Automatic Speaker Recognition System

based on
Machine Learning Algorithms
 ABSTRACT

Speaker recognition is a technique used to automatically recognize a speaker from a recording

of their voice or speech utterance. Speaker recognition technology has improved over recent
years and has become inexpensive and and reliable method for person identification and
verification. Research in the field of speaker recognition has now spanned over five decades
and has shown fruitful results, however there is not much work done with regards to South
African indigenous languages. This paper presents the development of an automatic speaker
recognition system that incorporates classification and recognition of Speedy home language
speakers. Four classifier models, namely, Support Vector Machines, K-Nearest Neighbors,
Multilayer Perceptions (MLP) and Random Forest (RF), are trained using WEKA data mining
tool. Auto-WEKA is applied to determine the best classifier model together with its best hyper-
parameters. The performance of each model is evaluated in WEKA using 10-fold cross
validation. MLP and RF yielded good accuracy surpassing the state-of-the-art with an accuracy
of 97% and 99.9% respectively; the RF model is then implemented on a graphical user
interface for development testing.
 INTRODUCTION OF THE PROJECT

1.1 INTRODUCTION:

DEFINITION: An Embedded system is a combination of computer firmware and hardware,

fixed in capability or programmable, designed for a specific function or functions within a
larger system.

An embedded system is a controller with a dedicated function within a larger

mechanical or electrical system, often with real-time computing constrains. It is embedded as
part of a complete device often including hardware and mechanical parts. Embedded systems
control many devices in common use today. Ninety -eighty percent of all microprocessors
manufactured are used in embedded systems.

Modern embedded systems are often based on microcontrollers

(i.e. microprocessors with integrated memory and peripheral interfaces), but ordinary
microprocessors (using external chips for memory and peripheral interface circuits) are also
common, especially in more complex systems.

Since the embedded system is dedicated to specific tasks, design engineers can
optimize it to reduce the size and cost of the product and increase the reliability and
performance. Some embedded systems are mass-produced, benefiting from economies of scale.

History: one of the very first recognizably modern embedded systems was the Apollo
guidance computer, developed ca.1965 by Charles stark draper at MIT instrumentation
laboratory. At the project’s inception, the Apollo guidance computer was considered the
riskiest item in the Apollo project as it employed the then newly developed monolithic
integrated circuits to reduce the size and weight. An early mass-produced embedded systems
was the automatics D-17 guidance computer for the minuteman missile, released in 1961.
When the minuteman 2went into production in 1966, the D-17 was replaced with a new
computer that was the first high-volume use of integrated circuits.
 Since these early applications in the 1960s, embedded systems have come down in
price and there has been a dramatic rise in processing power and functionality. An early
microprocessor for example, the INTEL 4004 was designed for calculators and other small
systems but still required external memory and support chips. In 1978 national Engineering
manufacturers association released a “standard” for programmable microcontrollers, including
almost any computer-based controllers, such as single board computers, numerical, and event-
based controllers.

Purpose of embedded systems: embedded systems are used in various domains like
consumer electronics, home automation, telecommunications, automotive industry, health care,
control and instrumentation, retail and banking applications etc.Within the domain itself
according to the application usage context, they may have different functionalities. Each
embedded systems are designed to serve the purpose of any one or combination of following
tasks:

 Data collection/storage/representation
 Data communication
 Data processing
 Monitoring
 Control
 Application specific user interface (ASIC).

Data collection/storage/representation: Embedded systems are designed for the

purpose of data collection performs acquisition of data from the extended world. Data
collection is usually done for storage analysis, manipulation and transmission. The term data
refers to all kinds of information text, voice, image, video and any other measurable quantities
data can be either analog or digital. Embedded with analog data capturing techniques collect
data directly in the form of analog signals. whereas embedded systems with digital data
collection mechanism converts the analog signal to digital converter and then collects the
binary equivalent of the analog data. If the data is digital, it can be directly captured without
any additional interface by embedded systems.
 The collected data may be stored in the system (or)may be transmitted to some other
system (or)it may be processed by the system (or)it may be deleted instantly after giving a
meaningful representation. These actions are purely dependent on the purpose for which the
embedded systems are designed. Embedded systems designed for pure measurement
applications without storage used in control and instrumentation domain. Collects data and
gives a meaningful representation of the collected data by means of graphical representation
(or) quantity value and deletes the collected data when new data arrives the data collection
terminal. Analog or digital without storage memory are typical examples of this. Any
measurement equipment used in the medical domain for monitoring without storage
functionality also comes under this category.

Some embedded systems stores the collected data for processing and analysis. Such
systems incorporate a built-in/plug in storage memory for storing the captured data. Some of
them give the user meaningful representation of the collected data by visual
(graphical/quantitative) or audible means using display units (lcd, led, buzzers etc).

Examples of this type are measuring instruments with storage memory and monitoring
instruments with storage memory used in applications.

Certain embedded systems store the data and will not give a representation of the same to
the user, whereas the data is used for internal processing.

A digital camera is typical example of embedded systems with data

collection/storage/representation of data. Images are captured and the captured image may be
stored within the memory of the camera. The captured image can also be presented to the user
through a graphic LCD unit.

Data communication: Embedded data communication systems are deployed in

applications ranging from complex satellite communication system to simple home networking
systems.

As the data collected by an embedded terminal may require transferring of the same to
some other system located remotely. The transmission is achieved either by a wire-line medium
or by a wireless medium. The wire-line medium was the most common choice in all olden days
embedded systems. As technology is changing, wireless communication is becoming the de-
facto standard for data communication in embedded systems.

A wireless medium offers cheaper connectivity solutions and make the communication
link free from the hassle of wire bundles. Data can either be transmitted by analog means (or)
by digital means modern industry trends are setting towards digital communication.

The data collecting embedded terminal itself can incorporate data communication units
like wireless modules (Bluetooth, zigbee, wifi, gprs etc….) or wire-line modules (rs-232, USB,
TCP/IP etc…).

Certain embedded systems act as dedicated transmission unit between the sending and
receiving terminals offering sophisticated functionalities like data packeting, encrypting and
decrypting ,network hubs ,routers, switches etc are typical examples of dedicated data
transmission embedded systems.

They act as a mediators in data communication and provide various features like data
security, monitoring etc.

Data (signal) processing: The data (voice, image, video, electrical signals and other
measurable quantities) collected by embedded systems may be used for various kinds of data
processing. Embedded systems with signal processing functionalities are employed in
applications demanding signal processing like speech coding synthesis, audio video codes,
transmission applications etc.

A digital hearing aid is a typical example of an embedded systems employing data

processing. Digital hearing aid improves the hearing capacity of hearing impaired persons.

Monitoring: Embedded systems falling under this category are specially designed for
monitoring purpose. Almost all embedded products coming under the medical domain are with
monitoring functions only. They are used for determining the state of some variables using
input sensors. They cannot impose control over variable. A very good example is the electro
cardiogram (ECG) machine for monitoring the heartbeat of a patient. The sensors used in ECG
are the different electrodes connected to the patient body.
 Some other examples of embedded systems with monitoring functions are measuring
instruments like digital CRO, digital millimeters, logic analysis etc used in control and
instruments application. They are used for knowing (monitoring)the status of some variables
like current, voltages etc . They cannot control the variables in turn.

Control: Embedded systems with control functions over some variables according to the
changes in the input variables. A system with control functionalities contains both sensors and
actuators. Sensors are connected to the input port for capturing the changes in the environment
variable (or ) measuring variable. The actuators connected to the output port are controlled
according to the changes in the input variable to put an impact on the controlling variable to
bring the controlled variable to the specified range. Ac’s used in our home to control the room
temperature to a specified limit is a typical example for embedded system for control purpose.
An ac contains a room temperature sensing element(sensor) which may be thermisted and a
hand held unit for setting up (feeding) the desired temperature.

The handled unit may be connected to the central embedded unit residing inside the ac
through a wireless link (or) through a wired link. The air compressor is controlled according to
the current room temperature and the desired temperature set by the end users.

Here the input variable is the current room temperature and the controlled variable is
also the room temperature . The controlling variable is cool air flow by the compressor unit. If
the controlled variable and the input variable are not at the same value .The controlling
variables tries to equalize them through taking actions on the cool air flow.

Application specific user interface (ASUI): These are embedded systems with
application specific user interfaces like buttons, switches, keypads, lights, bells, display units
etc..

Mobile phone is an example for this .In mobile phone the user interface is provided
through the keypad, graphic LCD module system speaker, vibration alert etc...

Characteristics of embedded systems: Characteristics of embedded systems are unlike

general purpose computing system. Embedded systems posses certain specific characteristics
and these are unique to each other embedded systems. Some of the important characteristics of
embedded systems are

 Application and domain specific

 Reactive and real time
 Operates in harsh environment
 Distributed
 Small size and weight
 Power concerns

Application and domain specific: if you closely observe any embedded system you will
find that each embedded system is having certain functions to perform and they are developed
in such a manner to do the intended functions only.

They cannot be used for any other purpose. It is the major criterion which distinguishes
an embedded system from a general purpose system. For example you cannot replace the
embedded control unit of your microwave oven with your air conditioners embedded control
unit. Because the ECU of the microwave oven and the air conditioners are specially designed to
perform certain tasks. Also you cannot replace an embedded control unit developed for a
particular domain say telecom with another control unit designed to serve another domain like
consumer electronics.

