International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 10 - Oct 2013
High Speed Architecture Design Of Viterbi Decoder
Using Verilog HDL
Karri Megha Shyam1, B.N. Srinivasa Rao2, R.Prasada Rao3
1
1,2,3
Final M.Tech Student,2Assistant Professor, 3 Assistant Professor
Dept of Electronics and Communication Engineering, Avanthi Institute of Engineering & Technology,
Narsipatnam,Visakhapatnam, Andhra Pradesh.
Abstract:- The main purpose of this study is to yield the gains
obtained by the developers with the usage of Viterbi algorithm.
Research mainly centers on the grandness of Viterbi algorithm
in the practical applications with the VHDL code. Research not
only helps the students related to the communications but it also
helps the people who are in the field of decoders as it is one of
the efficient method for reducing the errors while
communication procedure is in advance. Here, VHDL code is
used in order to implement the Viterbi algorithm in a proper
way. Apart from various codes, researcher selected VHDL code
for this research as it offers the high capability in designing the
electronic systems. Students and the business people and one can
easily understand and analyze the Viterbi algorithm concepts
and can gain more knowledge on the VHDL code and the tools
that are used in this research.
to-digital conversion. Most of the researchers uses bit error
rate to find the bit errors and it depends on the length of the
message.
Viterbi decoder implements a soft-decision
algorithm to demodulate digital data from an analog signal
corrupted by noise.
Forward Error Correction codes are divided into two
types.


I. Introduction
Error correcting code is a method to find the errors
sequentially and correct it based on the remaining constraints
such as bit length etc. This study is so called as coding
theory. Error detection is the simple method that is checking
of digits, applications are credit card number mistakes. It is
also called as the block codes. These are mainly used in the
CD players and mobile phones. The sum of the digits given
the module of some number.
In the coding theory forward error correction is a
method used for controlling errors in the data exchange in
noisy channels. The main theme is the sender encode his
message using error correcting codes. The receiver receives
the encoded message and find the errors in the message
avoiding the retransmission. This gives the ability to avoid
resending the data in the reverse channel in fixed band width.
This technique is applied where the situations are
about to retransmission is more cost effective and one way
communication. The information which is sending is stored in
the storage devices to reveal the corrupted data. It uses digital
bit stream with the modular carrier. It is used in multicast
communications also.
FEC processing in a receiver may be applied to a
digital bit stream or in the demodulation of a digitally
modulated carrier. It is an integral part of the initial analog-
ISSN: 2231-5381
Block codes work on fixed-size blocks (packets) of bits
or symbols of predetermined size. Practical block codes
can decoded in polynomial time to their block length.
Convolutional codes work on bit or symbol streams of
arbitrary length. They are decoded with the Viterbi
algorithm and other algorithms are used. Viterbi
decoding allows asymptotically optimal decoding
efficiency with increasing constraint length of the
convolutional
code
and
at
the
expense
of exponentially increasing complexity. It can be turned
into a block code, if desired, by "tail-biting".
II. RELATED WORK
There are more number of block coding techniques in that the
most of them is related to hamming distance. Hamming ECC
is commonly used to correct NAND flash memory errors. It
gives single-bit error correction and 2-bit error detection.
These are only compatible for more single level cell (SLC)
NAND. Denser multi level cell (MLC) NAND requires multibit correcting ECC such as BCH. NOR Flash typically does
not use any error correction.
The traditional block codes are usually implemented
the decision algorithms every input there is output signal.
That means wherever needs decision is 1 or 0 bit. Softdecision algorithms like the Viterbi decoder process analog
signals and allows for much higher error-correction
performance than hard-decision decoding. The traditional
block codes apply the algebraic properties of finite fields.
A convolutional code is a wing of error-correcting code in
which consists of

each m-bit information symbol converted into n bit
symbol, where m/n is the code rate (n ≥ m) and
http://www.ijettjournal.org
Page 4650
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 10 - Oct 2013

Converting is a function of the last k information
symbols, where k is the constraint length of the code.
