The implementation of fuzzy logic for a controller has been shown in U.S. Pat. No. 5,655,056 entitled “FUZZY CONTROLLER GROUP CONTROL SYSTEM”. This patent discloses a system in which a plurality of fuzzy controllers are interconnected by a communication line and the parameters of rules and membership functions of each fuzzy controller are capable of being adaptively changed, depending upon the state of the controlled system so as to improve the control performance of the overall system. In this patent, the CPU unit of each fuzzy controller acts to adaptively alter the presently prevailing rule and membership function based on the information from other fuzzy controllers in order to improve performance.
Another prior art controller is described in a publication of the IEEE Journal of Solid State Circuits, vol. 25, No. 2, April 1990 entitled “A VLSI FUZZY LOGIC CONTROLLER WITH RECONFIGURABLE, CASCADABLE ARCHITECTURE”, by Hiroyuki Watanabe et al. This article discloses a general background on fuzzy logic devices along with an illustration of a cascadable architecture. Still another implementation is shown in U.S. Pat. No. 5,131,071 entitled “FUZZY INFERENCE APPARATUS” which discloses multiple inference sections (
10
) whose output is integrally processed by a concluder section (
110
) and having a membership function generating circuit array which allows switching of the inference system in a time series fashion to enable a hierarchically oriented system to be achieved. However, each of these prior art control systems comprise complex multi-variable inputs and outputs for each fuzzy controller from/to sensors in order to operate and control a complex multi-variabled system. Furthermore, such prior art control systems do not provide a priority-based method of using an upstream fulfillment value to adaptively scale the incremental output of each subsequent fuzzy controller, thereby permitting the operation of certain fuzzy controllers to either dominate or be subordinate to other controllers within the system. Accordingly, it is highly desirable to obtain a priority constraint control system using single variable priority constraint controllers cascadable within an architecture to prioritize constraints for permitting the system to take into account multiple variables and constraints.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide a fuzzy logic controller system comprising a plurality of fuzzy logic controller units cascadably coupled to one another, each controller having membership processing means for receiving a single input signal corresponding to a sensed parameter from a controlled system and performing fuzzy inference operations based on the input signal for outputting an incremental output signal indicative of a requested change in a controlled output from a current level, permission means for scaling the incremental output signal in response to a fulfillment signal from another controller, fulfillment means responsive to the membership processing means and to the permission means for providing a fulfillment signal to another controller indicative of a permissible range of freedom for scaling the incremental output signal of another controller in the cascading sequence, and summing means for summing each of the incremental output signals from the plurality of controllers and applying the sum to the controlled output for providing a correction to the current level of the controlled output, wherein the fulfillment signal output from the controller operates to constrain the incremental output signal and influence of subsequent downstream controllers.
It is a further object of the present invention to provide a method for operating a fuzzy controller group control system, including a plurality of fuzzy controllers, for controlling an output parameter, comprising the steps of for each said fuzzy controller in the group, receiving a single input variable from a controlled system and performing fuzzy inference operations for outputting an incremental output indicative of a requested change in the output parameter from a current level, adaptively scaling the incremental output of each controller by an at least one signal output from a previous controller indicative of a permissible range of freedom for varying the incremental output, and summing the scaled incremental outputs from each of the plurality of fuzzy controllers and applying the sum to the current output parameter to obtain an updated output parameter.
REFERENCES:
patent: 4875184 (1989-10-01), Yamakawa
patent: 5131071 (1992-07-01), Tsutsumi
patent: 5159547 (1992-10-01), Chand
Davis George B.
Morris Duane
LandOfFree
Single variable priority constraint fuzzy 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 Single variable priority constraint fuzzy control system, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Single variable priority constraint fuzzy control system will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2912490