Reactive and real time: embedded systems are in constant interaction with the real world
through sensors and user defined input devices which are connected to the input port of the
systems. Any changes happening in the real world are captured by the sensors (or) input
devices in real time and the control algorithm running inside the unit reacts in a designed
manner to bring the controlled output variables to the desired level. The event may be periodic
one (or) an unpredicted one. If the event is an unpredicted one then such systems should be
scheduled to capture the events without missing them. Embedded systems produce changes in
output in response to the changes in the input. So they are generally referred as reactive
systems.
 Real time system operation means the timing behavior of the system should be
deterministic meaning the system should respond to requests or tasks in a known amount of
time. A real time system should not miss any deadlines for tasks for tasks (or) operations. It is
not necessary that all embedded systems should be real time in operations. Embedded systems
(or) applications which are mission critical like flight control systems, anti lock brake systems
(ABS) etc are examples of real time systems. The design of an embedded real time system
should take the worst case scenario into consideration.

Operates in harsh environment: It is not necessary that all embedded systems should be
deployed in controlled environments. The environments in which the embedded systems
deployed may be dusty one or a high temperature zone or an area subject to vibrations and
shock. Systems placed in such areas should be capable to with stand all these adverse operating
conditions. The design should take care of the operating conditions of the area where the
system is going to implement. For example if the system needs to be deployed in a high
temperature zone then all components used in the system should be of high temperature
grade .Here we cannot go for a compromise in cost. Also proper shock absorption technique
should be provided to systems which are going to be commissioned in places subject to high
shock. Power supply fluctuations, corrosion and components aging etc are the other factors that
need to be taken into consideration for embedded systems to work in harsh environment.

Distributed: The term distributed means that embedded system may be a part of larger
systems. May number of such distributed embedded systems from single large embedded
control units. An automatic vending machine is a typical example of this. The vending machine
contains a card reader, a vending unit etc. Each of them are independent embedded units but
they work together to perform the overall vending function. Another example is the ATM. Any
ATM contains a card reader embedded units, responsible for reading and validating the users
ATM card, Transaction unit for performing transactions, a currency counter for
dispatching/vending/currency to the authorized person and a printer unit for printing the
transaction details. We can visualize these as independent embedded systems But they work
together to achieve a common goal.
 Another typical example of a distributed embedded systems is the supervisory control and
the data acquisition (SCADA) system used in control for instrumentation applications, which
contains physically distributed individual embedded control connected to a supervisory module.

Small size and weight: product aesthetics is an important factor in choosing a product. For
example when you plan to buy a new mobile phone, you make a comparative study on the pros
and cons of the product available in the market. Definitely the product aesthetics (size, weight,
etc) will be one of the deciding factors to choose a product. People believe in the phrase “small
is beautiful”. Moreover it is convenient to handle a compact device than a bulky product. In
embedded domain also compactness is significant deciding factor. Most of the application
demands small sized and low weight products.

Power concerns: power management is another important factor that needs to be considered
in designing embedded systems. Embedded systems should be designed in a such way as to
minimize the heat dissipation by the system. The production of high amount of heat demands
cooling requirements like cooling fans which in turn occupies additional space and make the
system bulky. Now a day’s ultra low power components are available in the market. Select the
design according to the low dropout regulators and controller/processors with power saving
modes. Also power management is critical constraint in battery operated application. The more
the power consumption the less is the battery life.

Applications of embedded systems: Embedded systems are used in different

applications like automobiles, telecommunications, smart cards, missiles, satellites, computer
networking and digital consumer electronics.

Embedded systems in automobiles and telecommunications:

 Motor and cruise control system

 Body and engine safety
 Entertainment and multimedia in car
 E-com and mobile access
 Robotics in assembly line
 Wireless communication
 Mobile computing and networking
Embedded systems in smart cards, missiles and satellites:

 Security systems
 Telephone and banking
 Defense and aerospace
 Communication

Embedded systems in peripherals and computer networking:

 Display and monitors

 Networking systems
 Image processing
 Network cards and printers

Embedded systems in consumer applications:

 Digital cameras
 Set up boxes
 High definition TV’s
 DVD’s

1.2 Other Applications

Personal digital assistants (PDAs) or handheld computers are generally considered

embedded devices because of the nature of their hardware design, even though they are more
expandable in software terms.

This line of definition continues to blur as devices expand. With the introduction of the
OQO Model2 with the Windows XP operating system and ports such as a USB port—both
features usually belong to "general purpose computers", —the line of nomenclature blurs even
more.
 Physically, embedded systems ranges from portable devices such as digital watches and
MP3 players, to large stationary installations like traffic lights, factory controllers, or the
systems controlling nuclear power plants.

 Avionics, such as inertial guidance systems, flight control hardware/software and other
integrated systems in aircraft and missiles
 Cellular telephones and telephone switches
 Engine controllers and antilock brake controllers for automobiles
 Home automation products, such as thermostats, air conditioners, sprinklers, and
security monitoring systems
 Handheld calculators
 Handheld computers
 Household appliances, including microwave ovens, washing machines, television sets,
DVD players and recorders
 Medical equipment
 Personal digital assistant
 Videogame consoles
  Computer peripherals such as routers and printers.
 Industrial controllers for remote machine operation.

Wireless Sensor Networks (WSN), sometimes called Wireless Sensor and Actuator Network
(WSAN) is a wireless network consisting of spatially distributed autonomous sensors to
monitor physical or environmental conditions, such as temperature, humidity, pressure, etc.
A WSN system incorporates a gateway that provides wireless connectivity back to the wired
world and distributed nodes. To cooperatively pass their data through the network to a main
location. The more modern networks are bidirectional, also enabling control of sensor activity.

The increasing demand for energy, the continuous reduction in existing sources of fossil fuels
and the growing concern regarding environment pollution, have pushed mankind to explore
new non-conventional, renewable energy resources such as solar energy, wind energy, etc for
the production of electrical energy. Since India receives sunlight all 12 moths of a year. Hence
utilizing it in the different fields is a wise idea. India is an agricultural country. India ranks
second worldwide in farm output. At present, farmer manually irrigates land at regular interval.
This process sometimes consumes more water or sometimes the water reaches late due to
which the crops get dried. Solar powered smart farming irrigation system not only overcomes
this problem but also provide clean source of energy.

INTRODUCTION TO THE PROJECT

The human voice is a phenomenon that is highly dependent on the speaker who
produces it. Studies show that no two individuals sound exactly the same, the
acoustic aspects of what distinguishes the differences between the speakers voices
are uncertain and difficult to separate from signal aspects that reflect segment
recognition. According to Ramachandran et al, the sources of variation among
speakers are threefold, the difference in speaking styles (accent included), the
difference in vocal tract shapes and vocal cords, and how speakers express
themselves to convey a particular message (words or phrases used). However,
because a speaker's tendency to use certain phrases, words and syntactic
structures (referring to the third source) is not easy to quantify or control in an
experiment, automatic speaker recognition systems utilize the first two sources
only by exploring the low-level acoustic features of a speech signal. Speaker
recognition is an important topic in signal processing and has a variety of
applications, especially in security systems. Voice controlled systems and devices
rely heavily on speaker recognition. Some applications of speaker recognition
security control for confidential information, verifying customers for bank
transactions, forensics and remote access to computers. Researchers have
published a considerable amount of publications in the speaker recognition field,
however, very few attempts (if not none) have been published on the development
of speaker recognition systems developed using South African under-resourced
languages. According to, the South African official languages are still classified
as being highly under-resourced. Sepedi is one South African language that is
largely spoken in the Limpopo Province of South Africa. Census 2011 reports
that, Speedy is the language spoken by most persons in Limpopo province with
more than 2.8 million speakers, followed by Xitsonga with approximately one
million speakers..

EXISTING SYSTEM:

This paper presents the development of an automatic speaker recognition system that
incorporates classification and recognition of Sepedi home language speakers. The system uses
machine learning algorithms that learns features extracted from the Sepedi speech data to train
the classifier model. The system can be used to automatically authenticate speaker identities
using their voices to allow only the identified persons an access right to information systems or
to facilities that need to be protected from the intrusion of unauthorized persons.

PROPOSED SYSTEM
A. Fundamental tasks of Speaker Recognition
Speaker recognition has two fundamental tasks, namely speaker verification and speaker
identification, Speaker verification is the task of determining whether an unknown voice is
from a particular enrolled speaker, the speaker has to claim an identity and the system validates
the claimed identity. Applications of speaker verification include telephone banking, computer
login, cellular telephone fraud prevention and calling cards. Speaker identification is the task of
associating an unknown voice with one from a set of enrolled speakers. The speaker in this case
provides a voice sample (without claiming an identity) and the system determines to which of
the known set of enrolled speakers does the voice sample belongs to. Potential speaker
identification applications include automatic speaker labelling of recorded meetings for
speaker-dependent audio indexing and intelligent answering machines with personalized caller
greetings.
B. Classification of Speaker Recognition Systems
Speaker recognition systems are further classified by the constraints placed on the text of the
speech used in the system, the classification can either be text-dependent or text independent.
In the text-dependent case, the spoken text or phrase used to train and test the system is fixed
for each speaker. Text-dependent speaker recognition systems are used mostly in services such
as access control and telephone based services, where users are considered to be cooperative. In
the text-independent case, the spoken phrase or text used to train and test is not fixed. Text-
independent speaker recognition systems are the most flexible and widely used in events where
speakers can be considered non-cooperative users, as they do not specifically wish to be
recognized such as forensic analysis and surveillance procedures. Text-dependent recognition
achieves higher recognition performance than the text-independent recognition , However, due
to the flexibility that the text-independent recognition provides, the increasing development
trend is in the building of the textindependent recognition systems .
C. Phases of Speaker Recognition
A speaker recognition system is composed of two different phases, a training phase and a test
phase, In the training phase, a speakers voice is recorded and a number of audio features
vectors are extracted to form a unique model (voice-print) that uniquely identifies the speaker.
In the testing phase, (also known as the recognition phase) the voice sample provided is
compared against the previously created model.
D. Applications of Speaker Recognition Systems
The goal of speaker recognition is to automatically recognize who is speaking on the basis of
individual information included in the speakers utterances or voice. Research in the field of
speaker recognition has now spanned over five decades and has shown fruitful results Per Say
is Speaker Recognition fundamental tasks (Verification and Identification). The two phases of
Speaker Recognition (Training and Testing). of the leading provider of advanced voice
biometric recognition products used to verify speakers identities, securely, conveniently and
effectively Speaker recognition technology is applicable to a wide area of applications,
however, the two main areas of application are authentication and forensics. Speaker
recognition for authentication allows uses or automated systems to identify a person using their
voices. This type of authentication method is known as biometric person authentication. A
biometric person authentication can be used to supplement the knowledge-based (usernames
and passwords) or token-based (use of physical tokens such as identity cards) methods to allow
users safe and secure access to information or knowledge systems . Forensic is an important
applicable service that can be supported by speaker recognition technology. A lot of
information is exchanged between parties (law-abiding) in telephone conversations. If there is a
recorded speech sample during the commission of crime, the suspects voices can be compared
with the recorded voice to help resolve the crime. Proving the identity of a recorded voice can
help convict a person suspected of committing a crime or discharge an innocent suspect in
court.

