Artificial Intelligence in Tele-Vision

vinegarclothAI and Robotics

Jul 17, 2012 (6 years and 7 days ago)


S.PraveenKumar et. al. / International Journal of Engineering Science and Technology
Vol. 2(11), 2010, 6593-6597
Artificial Intelligence in Tele-Vision
, A.Anand
, S.M.Subramanian
Assistant professor
, Professor
Department of Electronics and Communication Engineering
Saveetha Engineering College


The digital television (DTV) semiconductor market is expected to grow to $7.7billion in 2010 and to $10
billion in 2013, an 18% CAGR. Flat screen DTV semiconductor revenue will grow to $6.9 billion in 2010 and to $9
billion in 2013, a 25% CAGR.. The above mentioned numbers provide huge opportunity for TV vendors but in a
highly-competitive environment. Introducing value added features is one of the ways to address the intense
competition. Increase in the number of television channels and programs have provided a number of choices to the
consumer, which has led to confusion. The user interested favorite item can be a song, a comedy scene. This may be
transmitted in a particular channel while he is viewing another one and the user might end up missing his favorite
program .This write-up proposes a solution by which the manufacturer will be able to provide an advanced feature,
which ensures the consumers get a great experience without missing their favorites.

1. Problem statement
When switching on the television, the user normally surfs all the channels for the favorite in a particular set of
channels. The favorite could be a particular song, album or a comedy scene. But most of the time the user finds none
of the channels is broadcasting the favorite, so chooses to view any one of the channels. In the mean time the
favorite might be broadcasted in another channel, which the user is not aware of. Because of this the user ends up
surfing the channels every now and then for the favorite. Some times the user might also end-up getting the favorite,
but watching it in the middle will create frustration.
2. Proposed solution
Artificial Intelligence in Tele-Vision (AITV) solution allows the users to surf the television channels on behalf of
the user in the background. This is done automatically based on the user’s choice. When the favorite is broadcasted
in any of the pre defined channels, i-TV switches to that channel automatically. This is achieved by a small change
in the hardware and

adding a minimal memory to the system. This memory will store the samples of user favorites.
The AITV is a feature which compares the incoming signals of pre defined channels and changes the channel
automatically if it matches with the users favorites. With the help of software different list of favorites for different
users and this list is created and managed. The AITV design chooses the sound signal for comparison in order to
minimize the memory required and to reduce the complexity. The project involves the Digital Signal Processing for
comparing the sound signals.

3. Competitive approaches
Based on the initial research it is found that no major Television Manufactures are providing this feature. However
the similar approach is exists in ‘smart remote controllers’ where the user can program and set a pre-defined timer to
switch the channel. For example when the prime time news is broadcasted at 9 PM in Channel number 15, it can be
automatically switched. But the limitation of this solution is only the pre-defined program with a known timing can
be chosen. With AITV, this is made completely dynamic. The user can store any of the favorite occurring in any of
the channels at any time. This basically would provide more freedom to the user and store the favorites as and when
it occurs. Another advantage of this solution is any sample can be stored and compared. For example in a top-10 list
of songs (which is a complete program) the user may be interested with only two of them. In order to watch the two
songs the user may end up watching whole program. Using AITV the user can configure and watch the program
only when the favorite songs are broadcasted. This provides huge amount of options to the user to watch a very
small segment in a program.

ISSN: 0975-5462
S.PraveenKumar et. al. / International Journal of Engineering Science and Technology
Vol. 2(11), 2010, 6593-6597
4. AITV Solution details
Introduction to Television Receiver:
The television receiver translates the pulses of electric current from the antenna or cable to images and sound. A

traditional television set integrates the receiver, audio system, and picture tube into one device. Following are the
main blocks of the television receiver:
 Tuner Module: The tuner blocks all the signals, other than that of the desired channel. It helps in surfing
the different channels depending on given control voltage from the control unit. It generates the
Intermediate Frequency (IF) signal of the particular tuned channel.
 Sound Section: The audio system consists of a discriminator, which translates the audio portion of the
carrier wave into an electronic audio signal. The amplifier strengthens the audio signal from the
discriminator and sends it to the speaker, which converts the electrical waves into sound waves.
 Video Section: This section is responsible for displaying the image from the received electrical signal with
the help of picture tube, set of amplifiers, filters and color decoders.
 Control Unit: This is a microcontroller based unit, which receives the user input and control the tuner and
other sections. This section performs operations initiated by the user, surfing the channels and other settings
like volume control, display settings etc.
AITV solution blocks:
To incorporate this solution into the television receiver, some more blocks needs to be added. These additional
blocks will be in parallel along with the existing modules. The additional modules will store and compare the
samples with the user’s choice. The detailed block diagram is mentioned below.
The additional tuner automatically switches between the predefined channels in a time sliced manner. It runs in the
background, controlled automatically by the control unit.
The additional sound section separates the sound signal from the output of the additional tuner. This block provides
the incoming audio stream (which is chosen by the tuner at that instance) to the processing unit in the time sliced

