What FuzzyDesigner is System Requirements

FuzzyDesigner is a software package for designing smart components • Intel Pentium (or compatible) 90 MHz or higher
implemented as Fuzzy Systems with optional hierarchical structure. • 64 MB of RAM
FuzzyDesigner generates code of the fuzzy system, which runs on PCs • Windows 98, Windows NT 4.0 SP6a, Windows Millennium
and various industrial hardware platforms, primarily on Rockwell Edition, Windows 2000, Windows XP Home Edition, Windows
Automation Logix5000 family of programmable logic controllers. XP Professional, or Windows .NET Server family platforms
FuzzyDesigner also generates platform independent C, C#, Java and
Matlab code of the fuzzy system and provides means for easy
communication with other applications using industrial standard OPC
protocol and optionally DDE protocol.
Supported Membership Functions
Each linguistic term defined for Input or Output Linguistic Variables is
defined by a membership function. FuzzyDesigner supports the following
membership functions:
Trapezoidal (Input and Output Linguistic Variable)
Inverse Trapezoidal (Input Linguistic Variable)
S-shape (Input Linguistic Variable)
Inverse S-shape (Input Linguistic Variable)
Singleton (Output Linguistic Variable)
Piece-wise linear (input fuzzy set)

A fuzzy system designed and generated by FuzzyDesigner might be

used for example as a direct nonlinear fuzzy-rule based controller, a
nonlinear (fuzzy rule based) PID controller supervisor, a smart switch
between conventional controllers (Takagi-Sugeno Controller), a
nonlinear process model for diagnostic or forecasting purposes, a large
hierarchical decision-making system etc.

Hierarchical Fuzzy Logic System

FuzzyDesigner includes library of components, which allows a user to
design a complex hierarchical fuzzy systems. The following
components are included:
Input Port - stores fuzzy or crisp input values of the fuzzy Fuzzy Rule Base
system, crisp value can be preprocesses by the user defined FuzzyDesigner supports:
linear digital filter
• two inference methods:
Input Linguistic Variable - stores linguistic terms and is Mamdani inference
used for classification of the actual component input (crisp or
Fuzzy Arithmetic Inference
fuzzy value) into reference fuzzy sets
• t-norms:
Rule Block - stores rules, performs fuzzy logic inference
based on fuzzy rules and computes degrees of fulfillment of minimum t-norm
consequent variables from degrees of fulfillment of premise product t-norm
variables • s-norms:
Intermediate Linguistic Variable - is a bridge for logical maximum s-norm
chaining of rule blocks Fuzzy rules can be designed manually by a user, or automatically, it means
Output Linguistic Variables - stores parameters of that FuzzyDesigner generates rules for all combinations of selected premise
linguistic terms and computes an output value from degrees variable and a user fills consequent fuzzy terms. Every rule can be
of fulfillment of all terms deactivated and independent rule weight can be defined for each consequent
variable.
Output Takagi-Sugeno Variable - stores parameters of
functional (linear or constant) terms and computes an output
value from degrees of fulfillment of all terms
PID controller - allows intelligent supervision of the
conventional PID controller
Output Port - stores fuzzy or crisp output values of the
fuzzy system

FuzzyDesigner – Brief Product Info August 2003

http://www.protys.cz
FuzzyDesigner@protys.cz
Defuzzification Methods Code Generators
FuzzyDesigner supports 5 defuzzification methods: FuzzyDesigner optionally generates designed fuzzy project as a:
• Centroid Average (CA) • C code
• Maximum Center Average (MCA) • Java code
• Mean of Maximum (MOM) • C# code
• Smallest of Maximum (SOM) • Matlab code
• Largest of Maximum (LOM) • Matlab C-mex code
Centroid Average and Maximum Center Average methods belong to • Matlab DDE s-function
continuous ones and are frequently used in control engineering and • Matlab OPC s-function
process modeling. The rest represent discontinuous methods, which are
mainly used in decision making and pattern recognition applications for Communication Protocols &
selecting the alternative.
External Interfaces
DOF(negative) DOF(negative) is maximal
FuzzyDesigner communication server supports following protocols:
DOF(zero) • DDE – classical Microsoft Dynamic Data Exchange protocol
• OPC– industrial standard based on DCOM protocol
y* y*
To quickly incorporate your designs into communication with real time
y* y* y*
(MCA) (CA) (SOM) (MOM) (LOM)

industrial information systems, FuzzyDesigner provides OPC client

component, which allows you to easily connect ports of your project to
PID controller Supervision data source and sinks of available OPC servers and periodically
evaluate your fuzzy system.
FuzzyDesigner allows an intelligent supervision of the conventional
FuzzyDesigner provides plug-in COM(ActiveX) component, DLL
PID controller. The PID component has following facultative features:
library and Microsoft .NET assembly that you can directly use in your
• equation format with applications. These components include OPC server which allows on-
independent or dependent gains line monitoring and tuning of fuzzy system parameters.
• dumping of the influence of the
setpoint
• two formats of derivative term
(the process variable or the
error)
• manual or automatic mode
• output limiting
• dead band control - zero or no
zero crossing
• gain forgetting factor for the
parameter tracking

Fuzzy System Analysis

FuzzyDesigner
FuzzyDesigner offers several methods for the fuzzy system analysis: FuzzyDesigner OPC reading from OPC items
OPC Client OPC
Client

OPC Server Server to input ports

• 2D and 3D graphs with many options concerning zoom, line, collect data
evaluation
- compute a control action

grid density , range, axis, texture, etc. write inputs

system DCOM
from factory - make a decision
- make a prediction ethernet
evaluation

• Simulation mode which allows to watch influence of individual DCOM ethernet OPC
Server
apply results

read outputs
component values. Value changes appear also in graphs as “path”. OPC OPC
writing to OPC items from
output ports
Client industrial network Server
industrial network

Data Base Data Base PLC I/O Modul

PLC

Download to Industrial Controllers

FuzzyDesigner allows to download fuzzy systems into Rockwell Automation industrial controllers
(Logix 5000 family, SLC500 and PLC5). The size of the fuzzy system (number of inputs, outputs,
membership functions, rules, etc.) is limited only by the memory of the controller.
FuzzyDesigner allows to:
• install/uninstall the fuzzy kernel into any RSLogix5000, RSLogix500, RSLogix5 project
• add/remove fuzzy systems
• change parameters of the fuzzy system on-line
• monitor the fuzzy system running on the controller in real-time

FuzzyDesigner – Brief Product Info August 2003

http://www.protys.cz
FuzzyDesigner@protys.cz