Glossary - Omron SYSMAC C200H-FZ001 Operation Manual

condition part
center of gravity method
conclusion part
defuzzification
fuzzy output
fuzzy variable
grade
label
knowledge base
maximum value method
membership function
rule
rule grade

Glossary

The "if" clauses of a rule. A single rule can have up to 8 condition parts in the
form "IF X AND X ... AND X
1 2
A method of defuzzification. The center of gravity of the fuzzy outputs (along
the x-axis) is calculated as the final result of fuzzy logic processing.
The "then" clauses of a rule. A single rule can have up to 2 conclusion parts
in the form "THEN Y AND Y
1
Defuzzification is the process that combines the fuzzy outputs to a single re-
sult that is output from the Fuzzy Logic Unit. There are two methods of defuz-
zification available in the C200H-FZ001: the center of gravity method and
maximum value method.
After all of the rule grades are calculated, the grade for each label is calcu-
lated. The grade for a label is called its fuzzy output. The fuzzy output is the
maximum rule grade for that label. The maximum value is taken because the
rules are linked by logical ORs.
A variable such as distance, temperature, or pressure, that is measured by
the PC's sensors. The measured values of the fuzzy variables are input to
the Unit for fuzzy logic processing.
A number between 0 and 1 that indicates how well the given value of the
fuzzy variable satisfies a label. A grade of 0 means the label is not satisfied
at all, and a grade of 1 means the label is perfectly satisfied at the given val-
ue of the fuzzy variable.
Labels are used to describe the state of a fuzzy variable. Labels for distance
might be "close," "OK," and "far." Labels for temperature might be "cold,"
"cool," "warm," and "hot." Labels for pressure might be "light," "moderate,"
and "hard." The C200H-FZ001 can have up to 7 labels for each fuzzy vari-
able.
The rules and membership functions.
A method of defuzzification. The position (along the x-axis) of the maximum
fuzzy output is used as the final result of fuzzy logic processing. Either the
leftmost or rightmost position will be used if there are two or more fuzzy out-
puts with the same maximum value.
Defines the relationship between the grade and the fuzzy variable for each
label. The most accurate shape for a membership function is usually a bell
curve, but triangles and trapezoids are used most often to simplify and speed
up processing.
An if/then statement that expresses a relationship between inputs and out-
puts. The C200H-FZ001 can have up to 128 rules.
Membership functions assign grades to the labels of the condition parts of a
rule according to the present conditions. The rule grade is the minimum of
these grades.
."
8
."
2
3