Products/fuzzyTECH Editions/Editions Overview
We have
created different Editions of fuzzyTECH to
complement the specific needs of your target hardware. This page
provides you with an
overview on the three groups of fuzzyTECH Editions:
Further down on this page you find detailed Technical Specifications for all Editions.
General Purpose fuzzyTECH Edition
The Edition
also support the special
"plugandplay" type runtime modules provided for specific process
control
software such as InTouch™, FactoryLink™,
TheFIX™, Genesis™, and WinCC™.
In addition to the M source code generation provided by every fuzzyTECH
Edition,
the fuzzyTECH Professional Edition also
contain a MEX runtime module
that provides highperformance computation of fuzzy logic systems in
the Matlab/Simulink™
environment. N.B.: Starting
with release 5.0, the fuzzyTECH Professional
Edition replaces the fuzzyTECH Precompiler Edition.
The following tables provide a summary of the various fuzzyTECH Editions specifications. For a detailed explanation of the terminology used here, download and install the fuzzyTECH Demo from the Download section and use the index function of its Online Help System.
Overview
fuzzyTECH is available in different editions to
provide
the most comprehensive support for your target platform and application
area. Due to
differences in the capabilities of the supported hardware platforms,
technical
restrictions apply to the size of the fuzzy systems. The following
table shows an
overview of the
maximum number of interfaces, variables, terms, rule blocks and rules
for the different fuzzyTECH
Editions. A "" sign indicates that no
practical limit exists. The total number of Variables
(Total) represent the
number of input, output and intermediate
variables of the entire fuzzy logic system. The columns Input
and Output show the
maximum amount of input and output
variables. Terms per Variable
relates to the total number of terms
for each variable. The column Total
Terms shows
the maximum number of terms that may occur in one fuzzy logic project.
The Rules
sections shows the maximum total number of Rule
Blocks and Rules
that a fuzzy logic project may contain, as well as the
maximum total number of input variables (Inputs
per RB) and
output variables (Outputs per RB)
that can be assigned to a
rule block.
Overview  Variables  Rules  
Feature
Edition