Block Diagram:

DISPLAY
POWER SUPPLY

RASPBERRY
SPEAKER PI 3

Motor

Block diagram of module

Block diagram explanation:

This Project mainly consists of Power Supply section, Raspberry pi3 , display,
speaker,motor.

Power supply:
This section is meant for supplying Power to all the sections mentioned above. It basically
consists of a Transformer to step down the 230V ac to 9V ac followed by diodes. Here diodes
are used to rectify the ac to dc. After rectification the obtained rippled dc is filtered using a
capacitor Filter. A positive voltage regulator is used to regulate the obtained dc voltage.

Raspberry Pi 3:

Quad-Core 64bit CPU, Wi-Fi & Bluetooth! The Raspberry Pi 3 Model B is the third generation
Raspberry Pi. This powerful credit-card sized single board computer can be used for many
applications. the Broadcom BCM2837 System-On-Chip (SoC) incorporates four superior ARM
Cortex-A53 process centers running at1.2GHz with 32Kb Level one and 512Kb Level a couple
of reserve memory, a Video Core IV illustrations processor, and is associated with a 1GB
LPDDR2 memory module on the back of the board. It also alternatives 40-pins broadly useful
info yields (GPIO) and enhanced property with Bluetooth Low Energy (BLE) and BCM43143
Wi-Fi on board.
 HARDWARE COMPONENTS
RASPBERRY PI 3 MODEL B

The Raspberry Pi is an index card-sized single-board computer developed in the UK by

the Raspberry Pi Foundation with the intention of promoting the teaching of basic computer
science in schools.

 The Raspberry Pi 3 delivers 10 times the processing capacity of Raspberry Pi 1 model.

 This second generation Raspberry Pi has an upgraded Broadcom BCM2837 processor,
which is a powerful ARM Cortex-A53 based 64 bit quad-core processor that runs at
1.2GHz. The board also features an increase in memory capacity to 1Gbyte.

Fig 2.1 Raspberry Pi-3 Board

The Raspberry Pi has a Broadcom BCM2836 system on a chip which includes an 900 MHz
32- bit quad-core ARM Cortex-A7, Video Core IV GPU, and was originally shipped with 256
megabytes of RAM, later upgraded (Model B & Model B+) to512 MB. It does not include a
built-in hard disk or solid-state drive, but it uses an SD card for booting and persistent storage,
with the Model B+ using a Micro SD as shown in figure 2.1.

The Foundation provides Debian and Arch Linux ARM distributions for download. Tools are
available for Python as the main programming language, with support for BBC BASIC (via
the RISC OS image or the Brandy Basic clone for Linux), C, Java, and Perl.

Please note: The new processor on the Pi 2 means that you will need to update your existing
SDcard or create a new SD card with your operating system (Raspbian, Arch, XBMC, NooBs,
etc) because you cannot plug in older cards from a Pi 2 into a Pi 3 without upgrading with sudo
apt-getupgradeon the Pi 1 first.

2. 1.1 SPECIFICATIONS AND CONNECTORS

 Chip Broadcom BCM2837 SoC

 Core architecture: Quad-core ARM Cortex-A53
 1.2GHz 32/64- bit quad-coreARM Cortex-A53
 GPU Dual Core Video Core IV® Multimedia Co-Processor
 Provides Open GL ES 2.0, hardware-accelerated Open VG, and 1080p30 H.264
 high-profile decode Capable of 1Gpixel/s, 1.5Gtexel/s or 24GFLOPs with texture
filtering and DMA infrastructure
 Memory 1 GB LPDDR2 RAM at 900 MHz
 Operating System Boots from Micro SD card, running a version of the Linux
operating system
 Dimensions 85 x 56 x 17mm
 Power Micro USB socket 5V, 2A
 Operating system-Linux, Fedora, RISC OS, Free BSD, Net BSD
 Power 2.5w, CPU-ARM116JZF-S.
 RAM memory 512mb
  Total data can be Storage in SD card slot.
 It Supports Broadcom video core graphics.

2.1.2 CONNECTORS

 Ethernet 10/100 Base T Ethernet socket.

 Video Output HDMI.
 Composite RCA (PAL and NTSC).
 Audio Output 3.5mm jack, HDMI.
 USB 4 x USB 2.0 Connector.
 GPIO Connector 40-pin 2.54 mm (100 mil) expansion header: 2x20 strip.
 Providing 40 GPIO pins as well as +3.3 V, +5 V and GND supply lines.
 Camera Connector 15-pin MIPI Camera Serial Interface (CSI-2).
 JTAG Not populated.
 Display Connector Display Serial Interface (DSI) 15 ways flat flex cable connector
with two data lanes and a clock lane.

2. 1.3 APPLICATIONS

 Living room PC.

 A mini Web browser.
 Arcade Pi.
 Robotics.
 Firefox OS on Pi.
 Bit Torrent Server.
 Weather Station.
 Home automation.
 Raspberry Pi Cluster.
 Device Information.
2.1.4ARCHITECTURAL OVERVIEW

The Raspberry Pi is based on the Broadcom BCM2837 system on a chip (SoC), which
includes an 64-bitquad-coreARM Cortex-A53, 1.2GHz processor, Video Core IV GPU, and
was originally shipped with 512 megabytes of RAM, later upgraded (Model B & Model
B+) to 1GB. The system has Secure Digital (SD) or Micro SD (Model B+) sockets for boot
media and persistent storage.

The Raspberry Pi 3 delivers 6 times the processing capacity of previous models. This
second generation Raspberry Pi has an upgraded Broadcom BCM2837 processor, which is a
powerful ARM Cortex-A53 based quad-core processor that runs at 1.2GHz. The board also
features an increase in memory capacity to 1Gbyte.

2.1.5 PERFORMANCE

The Raspberry Pi 3, with a quad-core Cortex-A53 processor, is described as 10 times the
performance of a Raspberry Pi 1. This was suggested to be highly dependent upon
task threading and instruction set use. Benchmarks showed the Raspberry Pi 3 to be
approximately 80% faster than the Raspberry Pi 2 in parallelized tasks.

Raspberry Pi 2 includes a quad-core Cortex-A7 CPU running at 900 MHz and 1 GB RAM.
It is described as 4–6 times more powerful than its predecessor. The GPU is identical to the
original. In parallelized benchmarks, the Raspberry Pi 2 could be up to 14 times faster than a
Raspberry Pi 1 Model B+. While operating at 700 MHz by default, the first generation
Raspberry Pi provided a real-world performance roughly equivalent to 0.041 GFLOPS. 

On the CPU level, the performance is similar to a 300 MHz Pentium II of 1997–99. The

GPU provides 1 Gpixel/s or 1.5 Gtexel/s of graphics processing or 24 GFLOPS of general
purpose computing performance. The graphical capabilities of the Raspberry Pi are roughly
equivalent to the performance of the Xbox of 2001.
2.1.6 GPIO PIN DESCRIPTION FOR RASPBERRY PI
 Fig 2.1.6 GPIO Pin Diagram of Raspberry Pi-3

In addition to the familiar USB, Ethernet and HDMI ports, the R-Pi offers lower-level
interfaces intended to connect more directly with chips and subsystem modules. These GPIO
(general purpose I/O) signals on the 2x13 header pins include SPI, I²C, serial UART, 3V3 and
5V power. These interfaces are not "plug and play" and require care to avoid miswriting. The
pins use a 3V3 logic level and are not tolerant of 5V levels, such as you might find on a 5V
powered Arduino. CSI (camera serial interface) can be used to connect the 5 MP camera
available. Not yet software-enabled is the flex cable connectors with DSI (display serial
interface) and a serial link inside the HDMI connector called CEC.

General Purpose Input/ Output is a generic pin on a chip whose behavior (including
whether it is an input or output pin) can be controlled (programmed) through software. The
Raspberry Pi allows peripherals and expansion boards to access the CPU by exposing the
inputs and outputs.

GPIO functions:
General-purpose input/output (GPIO) is a generic pin on an integrated circuit whose behavior,
including whether it is an input or output pin, can be controlled by the user at run time. GPIO
pins have no special purpose defined and go unused by default. The idea is that sometimes
the system integrator building a full system that uses the chip might find it useful to have a
handful of additional digital control lines, and having these available from the chip can avoid
the effort of having to arrange additional circuitry to provide them.