Convolutional codes are used in most of the applications and
it mainly used in the video communication. These are mainly
used in the hard decision codes. It encodes the data and
stored in 1 bit registers which is indexed and starts with 0.
The encoder has nmodulo-2 adders where the logic is:
0+0 = 0, 0+1 = 1, 1+0 = 1, 1+1 = 0), and n generator
polynomials one for each adder. An input bit m1 is stored in
leftmost register. Generator polynomials are used and the
encoder outputs n bits. Bit shift all register values to the right
represented as m1 moves to registers represented as
m0, m0 moves to m-1 and wait for the next input bit. If there
are no remaining input bits and the encoder continues output
until all registers have returned to the zero state.
Several algorithms exist for decoding convolutional
codes. For small values of k the Viterbi algorithm is
universally
used
as
it
provides maximum
likelihood performance and is highly parallelizable. We use
Viterbi decoders are thus easy to implement
in VLSI hardware
and
in
software
on
CPUs
with SIMD instruction sets.
Details of Errors occurred during the coding in
communication process:
Transferring of data from one person to another
person is called communication which requires more
programming for the mechanism and there is chance to get
errors in the signal. If there are simple errors there are
adjusted using the sequence.
But major problems can be solved randomly
utilizing few of the important features of Viterbi algorithm
which yields to the original sequence. The most important
feature for communication is to facilitate error free data
transmission among digital or analog functioning signals
along with amplification. Coding in communication system is
divided into four sections as
1) Encryption: mainly used for security of data,
2) Data compression: used for data streaming and to reduce
the space,
3) Data translation: used to demonstrate the data for
transmission of communication channels
4) Error Control: identifies the errors and correct them as
soon as possible.
Digital signals are represented as 0 and 1 and
detecting errors and the noise occurred during the
transmission is used for the correct those errors. Wired
communication can leads to the errors due to the random
motion and transmit through the conductors such as registers.
In wireless communication there are some resources for noise
such as mobiles gives disturbance to signal. The original
signal is normally added with the noise signal at receiver
input. The forward error correction auto repeat request
(ARQ), hybrid ARQ and error code correction (channel
coding) are the general methods used for error correction [9].
ISSN: 2231-5381
III. PROPOSED WORK
In our work we introduced a solution for block codes
and provide continues data stream. For we choose Viterbi
algorithm. In Viterbi algorithm we provide continues data
stream for bit streams. In Viterbi algorithm, we need to find
the shortest path for the set of observations. For finding the
shortest path we use trellis representation in this algorithm.
Trellis graph is used to represent the observation path. It uses
forward error correcting coded based on the convolutional
codes.
Viterbi algorithm uses syndrome trellis coding and
its is explained as follows.
The trellis will be assumed to have a periodic
structure and the meaning that the Viterbi decoding algorithm
operations will be the same for every state transition interval.
To construct such a trellis, we can use a memory-ν binary
shift register whose contents at any given time define the
state of the trellis. The states are represented as 00, 01, 10
and 11.
State
State
00
00
01
01
10
10
11
11
State transition when input is 0
State transition when input is 1
The Viterbi decoding algorithm is a decoding
process for convolutional codes for a memory-less channel. It
defines the normal flow of the noisy channel and error
recovery adds the duplicate information to the original
information I and output t is transformed through a
communication channel. Input at receiver is the information
with redundancy. The receiver extract the original
information through a decoding algorithm and generates an
estimate (e). A decoding algorithm increases the probability
p(r|e) is a maximum likelihood (ML) algorithm. An algorithm
which maximizes the p(e|r) through correct selection of the
estimate (e) is called a maximum a posteriori (MAP)
algorithm. Two methods have same results when the source
information i has a uniform distribution.
http://www.ijettjournal.org
Page 4651
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 10 - Oct 2013
i
Convolution
al Encoder
t
Channel
Convolution
al Decoder
r
e
The main tasks in the Viterbi decoding process are as
follows:
1. Quantization: Conversion of the analog inputs into digital.
2. Synchronization: Detection of the boundaries of frames
and code symbols.
