Sei sulla pagina 1di 2

Aide-d-voix

HOME AUTOMATION THROUGH SPEECH RECOGNITION


(Vision through speech)
Abstract - A speech recognition system for Home
Automation is developed in this paper. Home
automation is the technology that enhances the A grammar, the list of words known to the program,
interactivity and autonomy of a home. It is a field lets the engine to associate the phoneme with
with potential explosive growth due to the recent particular word.
rapid improvements in computing power. Speech
recognition is the ability of a computer system to Recognition procedure is divided into two
respond accurately to verbal commands. Speech consecutive stages depending on data set and test set.
recognition makes use of specific AI (artificial It consists of two stages:
intelligence) rules to determine what words the (1) Training (for data set)
speaker is speaking. Speech recognition programs, (2) Comparison and Classification (for test set)
allows people to give command and enter data using
their voices rather than a mouse or keyboard. (1) Training: The words, which have to be
recognized, need to be added in the database
Objective: The main aim of designing this software provided in the software. The words can be
is to provide a tool of accessibility to individuals dynamically added to the database. Some level of
who have physical or cognitive difficulties, training is required to be done for accurate
impairments, and disabilities. recognition.
A software program is developed for recognizing the
commands. It derives the input from the user in form (2) Comparison and Classification: At this stage,
of speech then recognizes it and performs according comparison is done with the help of the generated
to the conditions specified in the code and word and the words on the program. Based on the
corresponding appliance is activated. This method of result the appropriate function is performed.
atomizing the home appliances serves as a
communication aid to the visually impaired. System Architecture
We propose architecture for building speech
Process: A typical complete speech recognition
recognition system, which uses speech command
process consists of the following parts: (1) sound
given by the user and processes it to perform the
converter (2) Fragmentation, (3) Recognition.
activation of the desired appliance.
Sound Acquisition: The user voice is captured with
the help of mic in a handset.
Overview
Sound Conversion: The digital sound captured by
the sound card through a mic is converted in to a
It is architecture for controlling the home appliance
more manageable format. The converter translates
through the speech command. The user will give the
the stream of amplitudes that form the digital sound
command to a handset that contains the speech
wave in to its frequency components. It is still a
recognising software system. On receiving the sound
digital representation, but more akin to what a human
signal the software does the recognising process
ear really perceives.
described above.
Fragmentation: The next stage is the identification
The architecture of the system is described as in
of phonemes –the elementary sound that are building
figure 1. The input is derived from the user and when
blocks of words. Each frequency component of
the speech recognition component present in the user
sound is mapped to a specific phoneme. This process
program senses the sound it activates the speech
actually finishes the conversion from sounds to
recognition engine to process the input word and to
words.
do the comparison operation. After the comparison
Recognition: The final step is to analyze the string.
task is completed the matched word is transferred to
user program. On receiving the matched word the
program executes signal transmission function.
can be captured by the sensor we place for routing
c SENSOR the signal to the appliance.
INTERRUPT
Sensors: The sensors are used for forwarding the
SPEECH signal transmitted by handset to the desired
RECOGNITION appliance, which need to be controlled. The signal is
ENGINE sent to the relay, which supply the required voltage
RELAY to the appliance.
DATABASE
Relay: A relay is an electrically operated switch.
Current flowing through the coil of the relay creates
a magnetic field, which attracts a lever and changes
APPLIANCE the switch contacts. The coil current can be on or off
USER PROGRAM
so relays have two switch positions and they are
MIC
Audio double throw (changeover) switches. Whenever the
System relay gets the signal from the sensor it gets activated
AC and the voltage is passed on to the appliance.
The relay's switch connections are usually labeled
COM, NC and NO:
USER COM = Common, always connect to this; it is the
moving part of the switch.
Figure 1. Architecture of Speech Recognition System NC = Normally Closed, COM is connected to this
when the relay coil is off.
NO = Normally Open, COM is connected to this
Components of the architecture when the relay coil is on.

The components in the architecture are 1.Handset Appliance: The appliance that can be activated in
2.Sensors 3.Appliance. There are three essential normal 230V supply can be automised in this system.
components in the Handset:
1.User program 2.Speech recognition engine Future work: The following is the area, which we
3.Interrupt. have identified as relevant for future work:
• Multi lingual speech recognition
Speech recognition engine: The recogniser used
here is speaker-independent software. It has an
Conclusion: We propose, the architecture for
inbuilt database containing very large number of
automising the home appliances with the help of
words in it. The user program activates the
speech recognition. This way of automising the
recognition engine when it recognizes a sound. It
home appliance will be helpful for the visually
will process the sound into a word that may closely
impaired and for physically challenged people.
match the one you provided. Initially the desired
words for the control of each appliance provided in
References:
the user program are loaded in to the engine
database. As soon as the recogniser gets any sound
input, the software does the recognising process and • “Readings in Speech Recognition”
checks whether the recognised word matches with edited by Alexander Waibel, Kai-Fu Lee.
the any of the words that are loaded by the user
program. If it matches the matched word is returned
to the program by the recogniser. • “Computer Telephony Encyclopedia”
By Richard Grigonis.
User program: Now after the recognised word is
obtained from the recognition engine the instruction
for the appropriate word given in the program are
executed. For controlling the home appliances the
signal has to be transferred to the sensors, which then
forward the activation or deactivation signal to the
appliance. For transferring the wireless signal we use
an interrupt in the handset, it sends the signal, which

Potrebbero piacerti anche