GPIOs are used in:

 Devices with pin scarcity: integrated circuits such as system-on-a-chip, embedded and
custom hardware, and programmable logic devices (for example, FPGAs.
 Multifunction chips: power managers, audio codes, and video cards.
 Embedded applications (for example, Arduino, Beagle Bone, PSoC kits and Raspberry
Pi) make heavy use of GPIO for reading from various environmental sensors (IR, video,
temperature, 3-axis orientation, and acceleration), and for writing output to DC
motors (via PWM), audio, LCD displays, or LEDs for status.

GPIO capabilities may include:

 GPIO pins can be configured to be input or output

 GPIO pins can be enabled/disabled
 Input values are readable (typically high=1, low=0)
 Output values are writable/readable
 Input values can often be used as IRQs (typically for wakeup events)

GPIO peripherals vary quite widely. In some cases, they are very simple, a group of pins
that can be switched as a group to either input or output. In others, each pin can be set up
flexibly to accept or source different logic voltages, with configurable drive strengths and pull
ups/downs. The input and output voltages are typically, though not universally, limited to the
supply voltage of the device with the GPIOs on and may be damaged by greater voltages.

A GPIO pin's state may be exposed to the software developer through one of a number of
different interfaces, such as a memory mapped peripheral, or through dedicated IO port
instructions. Some GPIOs have 5 V tolerant inputs: even when the device has a low supply
voltage (such as 2 V), the device can accept 5 V without damage.

Using the Pi as a computer feels fast and 'desktop like' - not sluggish! Particularly for
developers, compiling code on the Pi 3 is 4x faster and the extra RAM helps a lot too, so most
programs can now be compiled directly on the Pi. In this, pi3 have inbuilt Bluetooth and Wi-Fi.

2.2 SDHC (Secure Digital High Capacity Card)

Secure Digital (SD) is a nonvolatile memory card used extensively in portable devices, such

as mobile phones, digital cameras, GPS navigation devices, handheld consoles, and tablet
computers. The Secure Digital standard was introduced in August 1999 as an evolutionary
improvement over Multi Media Cards(MMC). The Secure Digital standard is maintained by
the SD Association (SDA). SD technologies have been implemented in more than 400 brands
across dozens of product categories and more than 8,000 models.

The Secure Digital format includes four card families available in three different form
factors. The four families are the original Standard-Capacity (SDSC), the High-Capacity
(SDHC), the extended-Capacity (SDXC), and the SDIO, which
combines input/output functions with data storage. The three form factors are the original size,
the mini size, and the micro size. Electrically passive adapters allow a smaller card to fit and
function in a device built for a larger card.

2.2.1 SD card working

The first-generation Secure Digital (SDSC or Secure Digital Standard Capacity) card was
developed to improve on the Multimedia Card (MMC) standard, which continued to evolve, but
in a different direction. The SD cards changed the MMC design in several ways:
  Asymmetrical slots in the sides of the SD card prevent inserting it upside down, while
an MMC goes in most of the way but makes no contact if inverted.
 Most SD cards are 2.1 mm (0.083 inches) thick, compared to 1.4 mm (0.055 inches) for
MMCs. The SD specification defines a card called Thin SD with a thickness of 1.4 mm,
but they are rare, as the SDA went on to define even smaller form factors.
 The card's electrical contacts are recessed beneath the surface of the card, protecting
them from contact with a user's fingers.
 The SD specification envisioned capacities and transfer rates exceeding those of MMC,
and these have both grown over time.

 Full-sized SD cards do not fit in the slimmer MMC slots, and there are other issues that
affect the ability to use one format in a host device designed for the other.

2.2.2 SDHC card working

The Secure Digital High Capacity (SDHC) format, announced in January 2006 and defined in
Version 2.0 of the SD specification, supports cards with capacities upto 32 GB. The SDHC
trademark is licensed to ensure compatibility The major compatibility issues between SDHC
and SDSC cards are the redefinition of the Card-Specific Data (CSD) register in Version 2.0,
and the fact that SDHC cards are shipped preformatted with the FAT32 file system.

Fig2.2.2 SDHC card

Version 2.0 also introduces a High-speed bus mode for both SDSC and SDHC cards,
which doubles the original Standard Speed clock to produce 25 MB/s. SDHC host devices are
required to accept older SD cards. However, older host devices do not recognize SDHC or
SDXC memory cards, although some devices can do so through a firmware upgrade. Older
Windows operating systems released before Windows 7 require patches or service packs
support access to SDHC cards.

Table2.2.2 SD card Pin Specification

2.2.3 SD card Compatibility

Host devices that comply with newer versions of the specification provide backward
compatibility and accept older SD cards. For example, SDXC host devices accept all previous
families of SD memory cards, and SDHC host devices also accept standard SD cards. Older
host devices generally do not support newer card formats, and even when they might support
the bus interface used by the card, there are several factors that arise:

 A newer card may offer greater capacity than the host device can handle (over 4 GB for
SDHC, over 32 GB for SDXC).
 A newer card may use a file system the host device cannot navigate (FAT32 for
SDHC, exFAT for SDXC)
 Use of an SDIO card requires the host device be designed for the input/output functions
the card provides.
 The hardware interface of the card was changed starting with the Version 2.0 (new
high-speed bus clocks, redefinition of storage capacity bits) and SDHC family (Ultra-
high speed (UHS) bus)
  Some vendors produced SDSC cards above 1 GB before the SDA had standardized a
method of doing so.

2.2.4 Speeds

SD card speed is customarily rated by its sequential read or write speed. The sequential
performance aspect is the most relevant for storing and retrieving large files (relative to block
sizes internal to the flash memory), such as images and multimedia. Small data (such as file
names, sizes, and timestamps) falls under the much lower speed limit of random access, which
can be the limiting factor in some use cases. With early SD cards, the speed was specified as a
"times" ("×") rating, which compared the average speed of reading data to that of the
original CD-ROM drive. This was superseded by the Speed Class Rating, which guarantees a
minimum rate at which data can be written to the card.

The newer families of SD card improve card speed by increasing the bus rate (the frequency of
the clock signal that strobes information into and out of the card). Whatever the bus rate, the
card can signal to the host that it is "busy" until a read or a write operation is complete.
Compliance with a higher speed rating is a guarantee that the card limits its use of the "busy"
indication
2.3 Pi CAMERA:

Raspberry Pi camera module, a low cost hardware module that supports still picture and video
recording and is the first official hardware add-on for the Raspberry Pi. We'll also look at the
camera's features and specifications, how to set it up and examine the software that supports it.

Camera Details

Since 2012, the Raspberry Pi Foundation had been reporting that an official camera module
was in development. In May 2013, an announcement was made by RS Components and
Premier Farnell/Element 14, distribution partners of Raspberry Pi, that the camera module was
available (it is also available from other sources) and sells for retail €30 or US$25. 
The camera consists of a small (25mm by 20mm by 9mm) circuit board, which connects to the
Raspberry Pi's Camera Serial Interface (CSI) bus connector via a flexible ribbon cable. The
camera's image sensor has a native resolution of five megapixels and has a fixed focus lens.
The software for the camera supports full resolution still images up to 2592x1944 and video
resolutions of 1080p30, 720p60 and 640x480p60/90. The camera module is shown below:

Fig 2.3 PI Camera

Installation involves connecting the ribbon cable to the CSI connector on the Raspberry Pi
board. This can be a little tricky, but if you watch the videos that demonstrate how it is done,
you shouldn't have any trouble.

The OV5647 is a low voltage, high performance, 5 megapixel CMOS image sensor that
provides 2592x1944 video output using OmniBSI™ technology. It provides multiple resolution
raw images via the control of the serial camera control bus or MIPI interface.

The OV5647 has an image array capable of operating up to 15 fps in 2592x1944 resolution
with user control of image quality, data transfer, camera functions through the SCCB interface.
The OV5647 uses innovative OmniBSI technology to improve the sensor performance without
the physical and optical trade-off.

For customized application, the OV5647 includes a one-time programmable (OTP) memory.

Features:

 1.4 µm x 1.4 µm pixel with Omni BSI technology for high performance (high
sensitivity, low crosstalk, low noise)
  optical size of 1/4"
 automatic image control functions: automatic exposure control (AEC), automatic white
balance (AWB), automatic band filter (ABF), automatic 50/60 Hz luminance detection,
and automatic black level calibration (ABLC)
 programmable controls for frame rate, AEC/AGC 16-zone size/position/weight control,
mirror and flip, cropping, windowing, and panning
 image quality controls: lens correction, defective pixel canceling
 support for output formats: 8-/10-bit raw RGB data support for video or snapshot
operations support for LED and flash strobe mode
 support for internal and external frame synchronization for frame exposure mode
 support for horizontal and vertical sub-sampling standard serial SCCB interface
 digital video port (DVP) parallel output interface MIPI interface (two lanes)
 32 bytes of embedded one -time programmable (OTP) memory
 on- chip phase lock loop (PLL)
 embedded 1.5V regulator for core power
 programmable I/O drive capability, I/O tri-state configurability
 support for black sun cancellation
Specifications:
 active array size: 2592 x 1944 power supply:
 core: 1.5V + 5% (with embedded 1.5V regulator) analog: 2.6 ~ 3.0V (2.8V typical)
 I/O: 1.7V ~ 3.0V
 power requirements: active: TBD standby: TBD
 temperature range:
 operating: -30°C to 70°C stable image: 0°C to 50°C
 output formats: 8-/10-bit RGB RAW output
 lens size: 1/4"
 lens chief ray angle: 24 input clock frequency: 6~27 MHz
 S/N ratio: TBD
 dynamic range: TBD
 maximum image transfer rate:
  QSXGA (2592 x 1944): 15 fps 1080p: 30 fps
 960p: 45 fps 720p: 60 fps
 VGA (640 x 480): 90 fps QVGA (320 x 240): 120 fps
 sensitivity: TBD
 shutter: rolling shutter / global shutter
 maximum exposure interval: 1968 x tROW
 pixel size: 1.4 µm x 1.4 µm
 well capacity: TBD dark current: TBD
 fixed pattern noise (FPN): TBD
 image area: 3673.6 µm x 2738.4 µm
 die dimensions: 5520 µm x 4700 µm

2.4 Speakers or Head phones:

Fig: 2.4 Diagram of a dynamic loudspeaker.