Total  Input  Output  Terms per Variable Linguistic/Categorical 
Total Terms  Rule Blocks (RB) 
Inputs per RB  Outputs per RB  Total Rules 
Professional  255  255  32  32/255  65535  50  250  11   
Membership Functions
fuzzyTECH supports various fuzzy logic inference
methods and algorithms. The following table lists methods of
fuzzification and Types of membership functions
(MBF) supported by each
fuzzyTECH Edition. Standard
MBFs
are sometimes called "4point definitions". Arbitrary
MBFs can be defined with up to 16 points of definition. Inverse MBFs (inverse terms)
are useful for filling a rule part
with the negated form of an already existing term. The column MBF
Shapes
shows the available approximation functions for membership functions.
The column Fuzzification
Method lists the supported algorithms for the fuzzification
step of the fuzzy
logic inference. Fuzzy Input
indicates that variables
are inputted as fuzzy values (i.e. as vectors of membership degrees)
instead of crisp values. The standard fuzzification
method is computation at run time (MBF
Computation). For most target hardware implementations,
this is the most efficient approach.
Membership Functions (MBF)  Type  Shape  Fuzzification Methods  
Feature Edition 
Standard MBF  Arbitrary MBF  Inverse MBF  Linear  Spline  Fuzzy Input  MBF Computation  Categorical 
Professional  x  x  x  x  x  x  x  x 
Inference
and Defuzzification
The table below summarizes supported methods of fuzzy logic
inference and defuzzification. The fuzzy
inference consists of three computational steps: Aggregation,
Composition,
and Result Aggregation. Different operators can be
chosen for aggregation
(Input aggregation) and result aggregation. Fuzzy operators used for
aggregation (Minimum or
Maximum) combine the
preconditions of each fuzzy rule. Beside this standard operators, some fuzzyTECH
Editions support compensatory operators (Gamma,
MinAvg, MinMax),
that help to compute relations between rules formulated with the logic
standard operators
AND (Minimum) and NOR (Maximum). The second
step of the fuzzy rule inference, the composition, works generally with
the PRODOperator
as fixed operator. Standard Rules
are rules with a fixed rule weight (Degree of Support = 1.0) that
cannot be changed. FAM
Rules
stands for "Fuzzy Associative Maps" and refers to individually weighted
rules
(Degree of Support = DoS). The last step of fuzzy inference is the
socalled result
aggregation. Its MAX
operator selects the maximum firing degree of all rules matching to the
term. The BSUM
operator uses a bounded sum. Thus, all firing degrees are summed by
using a bound of one.
Note that BSUM result
aggregation is different from BSUM MoM
and BSUM CoA. The bounds are zero and one.
Inference & Defuzzification  Aggregation  Composition  Result Aggregation  Defuzzification  
Feature Edition 
Min.  Max.  MinMax  MinAvg  Gamma  Product  Standard Rules  FAM Rules (DoS)  Max  BSUM  CoM  CoA  MoM  Categorical MoM  Fuzzy Output  Hyper CoM^{ 1)} 
Professional  x  x  x  x  x  x  x  x  x  x  x  x  x  x  x  x 
1) Only available as addon module
The result of the fuzzy inference is a fuzzy value that has to be retransformed into a crisp value. This transformation is called Defuzzification. Different computation methods can be applied for defuzzification. The standard defuzzification is CoM (CenterofMaximum), delivering the "best compromise" for the inference result. It is equivalent to most implementations of CenterofArea (CoA) and CenterofGravity (CoG) methods. The MoM (MeanofMaximum) method delivers the "most plausible" result. Hence, it is mostly used in applications such as pattern recognition, data analysis, and decision support. The defuzzification method HyperCoM (see last column) is only available as addon module for a few editions. HyperCoM is a defuzzification method that takes both positive and negative experience into consideration (e.g. in the form of recommendations and warnings). A hyperdefuzzification strategy weights these recommendations and warnings against each other and computes a membership function, from which HyperCoM then computes the optimum based output value.
System Optimization and Analysis, Addon
Modules
All fuzzyTECH Editions come with a complete set of
debugging features (Debug Modes) and
analysis tools Analyzers (see
Table below). Remote debugging, where fuzzyTECH
running on the PC debugs
a fuzzy logic system running on a different computer/microcontroller,
is
facilitated by Online
and RTRCD
debug modes. The RTRCD debug mode (Real Time Remote Cross Debugging)
lets you analyze the running system and modify rules and membership
functions.
The Online debug mode
in addition allows for
all types of modifications onthefly. To expand the capabilities of fuzzyTECH,
several Addon Modules are available. The
HyperInference Module
expands traditional fuzzy logic rule inference with the ability of
"prohibitive" rules. Click on the [AddOn Modules] item in the
treeview to your left for more information.
Tools 
Debug Modes (internal) 
Communication Channels  Analyzer 
Addon Modules 

Feature Edition 
Interactive  File / Batch  DDE  RTRCD  Online  TCP/IP  Serial Interface (RS232)  User defined (FTOCC) 
TransferPlot, 3D Plot, Time Plot, Trace 
Hyper Inference  
Professional  x  x  x    x  x  x  x  x  x 
Code Generation
The table below lists supported code generator options. The File Code
option generates a complete C source code for an
executable program which accepts file data as input and writes outputs
to a file as well.
Since fuzzyTECH generates the complete fuzzy logic
system as a single function,
input and output values can be transferred as function
parameters
(Function Call) over
the system stack.
The C Code
column lists which C compiler standard is supported. Besides, Code
Interface Resolution
indicates the data types of the interfaces, too.
Code Generation  CodeOptions  CCode  JavaCode  ActiveX Java fuzzyTECH Runtime DLL Delphi ^{1)} Visual Basic ^{1)} C# ^{1)} 
Structured Text by IEC 611313 standard for CoDeSys 

Feature Edition 
I/O passing as function parameter 
ANSI  Java  FTR  ST  
Code Interface Resolution (8 Bit, 16 Bit, double)  
Professional  x  8/16/d  16/d  16/d  16/d 
^{1)}Using fuzzyTECH Runtime DLL.