3. Branch metric computation.
4. State metric update: Update the state metric using the new
branch metric.
5. Survivor path recording: Tag the surviving path at each
node.
6. Output decision generation: Generation of the decoded
output sequence based on the survivor path information.
The below block diagram shows the Viterbi
decoding algorithm performs the above tasks in the specified
order. It has different parts have different works. In this
process the signals are quantized and converted into digital
signals. The block detects the frame boundaries of code
words and symbol boundaries. We consider that the Viterbi
decoder receives successive code symbols and in which the
boundaries of the symbols and the frames have been
identified.
Received Signal
Receiver front
end
Signal timing recovery
Analog Signal
Quantization
Synchronization
Branch metric
computation
1. Branch Metric Unit (BMU): It calculates the shortest
length between the input pairs and ideal pairs.
2. Path Metrics Calculation: For every state it calculates for
survivor ending in a metrics needed.
3. Trace back: To get the expected results this is responsible
to be simulated at the hardware implementations that don‟t
store the actual information regarding the survivor paths. But
are stores one bit of information every time.
The branch metric computation block compares the
received code symbol with the expected code symbol and
counts the number of differing bits. The state metric update
block selects the survivor path and updates the state metric.
Each wing is usually implemented by a module
called Add-Compare-Select module. Both adders compute the
partial path metric of each branch and the comparator
compares the two partial metrics and the selector selects an
appropriate branch. Novel partial path metric updates the
state metric of state p and the respective survivor pathrecording block records the survivor path.
The number of necessary ACS module is equal to
half the number of total states. Time sharing of some ACS
modules is possible to save the hardware and such sharing
slows down the operation and dissipates more power. In This
gist we reckon replication of necessary ACS modules and
which is more power efficient.
IV.EXPERIMENTAL RESULTS
The input values for Viterbi encoder are specified in
the developed code. Providing input every time for the
developed design is time taking process. So, the input values
are included directly in the developed code, so that simulation
process can be executed directly. The Viterbi encoder input
values given in this project are:
0110100110010110100101100110100110010110011010010
1101001100101101001011001101001011010011001011001
1010
Each input value will be processed and a
corresponding output will be provided for the Viterbi
encoder. The output values will be specified in the form of
wave forms in both Xilinx and active HDL simulation
environments
Timing and Control
State Metric Update
Output decision
generation
Decoded Information bits
The Viterbi decoder architecture mainly consists of three
units.
ISSN: 2231-5381
http://www.ijettjournal.org
Page 4652
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 10 - Oct 2013
ANALYSIS OF SIMULATION USING XILINX
When clock signal is applied to the Viterbi decoder
reset button is set as 0,and the system reset button kept as
0,the cyclic encoded data starts after 100ns.by applying the
valid encoded pulse train to the decoder enc_symbol 0 and
enc_symbol 1 varies according to the periodic pulse train.
According to the enc_symbol the encoded output bit also a
pulse train.
The dec_symbol0 consists an 3‟h0 error bits at
decoder process up to 100ns,after 100ns it decodes 3bits at a
time changes continuously every 50ns.dec_valid_in is a
continuous pulse train when dec_symbol is change their bit
stream every after 100ns.thus we get dec_ out bit and
dec_valid _out bits same as 1.the count of decoder is 103 and
the buf_in_cnt as 102.from that the decoder buf_out_cnt and
total count can be represented as 0 bit count
The concerned output for the above encoder is
0312112110223322120221312311001110223322231100112
13112021022332212022131231100
112131120210223322231100
And for decoded symbol1 when the value is given as 7 then
the output is the demux value given as 0 7 0 7 0 7 0 7
When decoded valid input value is given as „0‟ then the
output is continues signal.
For a pattern value when given as „3‟ then there will
be no changes in the signal and similar appears when the
decoded bit output, decoded valid output and decoded output
error value is set to „0‟.