A loudspeaker (or "speaker") is an electroacoustic transducer that converts an electrical

signal into sound. The speaker moves in accordance with the variations of an electrical signal
and causes sound waves to propagate through a medium such as air or water.After the
acoustics of the listening space, loudspeakers (and other electroacoustic transducers) are the
most variable elements in a modern audio system and are usually responsible for most
distortion and audible differences when comparing sound systems.

It is an inexpensive, low fidelity 3½-inch speaker, which is typically found in small

radios.

Theory of speaker:
 The term "loudspeaker" may refer to individual transducers (known as "drivers") or to complete
speaker systems consisting of an enclosure including one or more drivers. To adequately reproduce a
wide range of frequencies, most loudspeaker systems employ more than one driver, particularly for
higher sound pressure level or maximum accuracy. Individual drivers are used to reproduce different
frequency ranges. The drivers are named subwoofers (for very low frequencies); woofers (low
frequencies); mid-range speakers (middle frequencies); tweeters (high frequencies); and sometimes super
tweeters, optimized for the highest audible frequencies.

The terms for different speaker drivers differ, depending on the application. In two-way systems
there is no mid-range driver, so the task of reproducing the mid-range sounds falls upon the woofer and
tweeter. Home stereos use the designation "tweeter" for the high frequency driver, while professional
concert systems may designate them as "HF" or "highs". When multiple drivers are used in a system, a
"filter network", called a crossover, separates the incoming signal into different frequency ranges and
routes them to the appropriate driver. A loudspeaker system with n separate frequency bands is described
as "n-way speakers": a two-way system will have a woofer and a tweeter; a three-way system employs a
woofer, a mid-range, and a tweeter.

Driver design:

Fig:2.4.1 Diagram of cut away view of a dynamic loudspeaker.

Application tips of APR9600 for better sound replay quality:

1. Use a good quality 8 Ohm speaker with a cavity such as speakers for computer sound systems. Do not
use a bare speaker which gives you degraded sound.
2. For better sound replay quality, speak with a distance to the on-board microphone and speak clearly.
Also, keep the background noise as low as possible.

3. For even better sound replay quality, use microphone input or Audio Line In input. If Audio Line In is
used, the amplitude of input signal should be < 100 mV p-p.

2.5 Power Supply

All digital circuits require regulated power supply. In this article, we are going to learn how
to get a regulated positive supply from the mains supply.
 

Figure 2.5.1 block diagram of a fixed regulated power supply

TRANSFORMER

A transformer consists of two coils also called as “WINDINGS” namely PRIMARY &
SECONDARY.
They are linked together through inductively coupled electrical conductors also called as
CORE. A changing current in the primary causes a change in the Magnetic Field in the core &
this, in turn, induces an alternating voltage in the secondary coil. If load is applied to the
secondary then an alternating current will flow through the load. If we consider an ideal
condition then all the energy from the primary circuit will be transferred to the secondary
circuit through the magnetic field.

So

The secondary voltage of the transformer depends on the number of turns in the Primary as well as in the
secondary.

RECTIFIER
A rectifier is a device that converts an AC signal into DC signal. For rectification purpose
we use a diode, a diode is a device that allows current to pass only in one direction i.e. when the
anode of the diode is positive with respect to the cathode also called as forward biased
condition & blocks current in the reversed biased condition.
Rectifier can be classified as follows:
1)      Half Wave rectifier.

This is the simplest type of rectifier as you can see in the diagram a half wave rectifier consists
of only one diode. When an AC signal is applied to it during the positive half cycle the diode is
forward biased & current flows through it. But during the negative half cycle diode is reverse
biased & no current flows through it. Since only one half of the input reaches the output, it is
very inefficient to be used in power supplies.

  2)      Full wave rectifier.

Half wave rectifier is quite simple but it is very inefficient, for greater efficiency we would like
to use both the half cycles of the AC signal. This can be achieved by using a center tapped
transformer i.e. we would have to double the size of secondary winding & provide connection
to the center. So during the positive half cycle diode, D1 conducts & D2 is in reverse biased
condition. During the negative half cycle diode, D2 conducts & D1 is reverse biased. Thus we
get both the half cycles across the load.
One of the disadvantages of Full Wave Rectifier design is the necessity of using a center
tapped transformer, thus increasing the size & cost of the circuit. This can be avoided by using
the Full Wave Bridge Rectifier.

 3)      Bridge Rectifier.

 As the name suggests it converts the full wave i.e. both the positive & the negative half
cycle into DC thus it is much more efficient than Half Wave Rectifier & that too without using
a center tapped transformer thus much more cost effective than Full Wave Rectifier.

Full Bridge Wave Rectifier consists of four diodes namely D1, D2, D3, and D4. During the
positive half cycle diodes, D1 & D4 conduct whereas in the negative half cycle diodes D2 &
D3 conduct thus the diodes keep switching the transformer connections so we get positive half
cycles in the output.

  If we use a center tapped transformer for a bridge rectifier we can get both positive &
negative half cycles which can thus be used for generating fixed positive & fixed negative
voltages.

FILTER CAPACITOR

Even though half wave & full wave rectifier give DC output, none of them provides a
constant output voltage. For this, we require to smoothen the waveform received from the
rectifier. This can be done by using a capacitor at the output of the rectifier this capacitor is also
called as “FILTER CAPACITOR” or “SMOOTHING CAPACITOR” or “RESERVOIR
CAPACITOR”. Even after using this capacitor a small amount of ripple will remain.
We place the Filter Capacitor at the output of the rectifier the capacitor will charge to the
peak voltage during each half cycle then will discharge its stored energy slowly through the
load while the rectified voltage drops to zero, thus trying to keep the voltage as constant as
possible.
 If we go on increasing the value of the filter capacitor then the Ripple will decrease. But
then the costing will increase. The value of the Filter capacitor depends on the current
consumed by the circuit, the frequency of the waveform & the accepted ripple.

 
Where,
Vr= accepted ripple voltage. ( should not be more than 10% of  the voltage)
I= current consumed by the circuit in Amperes.
F= frequency of the waveform. A half wave rectifier has only one peak in one cycle so F=25hz
Whereas a full wave rectifier has Two peaks in one cycle so F=100hz.

VOLTAGE REGULATOR 

A Voltage regulator is a device which converts varying input voltage into a constant

regulated output voltage. Voltage regulator can be of two types

1)      Linear Voltage Regulator

      Also called as Resistive Voltage regulator because they dissipate the excessive voltage
resistively as heat.
2)      Switching Regulators.
 They regulate the output voltage by switching the Current ON/OFF very rapidly. Since their
output is either ON or OFF it dissipates very low power thus achieving higher efficiency as
compared to linear voltage regulators. But they are more complex & generate high noise due to
their switching action. For low level of output power, switching regulators tend to be costly but
for higher output wattage they are much cheaper than linear regulators.
The most commonly available Linear Positive Voltage Regulators are the 78XX series
where the XX indicates the output voltage. And 79XX series is for Negative Voltage
Regulators.

After filtering the rectifier output the signal is given to a voltage regulator. The
maximum input voltage that can be applied at the input is 35V. Normally there is a 2-3 Volts
drop across the regulator so the input voltage should be at least 2-3 Volts higher than the output
voltage. If the input voltage gets below the Vmin of the regulator due to the ripple voltage or
due to any other reason the voltage regulator will not be able to produce the correct regulated
voltage.

CIRCUIT DIAGRAM
 Fig 2.5.2. Circuit Diagram of power supply

IC 7805:

7805 is an integrated three-terminal positive fixed linear voltage regulator. It supports an

input voltage of 10 volts to 35 volts and output voltage of 5 volts. It has a current rating of 1
amp although lower current models are available. Its output voltage is fixed at 5.0V. The 7805
also has a built-in current limiter as a safety feature. 7805 is manufactured by many companies,
including National Semiconductors and Fairchild Semiconductors.

The 7805 will automatically reduce output current if it gets too hot. The last two digits
represent the voltage; for instance, the 7812 is a 12-volt regulator. The 78xx series of regulators
is designed to work in complement with the 79xx series of negative voltage regulators in
systems that provide both positive and negative regulated voltages since the 78xx series can't
regulate negative voltages in such a system.

The 7805 & 78 is one of the most common and well-known of the 78xx series regulators, as
it's small component count and medium-power regulated 5V make it useful for powering TTL
devices.

