|
Customer Profile
Winbond Electronics
Corporation America
www.winbond.com
Our Solution
The WINBOND SR-ADT development tool will allow the user
to build and define his speech recognition system in
a fully graphical environment.
Tools and Technologies
Windows NT, VC++ and some executables, dlls/libraries
etc. given by the client. Visual Source Safe for version
control. Rational Rose for Design. Rational Robot for
testing.
Methodology
We developed the Software using the VC++ and Object
Oriented Methodology.
Benefits
The tool enables the user to "Voice Simulate"
his design in the PC environment before downloading
it to the embedded system. With this the tool enables
the user to test and define the VUI even before starting
on his hardware design or in parallel to it
|
Introduction
The WINBOND SR-ADT is a sophisticated
development tool used to generate code and Data images
for voice user interfaces on embedded systems. This
tool can be used with the client’s SRXXXX series
of chips in embedded voice recognition systems.
Situation
To develop Voice User Interfaces
(VUI) requires a lot of effort for writing, debugging
and testing the software because every time we have
to burn the EPROM for testing the embedded system. Our
tool makes the life of the developer easier by providing
an environment where VUI can be designed and implemented
in a much easier fashion.
Features
WINBOND SR-ADT Software mainly contains
the following modules.
- Graphical User Interface
- Topic Compiler
- Simulator
- IVS Module
- Flashit Module
- Code generation module
Graphical User Interface module:
This allows the user
to create the state machine representation using the
building blocks provided by the tool.
Topic Compiler: This
compiles the state machine design and generates the
list of errors if any.
Simulator: The
Simulator is used to simulate and test the design before
downloading it into the actual embedded system.
IVS Module:
This module allows the user to create and edit word
tables. It compresses all the word tables, sentence
tables and the voice data images and creates a binary
file to be downloaded to the chip.
Flashit module: This
module sends all the output images generated by the tool
to the Speech recognition Chip based on a memory map
file.
Code generation module:
This module automatically generates the C Program based
on the state machine design prepared by the user that
can be compiled into 8051 code and downloaded to the
embedded system. This is an excellent feature that will
save a lot of effort for VUI system developers who currently
have to write the code manually.
Our Process
The steps involved in Generating the
Code and Data for SRXXXX series of chips are shown below
:
- Creating the project with State
machine representation
- Creating the Word tables and Sentence
tables using the IVS Module
- Compiling the project
- Voice Simulation and Testing
- Building the IVS Vocabulary
- Generating the ‘C’ code
using the Code generation module
- Sending the data images to the SRXXXX
series of chips
We played a considerable
role in identifying the requirements for the project.
We added many features creatively so that it will be
really useful to the developer.
Conclusion
Winbond SR-ADT reduces the efforts
required for developing the embedded voice recognition
systems to a great extent and makes the life of the developer
much easier.
|