When glb_seed value is set to 000E4048 then there appear a
demux in the result. For ccnt, value is set to 1 then the result
104101010101010101010101010101010
And finally for buffer output count, total count and
sim done value when set to „0‟ the n there will be no changes
in the signal.
Implementation of Viterbi encoder can also be
simulated with the help of active HDL simulation
environments. Active-HDL is an Aldec product developed
using Field Programmable Gate Array (FPGA) with HDL
simulator. The main reason for choosing this tool for
simulation is that it can support graphical and text based
designs with various simulation languages.
Synthesis and Implementation
The design can be synthesized and implemented
after verifying the design behavior with simulations. The
process of synthesis and implementation can provide a clear
view on how the developed code is working for obtaining
simulation results
Analysis of simulation
For clock signal when set to value „1‟ then it results in a
continuous signal. When the signal is set to reset value „0‟
then it is a dc signal and there will be no changes in the
signal. When srst that is system reset value is set to „0‟. For
an encoded bit input value is set to „0‟ then the results can be
obtained as
100110100111000110111000
When an encoded valid input value is set to „0‟ then the result
can be obtained as
101010101010101010101010
When an encoded symbo0 value is given as „0‟ then the
results are
110111000000001111111111
For an encoded symbol 1 and if the value is set to 1 then the
analyzed results are
111011110001111000111111
Now for a decoded symbol 0 and when the value is set to 7
then there will be periodic changes obtained in a de-mux as 7
0707
ISSN: 2231-5381
The entire code developed for this project can be
visualized using Xilinx hardware device. How the Viterbi
algorithm code for encoder and decoder is working in a step
by step process can be analyzed based on results of
implementation.
http://www.ijettjournal.org
Page 4653
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 10 - Oct 2013
Checking the details of the design after
implementing each and every action is necessary. Synthesis
report can provide an opportunity to view all the resources
utilized by the developed design
Implementing the design in FPGA editor can
provide the below results. This view is providing in-built
working process and routing information.
A user can verify the design before using it in a
device; floor planner view can provide the opportunity to
identify the Viterbi decoder position. Positions of all
components used while developing the design can be
obtained by just a click.
All the trellis occurred are arbitrarily solved even in
presence of two or more simple errors in input string using
Viterbi algorithm and at same time the in presence of more
errors the decryption will be low even then the algorithm
ISSN: 2231-5381
works effectively this is the main advantage in
implementation of Viterbi algorithm. The Use of this
algorithm is found to be advantageous due to its cost
effectiveness in modulated minimize at the same time the
functional performance in some situation would modulate in
maintaining the original cost. Linear functioning of linear
pulse distance is due to convenient source sequence.
Analysis:
The implementation of Viterbi algorithm is
somewhat difficult even though the process of algorithm is
simple. Conventional encoding can be easily implemented on
Viterbi algorithm even though there exist a large gap in
complexity with the transmission side. State trellis uses
conventional encoding, the decoder explores rotates between
states because it is a finite state machine. It requires large
memory registers for storing results. There is some delay in
final decision on a sequence of transitions because of the size
of the input code is very high. By observing the transition
metrics between states the decision can be done and the
results are updated in the form of Hamming distance or
Euclidean with the error-corrupted received sequence.
The computation process of conventional codes
depends on the minimum distance and then on the constraint
size and coding speed. There are continuous changes in
implementation process in order to increase the parameters
like Gain. Complexity raise up to maximum limit in order
change the old techniques. Now a day‟s some algorithmic
part and Systolic architectures in implementing on various
devices Adaptive Viterbi decoding (AVD). Only subsets of
the states are used in AVD algorithm for storing and
implementation process. Due to decrease in size of states, the
performance also reduces. In order to increase the
performance of decoder, it is implemented at shorter
distances. When channel status information is available,
Replace stronger codes with simpler codes for high speed
switching between different coding rates.
For better use of power and coding rates in circuits‟
adaptive channel coding systems are used. The adaptive
decoder algorithm must contain programmability or reprogrammability at hardware logic level or algorithmic level
in order to increase the computation levels. Field
programmable gate array device is mentioned here for
achieving both parameters power and coding rate.