MOTORS

DEFINITION

Motor is a device that creates motion, not an engine; it usually refers to either an electrical
motor or an internal combustion engine.
It may also refer to:

 Electric motor, a machine that converts electricity into a mechanical motion

o AC motor, an electric motor that is driven by alternating current
 Synchronous motor, an alternating current motor distinguished by a rotor
spinning with coils passing magnets at the same rate as the alternating
current and resulting magnetic field which drives it
 Induction motor, also called a squirrel-cage motor, a type of
asynchronous alternating current motor where power is supplied to the
rotating device by means of electromagnetic induction
o DC motor, an electric motor that runs on direct current electricity
 Brushed DC electric motor, an internally commutated electric motor
designed to be run from a direct current power source
 Brushless DC motor, a synchronous electric motor which is powered by
direct current electricity and has an electronically controlled
commutation system, instead of a mechanical commutation system based
on brushes
o Electrostatic motor, a type of electric motor based on the attraction and repulsion
of electric charge
o Servo motor, an electric motor that operates a servo, commonly used in robotics
o Internal fan-cooled electric motor, an electric motor that is self-cooled by a fan,
typically used for motors with a high energy density

TYPES OF MOTORS

Industrial motors come in a variety of basic types. These variations are suitable for many
different applications. Naturally, some types of motors are more suited for certain applications
than other motor types are. This document will hopefully give some guidance in selecting these
motors.

AC Motors

The most common and simple industrial motor is the three phase AC induction motor,
sometimes known as the "squirrel cage" motor. Substantial information can be found about any
motor by checking its (nameplate).

Advantages

 Simple Design
 Low Cost
 Reliable Operation
 Easily Found Replacements
 Variety of Mounting Styles
 Many Different Environmental Enclosures
Simple Design

The simple design of the AC motor -- simply a series of three windings in the exterior (stator)
section with a simple rotating section (rotor). The changing field caused by the 50 or 60 Hertz
AC line voltage causes the rotor to rotate around the axis of the motor.

The speed of the AC motor depends only on three variables:

1. The fixed number of winding sets (known as poles) built into the motor, which
determines the motor's base speed.
2. The frequency of the AC line voltage. Variable speed drives change this frequency to
change the speed of the motor.
3. The amount of torque loading on the motor, which causes slip.

Low Cost

The AC motor has the advantage of being the lowest cost motor for applications requiring more
than about 1/2 hp (325 watts) of power. This is due to the simple design of the motor. For this
reason, AC motors are overwhelmingly preferred for fixed speed applications in industrial
applications and for commercial and domestic applications where AC line power can be easily
attached. Over 90% of all motors are AC induction motors. They are found in air conditioners,
washers, dryers, industrial machinery, fans, blowers, vacuum cleaners, and many, many other
applications.

Reliable Operation

The simple design of the AC motor results in extremely reliable, low maintenance operation.
Unlike the DC motor, there are no brushes to replace. If run in the appropriate environment for
its enclosure, the AC motor can expect to need new bearings after several years of operation. If
the application is well designed, an AC motor may not need new bearings for more than a
decade.

Easily Found Replacements

The wide use of the AC motor has resulted in easily found replacements. Many manufacturers
adhere to either European (metric) or American (NEMA) standards. (For Replacement Motors)

Variety of Mounting Styles

AC Motors are available in many different mounting styles such as:

 Foot Mount
 C-Face
 Large Flange
 Vertical
 Specialty

DC Motors

The brushed DC motor is one of the earliest motor designs. Today, it is the motor of choice in
the majority of variable speed and torque control applications.

Advantages

 Easy to understand design

 Easy to control speed
 Easy to control torque
 Simple, cheap drive design

Easy to understand design

The design of the brushed DC motor is quite simple. A permanent magnetic field is created in
the stator by either of two means:

 Permanent magnets
 Electro-magnetic windings

If the field is created by permanent magnets, the motor is said to be a "permanent magnet DC
motor" (PMDC). If created by electromagnetic windings, the motor is often said to be a "shunt
wound DC motor" (SWDC). Today, because of cost-effectiveness and reliability, the PMDC
motor is the motor of choice for applications involving fractional horsepower DC motors, as
well as most applications up to about three horsepower.

At five horsepower and greater, various forms of the shunt wound DC motor are most
commonly used. This is because the electromagnetic windings are more cost effective than
permanent magnets in this power range.

Caution: If a DC motor suffers a loss of field (if for example, the field power connections are
broken), the DC motor will immediately begin to accelerate to the top speed which the loading
will allow. This can result in the motor flying apart if the motor is lightly loaded. The possible
loss of field must be accounted for, particularly with shunt wound DC motors.

Opposing the stator field is the armature field, which is generated by a changing
electromagnetic flux coming from windings located on the rotor. The magnetic poles of the
armature field will attempt to line up with the opposite magnetic poles generated by the stator
field. If we stopped the design at this point, the motor would spin until the poles were opposite
one another, settle into place, and then stop -- which would make a pretty useless motor!

However, we are smarter than that. The section of the rotor where the electricity enters the rotor
windings is called the commutator. The electricity is carried between the rotor and the stator by
conductive graphite-copper brushes (mounted on the rotor) which contact rings on stator.
Imagine power is supplied:

The motor rotates toward the pole alignment point. Just as the motor would get to this point, the
brushes jump across a gap in the stator rings. Momentum carries the motor forward over this
gap. When the brushes get to the other side of the gap, they contact the stator rings again and --
the polarity of the voltage is reversed in this set of rings! The motor begins accelerating again,
this time trying to get to the opposite set of poles. (The momentum has carried the motor past
the original pole alignment point.) This continues as the motor rotates.

In most DC motors, several sets of windings or permanent magnets are present to smooth out
the motion.
Easy to control speed

Controlling the speed of a brushed DC motor is simple. The higher the armature voltage, the
faster the rotation. This relationship is linear to the motor's maximum speed.

The maximum armature voltage which corresponds to a motor's rated speed (these motors are
usually given a rated speed and a maximum speed, such as 1750/2000 rpm) are available in
certain standard voltages, which roughly increase in conjuntion with horsepower. Thus, the
smallest industrial motors are rated 90 VDC and 180 VDC. Larger units are rated at 250 VDC
and sometimes higher.

Specialty motors for use in mobile applications are rated 12, 24, or 48 VDC. Other tiny motors
may be rated 5 VDC.

Most industrial DC motors will operate reliably over a speed range of about 20:1 -- down to
about 5-7% of base speed. This is much better performance than the comparible AC motor.
This is partly due to the simplicity of control, but is also partly due to the fact that most
industrial DC motors are designed with variable speed operation in mind, and have added heat
dissipation features which allow lower operating speeds.

Easy to control torque

In a brushed DC motor, torque control is also simple, since output torque is proportional to
current. If you limit the current, you have just limited the torque which the motor can achieve.
This makes this motor ideal for delicate applications such as textile manufacturing.

Simple, cheap drive design

The result of this design is that variable speed or variable torque electronics are easy to design
and manufacture. Varying the speed of a brushed DC motor requires little more than a large
enough potentiometer. In practice, these have been replaced for all but sub-fractional
horsepower applications by the SCR and PWM drives, which offer relatively precisely control
voltage and current. Common DC drives are available at the low end (up to 2 horsepower) for
under US$100 -- and sometimes under US$50 if precision is not important.
Large DC drives are available up to hundreds of horsepower. However, over about 10
horsepower careful consideration should be given to the price/performance tradeoffs with AC
inverter systems, since the AC systems show a price advantage in the larger systems. (But they
may not be capable of the application's performance requirments).

Disadvantages

 Expensive to produce
 Can't reliably control at lowest speeds
 Physically larger
 High maintenance
 Dust

WORKING OF DC MOTOR

In any electric motor, operation is based on simple electromagnetism. A current-

carrying conductor generates a magnetic field; when this is then placed in an external magnetic
field, it will experience a force proportional to the current in the conductor, and to the strength
of the external magnetic field. As you are well aware of from playing with magnets as a kid,
opposite (North and South) polarities attract, while like polarities (North and North, South and
South) repel. The internal configuration of a DC motor is designed to harness the magnetic
interaction between a current-carrying conductor and an external magnetic field to generate
rotational motion.
Principle

When a rectangular coil carrying current is placed in a magnetic field, a torque acts on the coil
which rotates it continuously.

When the coil rotates, the shaft attached to it also rotates and thus it is able to do mechanical
work.

Every DC motor has six basic parts -- axle, rotor (a.k.a., armature), stator, commutator, field
magnet(s), and brushes. In most common DC motors (and all that BEAMers will see), the
external magnetic field is produced by high-strength permanent magnets 1. The stator is the
stationary part of the motor -- this includes the motor casing, as well as two or more permanent
magnet pole pieces. The rotor (together with the axle and attached commutator) rotate with
respect to the stator. The rotor consists of windings (generally on a core), the windings being
electrically connected to the commutator. The above diagram shows a common motor layout --
with the rotor inside the stator (field) magnets.

The geometry of the brushes, commentator contacts, and rotor windings are such that when
power is applied, the polarities of the energized winding and the stator magnet(s) are
misaligned, and the rotor will rotate until it is almost aligned with the stator's field magnets. As
the rotor reaches alignment, the brushes move to the next commentator contacts, and energize
the next winding. Given our example two-pole motor, the rotation reverses the direction of
current through the rotor winding, leading to a "flip" of the rotor's magnetic field, driving it to
continue rotating.

In real life, though, DC motors will always have more than two poles (three is a very common
number). In particular, this avoids "dead spots" in the commutator. You can imagine how with
our example two-pole motor, if the rotor is exactly at the middle of its rotation (perfectly
aligned with the field magnets), it will get "stuck" there. Meanwhile, with a two-pole motor,
there is a moment where the commutator shorts out the power supply (i.e., both brushes touch
both commutator contacts simultaneously). This would be bad for the power supply, waste
energy, and damage motor components as well. Yet another disadvantage of such a simple
motor is that it would exhibit a high amount of torque "ripple" (the amount of torque it could
produce is cyclic with the position of the rotor).

