addMF
Add membership function to fuzzy variable
Syntax
Description
fis = addMF(fis,varName,type,parameters)varName in the fuzzy inference system
fis. No other properties of fis are
changed.
fis = addMF(___,Name=Value)
var = addMF(var)var and returns the
resulting fuzzy variable in var. No other properties of
var are changed.
If var does not contain any membership functions, this syntax
adds a default type-1 membership function. Otherwise, the type of the added membership
function matches the type of the existing membership functions in
var.
var = addMF(var,type,parameters)
var = addMF(___,Name,Value)
To add a type-2 membership function to a fuzzy variable with no existing membership
functions, you must specify either the LowerLag or
LowerScale name-value argument.
Examples
Create a Mamdani fuzzy system, and add three input variables and one output variable. For this example, give the second input variable and the output variable the same name.
fis = mamfis; fis = addInput(fis,[0 80],Name="speed"); fis = addInput(fis,[0 100],Name="throttle"); fis = addInput(fis,[0 10],Name="distance"); fis = addOutput(fis,[0 100],Name="throttle");
Add a membership function to the first input variable, specifying a trapezoidal membership function, and set the membership function parameters.
fis = addMF(fis,"speed","trapmf",[-5 0 10 30]);
You can also specify the name of your membership when you add it to a fuzzy system. Add a membership function called "high" to the first input variable.
fis = addMF(fis,"speed","trapmf",[50 70 80 85],Name="high");
View the membership functions for the first input variable.
plotmf(fis,"input",1)
If your system has an input variable with the same name as an output variable, you must specify the variable type when adding a membership function. For example, add a membership function to the output variable.
fis = addMF(fis,"throttle","trimf",[0 20 40],VariableType="output"); plotmf(fis,"output",1)
Alternatively, you can add a default membership function to a fuzzy system and set its parameters using dot notation. For example, add and configure a membership function for the third input variable.
fis = addMF(fis,"distance"); fis.Inputs(3).MembershipFunctions(1).Type = "trapmf"; fis.Inputs(3).MembershipFunctions(1).Parameters = [-1 0 2 4]; plotmf(fis,"input",3)
Create a type-2 Sugeno fuzzy system, and add two input variables and one output variable.
fis = sugfistype2; fis = addInput(fis,[0 80],Name="speed"); fis = addInput(fis,[0 10],Name="distance"); fis = addOutput(fis,[0 100],Name="braking");
Add a membership function to the first input variable, specifying a trapezoidal membership function, and set the membership function parameters. This type-2 membership function uses default lower membership function lag and scale parameters.
fis = addMF(fis,"speed","trapmf",[-5 0 10 30]);
You can also specify the configuration of the lower MF when adding a type-2 membership function.
fis = addMF(fis,"speed","trapmf",[10 30 50 70],LowerScale=0.8,LowerLag=0.1);
You can also specify the name of your membership function when you add it to a fuzzy system. Add a membership function called "high" to the first input variable.
fis = addMF(fis,"speed","trapmf",[50 70 80 85],Name="high");
View the membership functions for the first input variable.
plotmf(fis,"input",1)
Create a fuzzy variable with a specified range.
var = fisvar([0 1]);
Add a membership function to the variable, specifying a trapezoidal membership function, and set the membership function parameters.
var = addMF(var,"trapmf",[-0.5 0 0.2 0.4]);You can also specify the name of your membership when you add it to a fuzzy variable. For example, add a membership function called "large".
var = addMF(var,"trapmf",[0.6 0.8 1 1.5],Name="large");
View the membership functions.
var.MembershipFunctions
ans =
1×2 fismf array with properties:
Type
Parameters
Name
Details:
Name Type Parameters
_______ ________ ____________________________
1 "mf1" "trapmf" -0.5 0 0.2 0.4
2 "large" "trapmf" 0.6 0.8 1 1.5
Alternatively, you can add a default membership function to a fuzzy variable and set its parameters using dot notation.
var = fisvar([0 1]);
var = addMF(var);
var.MembershipFunctions(1).Type = "trapmf";
var.MembershipFunctions(1).Parameters = [-0.5 0 0.2 0.4];Create a fuzzy variable with a specified range. By default, this variable has no membership functions.
var = fisvar([0 9]);
To add a type-2 membership function to a variable with no existing membership functions, specify either a LowerLag or LowerScale value for the membership function. For example specify a lower scale value.
var = addMF(var,"trimf",[0 3 6],LowerScale=1);Once a variable contains a type-2 membership function, you can add additional type-2 membership functions without specifying one of these parameters.
var = addMF(var,"trimf",[3 6 9]);View the membership functions.
var.MembershipFunctions
ans =
1×2 fismftype2 array with properties:
Type
UpperParameters
LowerScale
LowerLag
Name
Details:
Name Type Upper Parameters Lower Scale Lower Lag
_____ _______ ________________ ___________ __________
1 "mf1" "trimf" 0 3 6 1 0.2 0.2
2 "mf2" "trimf" 3 6 9 1 0.2 0.2
To use a custom membership function when designing a FIS at the MATLAB® command line, specify the name and parameters when adding the membership function using addMF. For example, the following command adds custom membership function custmf1 to the first input variable, input1, of FIS myFIS and names it customMF1.
Create a FIS and add an input variable.
myFIS = mamfis; myFIS = addInput(myFIS);
Add custom membership function custmf1 to the first input variable, input1, of this FIS and name it customMF1.
myFIS = addMF(myFIS,"input1","custmf1",[0 1 2 4 6 8 9 10],... Name="customMF1");
Ensure that the range of the input variable matches the expected range of your membership function. For this example, set the range of the first input to [0 10].
myFIS.Inputs(1).Range = [0 10];
To verify the appearance of your membership function, you can plot it using plotMF.
plotmf(myFIS,"input",1) 
