Gain control for fuzzy logic controller

• US 5,397,973 A
• Filed: 09/16/1993
• Issued: 03/14/1995
• Est. Priority Date: 09/16/1993
• Status: Expired due to Term

• Current Assignee: Allen-Bradley Company (Rockwell Automation, Inc.)
• Original Assignee: Allen-Bradley Company (Rockwell Automation, Inc.)
• Inventors: Dummermuth, Ernst
• Primary Examiner(s): Shoop, Jr., William M.
• Assistant Examiner(s): Masih, Karen
• Application Number:
  US08/122,855
• Time in Patent Office: 544 Days
• Field of Search: 318/560, 318/628, 318/568.11, 395/1
• US Class Current: 318/628
• CPC Class Codes: G05B 13/0275 using fuzzy logic only

Abstract

A method of fine tuning the "gain" of a fuzzy logic controller, the controller having input and output membership functions where the input membership functions receive one or more inputs and are grouped in sets according to each input, where the gain is tuned by uniformly scaling the domains of one or more of the sets of the input membership functions.

Claims

1. A method of tuning a feedback controller employing fuzzy logic, the controller having a set of inputs and at least one output, the output being the result of fuzzy logic applied to the inputs, the fuzzy logic having a plurality of input and output membership functions arranged in sets, each input membership function of a set being identified to a single one of the inputs and the membership functions of the set together being defined between domain limits, the method comprising the steps of:

selecting at least one set; and

tuning the gain of the fuzzy logic for the selected set by changing the domain limits to change the domain of the input membership functions associated with that set

wherein the domain is changed incrementally by a predetermined expansion factor.

2. A method of tuning a feedback controller employing fuzzy logic, the controller having a set of inputs and at least one output, the output being the result of fuzzy logic applied to the inputs, the fuzzy logic having a plurality of input and output membership functions arranged in sets, each input membership function of a set being identified to a single one of the inputs and the membership functions of the set together being defined between domain limits, the method comprising:

selecting at least one set; and

tuning the gain of the fuzzy logic for the selected set by changing the domain limits to change the domain of the input membership functions associated with that set,

wherein the membership functions are symmetrical with respect to zero and the domain limits of each membership function are scaled by a constant value.

3. A feedback controller for controlling a process having a process input and process output, the feedback controller receiving a reference input related to a desired process output and producing a process input, the feedback controller comprising:

a comparator for comparing the reference input to the process output to produce an error signal;

a differentiator for differentiating a function of the process output to create feedback difference signal;

fuzzy logic receiving the error signal and the feedback difference signal to produce at least one output, the output being the result of fuzzy logic applied to the error signal and the feedback difference signal, the fuzzy logic having a plurality of input and output membership functions arranged in sets, each input membership functions identified to one of the error signal and the feedback difference signal and being defined between domain limits;

a set selector for selecting at least one membership function set;

a gain adjustment input for receiving a gain adjustment command from an operator; and

a scaler for scaling, by the gain adjustment command, the domain of the membership function set selected by the set selector.