Construction and Working

Parts of a DC Motor

Armature

A D.C. motor consists of a rectangular coil made of insulated copper wire wound on a soft iron
core. This coil wound on the soft iron core forms the armature. The coil is mounted on an axle
and is placed between the cylindrical concave poles of a magnet.

Commutator

A commutator is used to reverse the direction of flow of current. Commutator is a copper ring
split into two parts C1 and C2. The split rings are insulated form each other and mounted on the
axle of the motor. The two ends of the coil are soldered to these rings. They rotate along with
the coil. Commutator rings are connected to a battery. The wires from the battery are not
connected to the rings but to the brushes which are in contact with the rings.

Brushes

Two small strips of carbon, known as brushes press slightly against the two split rings, and the
split rings rotate between the brushes.

The carbon brushes are connected to a D.C. source.

Working of a DC Motor

When the coil is powered, a magnetic field is generated around the armature. The left side of
the armature is pushed away from the left magnet and drawn towards the right, causing
rotation.
When the coil turns through 900, the brushes lose contact with the commutator and the current
stops flowing through the coil.

However the coil keeps turning because of its own momentum.

Now when the coil turns through 1800, the sides get interchanged. As a result the commutator
ring C1 is now in contact with brush B2 and commutator ring C2 is in contact with brush B1.
Therefore, the current continues to flow in the same direction.
PARAMETRS OF THE DC MOTRS

1. Direction of rotation
2. Motor Speed
3. Motor Torque
4. Motor Start and Stop

Direction of Rotation

A DC Motor has two wires. We can call them the positive terminal and the negative terminal,
although these are pretty much arbitrary names (unlike a battery where these polarities are vital
and not to be mixed!). On a motor, we say that when the + wire is connected to + terminal on a
power source, and the - wire is connected to the - terminal source on the same power source,
the motor rotates clockwise (if you are looking towards the motor shaft). If you reverse the wire
polarities so that each wire is connected to the opposing power supply terminal, then the motor
rotates counter clockwise. Notice this is just an arbitrary selection and that some motor
manufacturers could easily choose the opposing convention. As long as you know what rotation
you get with one polarity, you can always connect in such a fashion that you get the direction
that you want on a per polarity basis.
DC Motor Rotation vs Polarity

Facts:

 DC Motor rotation has nothing to do with the voltage magnitude or the current
magnitude flowing through the motor.
 DC Motor rotation does have to do with the voltage polarity and the direction of the
current flow.

DC Motor Speed

Whereas the voltage polarity controls DC motor rotation, voltage magnitude controls motor
speed. Think of the voltage applied as a facilitator for the strengthening of the magnetic field.
In other words, the higher the voltage, the quicker will the magnetic field
become strong. Remember that a DC motor has an electromagnet and a series of permanent
magnets. The applied voltage generates a magnetic field on the electromagnet portion. This
electromagnet field is made to oppose the permanent magnet field. If the electromagnet field is
very strong, then both magnetic entities will try to repel each other from one side, as well as
atract each other from the other side. The stronger the induced magnetic field, the quicker will
this separation/attaction will try to take place. As a result, motor speed is directly proportional
to applied voltage.
Motor Speed Curve

One aspect to have in mind is that the motor speed is not entirely lineal. Each motor will have
their own voltage/speed curve. One thing I can guarantee from each motor is that at very low
voltages, the motor will simply not move. This is because the magnetic field strength is not
enough to overcome friction. Once friction is overcome, motor speed will start to increase as
voltage increase.

The following video shows the concept of speed control and offers some ideas on how this can
be achieved.

Motor Torque

In the previous segment I kind of described speed as having to do with the strength of the
magnetic field, but this is in reality misleading. Speed has to do with how fast the magnetic
field is built and the attraction/repel forces are installed into the two magnetic structures. Motor
strength, on the other hand, has to do with magnetic field strength. The stronger the
electromagnet attracts the permanent magnet, the more force is exerted on the motor load.

Per example, imagine a motor trying to lift 10 pounds of weight. This is a force that when
multiplied by a distance (how much from the ground we are lifting the load) results in WORK.
This WORK when exerted through a predetermined amount of time (for how long we are
lifting the weight) gives us power. But whatever power came in, must come out as energy can
not be created or destroyed. So that you know, the power that we are supplying to the motor is
computed by

P = IV

Where P is power, I is motor current and V is motor voltage

Hence, if the voltage (motor speed) is maintained constant, how much load we are moving must
come from the current. As you increase load (or torque requirements) current must also
increase.

Motor Loading

One aspect about DC motors which we must not forget is that loading or increase of torque can
not be infinite as there is a point in which the motor simply can not move. When this happens,
we call this loading “Stalling Torque”. At the same time this is the maximum amount of current
the motor will see, and it is refer to Stalling Current. Stalling deserves a full chapter as this is a
very important scenario that will define a great deal of the controller to be used. I promise I will
later write a post on stalling and its intricacies.

Motor Start and Stop

You are already well versed on how to control the motor speed, the motor torque and the motor
direction of rotation. But this is all fine and dandy as long as the motor is actually moving. How
about starting it and stopping it? Are these trivial matters? Can we just ignore them or should
we be careful about these aspects as well? You bet we should!

Starting a motor is a very hazardous moment for the system. Since you have an inductance
whose energy storage capacity is basically empty, the motor will first act as an inductor. In a
sense, it should not worry us too much because current can not change abruptly in an inductor,
but the truth of the matter is that this is one of the instances in which you will see the highest
currents flowing into the motor. The start is not necessarily bad for the motor itself as in fact
the motor can easily take this Inrush Current. The power stage, on the other hand and if not
properly designed for, may take a beating.

Once the motor has started, the motor current will go down from inrush levels to whatever load
the motor is at. Per example, if the motor is moving a few gears, current will be proportional to
that load and according to torque/current curves.

Stopping the motor is not as harsh as starting. In fact, stopping is pretty much a breeze. What
we do need to concern ourselves is with how we want the motor to stop. Do we want it to coast
down as energy is spent in the loop, or do we want the rotor to stop as fast as possible? If the
later is the option, then we need braking. Braking is easily accomplished by shorting the motor
outputs. The reason why the motor stops so fast is because as a short is applied to the motor
terminals, the Back EMF is shorted. Because Back EMF is directly proportional to speed,
making Back EMF = 0, also means making speed = 0.

MOTORDRIVER CIRCUIT

The name "H-Bridge" is derived from the actual shape of the switching circuit which control
the motoion of the motor. It is also known as "Full Bridge". Basically there are four switching
elements in the H-Bridge as shown in the figure below.
As you can see in the figure above there are four switching elements named as "High side left",
"High side right", "Low side right", "Low side left". When these switches are turned on in pairs
motor changes its direction accordingly. Like, if we switch on High side left and Low side right
then motor rotate in forward direction, as current flows from Power supply through the motor
coil goes to ground via switch low side right. This is shown in the figure below.

Similarly, when you switch on low side left and high side right, the current flows in opposite
direction and motor rotates in backward direction. This is the basic working of H-Bridge. We
can also make a small truth table according to the switching of H-Bridge explained above.

Truth Table
High Left High Right Low Left Low Right Description
On Off Off On Motor runs clockwise
Off On On Off Motor runs anti-clockwise
On On Off Off Motor stops or decelerates
Off Off On On Motor stops or decelerates

As already said, H-bridge can be made with the help of trasistors as well as MOSFETs, the only
thing is the power handling capacity of the circuit. If motors are needed to run with high current
then lot of dissipation is there. So head sinks are needed to cool the circuit.

Now you might be thinkin why i did not discuss the cases like High side left on and Low side
left on or high side right on and low side right on. Clearly seen in the diagra, you don't want to
burn your power supply by shorting them. So that is why those combinations are not discussed
in the truth table.

SOFTWARE DESCRIPTION

3.1 REQUIREMENTS FOR SOFTWARE INSTALLATION

The Raspberry Pi is a very capable minicomputer that's small enough to fit in your pocket (even
though you'd probably never keep it there). You can add up to two USB devices (which, in
most cases, will be a mouse and keyboard) and connect it to the internet via wired Ethernet.
The Raspberry Pi is powerful and inexpensive, allowing you to create a home media center,
internet radio, or even your own VPN server on the cheap. That said, a little setup and a lot of
other parts are required to get it up and running. In this post, we'll take a look at how to get a
Raspberry Pi set up so you can start using it for a variety of purposes.

The Raspberry Pi may be a computer all on its own, but you'll still need a lot of other things to
make it work:

 A Raspberry Pi model B

 An HDMI- or composite video-capable television or monitor: An HDMI-capable

monitor because it offers better resolution and built-in sound. You can use analog if you want,
however.
 An HDMI or composite video cable: You'll need this cable to connect Raspberry Pi to
television or monitor.

 A 4GB Class 4 SD card (or better) and a card reader: Most SD cards will work, but some
aren't compatible and will, therefore, cause issues.

 A USB keyboard and mouse: Any standard USB keyboard or mouse will do. Wireless
(non-Bluetooth) peripherals worked for me, but I had to unplug them and plug them back in
after the Raspberry Pi booted.

 An Ethernet cable: Any standard Ethernet cable will do. We need this to connect to
network and the internet, however, so we can leave this item out if desired.

 A good quality, micro USB power supply that can provide at least
700mA at 5V: Most modern Smartphone chargers supply 700mA at 5V, but not all do. If it
offers at least that much power, probably good to go.