As shown in the block diagram, the additional tuner will provide the information of the programs telecasted in the
other predefined channels.
ISSN: 0975-5462
S.PraveenKumar et. al. / International Journal of Engineering Science and Technology
Vol. 2(11), 2010, 6593-6597
The processing unit will compare the incoming audio signal with the stored samples of the favorite list (internal
memory) and initiate the control unit.
In order to store the user favorites, a memory unit needs to be added. When the user stores the favorite, the sample is
stored in the memory. This memory unit is basically a Non Volatile Memory (NVM) where the contents can be
retrieved as and when required.
The control unit takes care of the controlling both the tuner operations as per the user input. It also switches the
additional tuner between the predefined channels periodically.
When the favorite is appearing, the user will activate the ‘store mode’ by pressing the button in the TV remote. On
pressing the button, the control unit routes the current audio signal to the processing unit and the sample get stored.
When the favorite is broadcasted once again in any of the predefined channels, the processing unit will compare the
incoming signal with the stored samples.
When it matches, it will issue an interrupt to the control unit. Upon receiving the interrupt, the control unit will
switch the television to the channel where the favorite is broadcasted.

5. RC-5 Coding
The IR receiver module at the TV "tunes" to a certain frequency and ignores all other IR received. The best
frequency for the job is between 30 and 60 kHz, the most used is around 36 kHz.
First of all, RC5 standard that uses fixed bit length and fixed quantity of bits. Each time you press a button at the
remote control, it sends a train of 14 bits, 1.728ms per bit, the whole train is repeated every 130ms if you keep the
button pressed.
The quantity of pulses used is 32 pulses per each half of the bit, 64 pulses per bit. So, a bit "0" to be transmitted it
means 32 square pulses of 27µs each, then 32 x 27µs of silence.
The bit "1" is the opposite, 32 x 27µs of silence followed

by 32 square pulses of 27µs (i.e.) also called as

You can see the 14 bits of the RC-5 system above. The RED bits are level "ON", while Blue are "OFF". The first
two bits, #1 and #2, are called AGC calibration.
They are "ON" level, and serve to calibrate the IR Receivers Auto Gain Control. In the Philips remotes, the bit #3 is
the CHECK bit, every time you press a key at the remote, even pressing repeatedly the same key, this bit flips state.
This feature is interesting. Suppose you pressed number "1" at the remote (trying to select channel 15 at TV) and
holding it for 2 seconds, then your other hand just blocks the InfraRed signal.
The next 5 bits, #4 to #8, are used for SYSTEM ADDRESS, or to identify which kind of device should execute the
COMMAND bits. For example, TV set uses ADDRESS ZERO.
Bit #8 is the Less Significant Bit. The next 6 bits, #9 to #14, are used for COMMAND information to the device
selected at the ADDRESS bits.
Bit #14 is the LESS SIGNIFICANT BIT, and it is last transmitted. Remember that there is a delay of 105ms
between pulse trains.

ISSN: 0975-5462
S.PraveenKumar et. al. / International Journal of Engineering Science and Technology
Vol. 2(11), 2010, 6593-6597

6. Schematic


ISSN: 0975-5462
S.PraveenKumar et. al. / International Journal of Engineering Science and Technology
Vol. 2(11), 2010, 6593-6597

7. Logic

8. Design challenges
Storing the samples in an optimal way is a big challenge. As the additional memory unit cannot be of large storage
capacity, it mandates efficient storage algorithms or methods.
Matching the stored signals with the real time signals is another challenge. It is quiet possible that the solution will
end up choosing the wrong channel because of the incorrect comparisons.
[1] Gulati.R.R, “Modern Television Practice, Principle of Technology and Servicing “, New age International Pvt., Ltd., 2002.
[2] Grob.B, Herndon. C.E., “Basic Television and Video Systems”, McGraw-Hill, 1999.
[3] John G. Proakis and Dimitris G.Manolakis, “Digital Signal Processing, Algorithms and Applications”, PHI of India Ltd., New Delhi, 3rd
Edition, 2000
[4] Sanjit K.Mitra “Digital Signal Processing: A Computer Based Approach” Tata McGraw-Hill, New Delhi, 1998.
[5] Kenneth J.Ayala., “The 8051 Microcontroller Architecture Programming and Applications”, Penram International Publishing (India), 1996.
[8] The Television as Robot Servant Robert Thibadeau, Ph.D.August, 1993 CMU-RI-TR-93-22The Robotics Institute Carnegie Mellon
University Pittsburgh, Pennsylvania 15213 Copyright (c) 1993 This research was sponsored in part by the Advanced Research Projects
Agency (DOD) ARPA order 6873, under contract #MDA972-92--1010.

Samples the incoming audio signal and
stores in temporary memory

Ask the user to select his directory
Ask the user to select the directory
Stores the audio sample in his directory as
his favorite clipping

Compares his favorite clipping’s with the
incoming audio signal

Once the incoming signal is matched a control
signal is sent to control unit.

Control unit tunes the television to channel
displaying the clipping automatically

ISSN: 0975-5462