3.2 SET UP RASPBERRY PI WITH A BASIC OPERATING SYSTEM

Fig 3.1 Selection Of Ultra SD Card For Os Installation

We can use the Raspberry Pi for all sorts of different thing some of which may require their
own special operating systems but to start out, it's a good idea to get acquainted with the Pi by
installing Raspbian, a Raspberry Pi-focused version of Linux. Here's what you need to do.
3.2.1 PREPARE SD CARD

First things first, we need to prepare SD card for the Raspberry Pi. This involves formatting it
properly and putting Raspbian, the standard Raspberry Pi operating system, on it. The steps
vary depending on whether or not on a Mac, Linux, or Windows PC, so just follow the
instructions for operating system below.

3.2.2 WINDOWS

Fig 3.2 Windows Image File

 Download the latest version of Raspbian and unzip the .img file inside. (It's almost 500MB
so it may take a little while to download.)

 Download Win32DiskImager and unzip the application (.exe file) inside.

 Insert SD card into Windows PC using a card reader.

 Open Win32DiskImager.exe, the application just downloaded, by double-clicking on it. If

we are running Windows 7 or 8, right click on it and choose "Run as Administrator" instead.

 If SD card isn't automatically detected by the application, click on the drop-down menu at
the top right (labeled "Device") and choose it from the list.

 In the image file section of the application, click the little folder icon and choose the
Raspbian .img file just downloaded.

 Click the Write button and wait for Win32DiskImager to do its thing. When it finishes, we
can safely eject SD card and insert it into your Raspberry Pi.
3.3 OPERATING SYSTEM

Fig 3.3 Image File Selection

 Download the latest version of Raspbian  and unzip the .img file inside. (It's almost
500MB so it may take a little while to download.)

 Download RPi- sd card builder (be sure to pick the appropriate version for installed
version of OS X) and unzip the application.

 Insert SD card into Mac using a card reader.

 Open RPi-sd card builder. You'll immediately be asked to choose a Raspbian image.
Choose the .img file you downloaded earlier.
 It will ask SD card is connected. Since we inserted it earlier, it is, so go ahead and click
Continue. It will present SD card options. If you only have one inserted, you won't see
anything else in the list and it'll be checked. If not, just check only the card you want to
use and click OK.

 Enter your administrator password and click OK.

 You'll be asked if the SD card was ejected. This is supposed to happen, as the application
needs to unamount it so it can perform a direct copy. Double-check that your SD card is
no longer available in the Finder. Do not remove it from your USB port. When you're
sure, click Continue.

 RPi-sd card builder finishes preparing your SD card, safely eject it and insert it into your
Raspberry Pi unit.

 RPi-sd card builder isn't an app so much as it's an Automator action that acts like one.
Some people have reported issues using it, so if you run into problems just open up the
Terminal app (Hard Drive → Applications → Utilities → Terminal) and follow the
instructions for Linux.
3.4 LINUX

Fig 3.4 Linux Image file

 Download the latest version of Raspbian and unzip the .img file inside. (It's almost 500MB
so it may take a little while to download.)

 Open up Linux terminal emulator of choice.

Insert SD card. If it mounts automatically, unmount it but make note of its location (e.g.
/dev/disk2s1). Even though the card is unmounted, do not remove it from the reader

In the command line, you're going to need to type a single line to copy the contents of the
Raspbian .img file to your SD card. It'll look like this

Sudodd if=PATH_TO_IMG_FILE of=

PATH_TO_SD_CARD_MOUNT_POINT bs=1m
 You'll need to replace PATH_TO_IMG_FILE with the path to the .img file and
PATH_TO_SD_CARD_MOUNT_POINT with the path to the SD card mount point. (The
bs=1m just specifies the byte size and you don't need to mess with it.) When you're done, the
command should look something like this

Sudoddif=/Users?adachis/Downloads/2012-12-16-

wheezy-raspbian.imgof=/dev/disk2s1bs=1m

It's very important that you do not get this information wrong or you could end up writing
Raspbian to the wrong disk and cause serious data loss. Be careful! When you're sure you've
got everything right, press enter. You will be prompted for the root password (or your
administrator password for those of you on OS X who are using this method). Type it and press
enter.

It'll take some time for the dd command to copy everything over to your SD card. While
that happens, it'll probably look like the Terminal froze up. Don't worry, it's still working and
will likely take longer than an average 2GB copy to an SD card (so be patient). When it
finishes, the command prompt will return and your SD card will be ready. Eject it safely and
stick it in your Raspberry Pi.

3.5 HOOK UP RASPBERRY PI

Fig 3.5 Raspberry Pi Connection

Connecting everything to Raspberry Pi is pretty easy. Start by connecting the HDMI cable from
the unit to your television (or monitor). If you're using the analog composite video connection
instead, connect it to your television (or monitor) and plug in a set of speakers to the 3.5mm
audio jack. If you want your Raspberry Pi to connect to your network and/or the internet, use an
Ethernet cable to connect it to your router. If you've set up a computer before, this should all
seem familiar.

Finally, you need to connect Raspberry Pi with a micro USB power adapter. As mentioned
earlier, you need an adapter that can provide at least 700mA at 5V. Fortunately, you probably
already have one. Many Smartphone and tablet chargers utilize micro USB and provide 700mA
at 5V (or more). You can find out by reading the small text on the plug and looking for the
output section. (Note: it may say 0.7A (or higher) instead of 700mA.) When you have a
compatible power adapter on hand, connect it to a micro USB cable and then connect that cable
to the micro port on your Raspberry Pi unit. It'll turn on all by itself and you should see it boot
for the first time.

3.6CONFIGURE RASPBERRY PI

RASPBERRY PI CONFIGURATION
Once Raspberry Pi boots for the first time you'll need to configure a few things. You'll know
it's ready for you when you see a Raspi-config window with a pretty big list of settings. You
can mess with the others if you feel so inclined, but the only thing you really need to do is
expand the file system so you can use up all the space on your SD card. To do so, follow these
steps:

1. Select expand rootfs (the second option) and press enter.

2. Confirm that you want to expand the file system and let Raspbian do its thing.

3. When you're returned to the configuration list, go all the way to the bottom and select the
Finish option.
4. It'll ask you if you want to reboot. Choose yes.

Your Raspberry Pi will reboot and take a little longer this time because it needs to expand the
file system. After a about a minute, you'll be asked to log in.

You'll need to use the default username and password:

Username: pi

Password: raspberry

Once you've logged in, you've got yourself a working Raspberry Pi.

CONCLUSION:
This paper reported on the development of a textindependent speaker recognition system based
on machine learning algorithms. The paper briefly described all the stages (training and testing)
which covered voice activity detection, feature extraction, model training, evaluation and
graphical user interface. The dataset of Sepedi speech data was obtained from the NCHLT
project. Voice activity detection was performed using Long-Term Spectral Divergence
algorithm. Features were extracted using pyAudioAnalysis library. We used the SVM, KNN,
RF and MLP implemented in WEKA to train the models. We also applied Auto-WEKA to
determine the best algorithm. It was observed that MLPs performed well on the given dataset,
however, Auto-WEKA selected Random Forest as the best algorithm, which was then
implemented on the GUI. The future work will focus on:- • Increasing the data set to test the
effect of data size on performance. • Increasing number of speakers to test the complexity of
our approach. • Adding noise to speech samples to test the scalability and robustness of the
models. • Adding recent deep learning algorithms.

REFERENCES
[1] T. Kinnunen and H. Li, "An overview of text-independent speaker recognition: From
features to supervectors," Speech communication, vol. 52, no. 1, pp. 12-40, 2010.
[2] R. R Ramachandran, K. R. Farrell, R. Ramachandran, and R. J. Mammone, "Speaker
recognitiongeneral classifier approaches and data fusion methods," Pattern Recognition, vol.
35, no. 12, pp. 2801-2821, 2002.
[3] N. Singh, R. Khan, and R. Shree, "Applications of speaker recognition," Pmcedia
engineering, vol. 38, pp. 3122-3126, 2012.
[4] A. Larcher, K. A. Lee, B. Ma, and H. Li, "Text-dependent speaker verification: Classifiers,
databases and rsr2015," Speech Communication, vol. 60, pp. 56-77, 2014.
[5] E. Aliyu, O. Adewale, and A. Adetunmbi, "Development of a textdependent speaker
recognition system," International Journal of Computer Applications, vol. 69, no. 16, 2013.
[6] E. Variani, X. Lei, E. McDermott, I. Lopez-Moreno, and J. Gonzalez- Dominguez, "Deep
neural networks for small footprint text-dependent speaker verification." in ICASSP, vol. 14,
2014, pp. 4052-4056.
[7] M. J. Adamski, "A speaker recognition solution for identification and authentication,"
Master's thesis, University of Johannesburg, Gauteng Province, South Africa, 2013.
[8] S. Furui, "50 years of progress in speech and speaker recognition research," ECTI
Transactions on Computer and Information Technology (ECTI-CIT), vol. 1, no. 2, pp. 64-74,
2005.
[9] F. Bimbot, J.-F. Bonastre, C. Fredouille, G. Gravier, I. Magrin- Chagnolleau, S. Meignier,
T. Merlin, J. Ortega-Garcfa, D. Petrovska- Delacr£taz, and D. A. Reynolds, "A tutorial on text-
independent speaker verification," EURASIP Journal on Advances in Signal Processing, vol.
2004, no. 4, p. 101962, 2004.
[10] D. A. Reynolds, "Automatic speaker recognition using gaussian mixture speaker models,"
in The Lincoln Laboratory Journal. Citeseer, 1995