The fundamental concept used in Software radio
terminal is re-programmability which is present either at
hardware level or algorithmic level. An interface with
different communication standards are used in Software
defined radio terminals and it must consist of high decoding
capability for reprogrammable devices. Viterbi decoder use
convolution codes for channel decoding in its programmable
architecture and is similar to a special class of standards and
specifications.
Viterbi decoder program is used for switching
between UMTS and GPRS .whenever the implementation
uses only decoder, reconfiguration is not a major problem.
FPGA holds the total transceiver digital processing subsystem in real time applications. In order to decrease the
http://www.ijettjournal.org
Page 4654
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 10 - Oct 2013
problems in design process re use the same building blocks.
At
transmission
time
for
particular
application
programmability is used where data coding rate is
continuously changed at run time.
The Viterbi decoding process conventional codes of
fixed size of 5 and 9 and the code rate ½ and 1/3.by using
FPGA there is possibility of implementing extra features in
the Viterbi decoder (e.g. different standards and new
functions).
IV.CONCLUSION
As Viterbi algorithm is conceived more interesting and
challenging for this research topic, it is considered, and also it
has wide variety of applications in digital communications
field. This research helps to generate more profits by the
developers using Viterbi algorithm. This research mainly
concerned with implementation of Viterbi algorithm using
VHDL coding. Viterbi algorithm has many advantages like
low power consumption and main advantage is error
correcting using VHDL. Anyone reading this document will
have to gain the cognition of working with different tools like
Xilinx ISE and MODELSIM.
References:
[1] R. E. Stake (1995) “The art of case study research”.
USA: Sage.
[2] K.Aleksandar, M.F. Jose (2000) “The Viterbi Algorithm
and Markov Noise Memory”. [Internet] available at URL:
<http://www.ece.cmu.edu/~moura/papers/kavcic_viterbinoise
memory.pdf>, [accessed on20th November 2010].
[3] K. Hueske, J. Geldmacher, J. Gotze (2007) Adaptive
decoding of convolutional codes, Copernicus Publications,
pp. 209-214. Internet. (N.D) Coding in communication
system.[Internet]
available
at
URL:
<http://www.mathdb.org/notes_download/elementary/number
/ne_N6.pdf>, [accessed on20th November 2010].08)e:
[7] B. Tristan (2006) IMPLEMENTATION OF THE
VITERBI
ALGORITHM
USING
FUNCTIONAL
PROGRAMMING LANGUAGES. [Internet] available at
URL:
http://people.eecs.ku.edu/~esp/publications/c2009ItcTristan.p
df>, [accessed on 20November 2010].
[8] K. Hucker (2001) Research Methods in Health, Care and
Early Years. UK: Heinemann Publishers.
[9] Y. K. P.Cheung, G. A. Constantinides, J.T. D. Sousa
(2003) Field-programmable logic and applications: 13th
International Conference, FPL 2003, Lisbon, Portugal. New
York: Springer.
[10] J. Kia (2005) What Is a Convolutional Encoder /
Convolutional Encoding. [Internet] available at URL:
<http://www.askkia.com/articles/what-is-a-convolutionalencoderviterbi-algorithm.html>, [accessed on 20 th
November 2010].
17. G. Fettweis and H. Meyr, “Feedforward architecture for
parallel viterbi decoding,” J. VLSI Signal Processing, vol. 3,
pp. 105-119, 1991.
ISSN: 2231-5381
18. G. Fettweis, H. Meyr, “High rate Viterbi processor: a
systolic array solution,” IEEE J. SAC, Oct. 1990.
19. G. Fettweis, H. Dawid, and H. Meyr, “Minimized method
for Viterbi decoding: 600 Mb/s per chip,” in Proc.
GLOBECOM 90, vol. 3, pp. 1712-1716,Dec. 1990.
20. G. Fettweis, L. Thiele, H. Meyr, “Algorithm
transformations for ulimited parallelism,” Proc.
IEEE ISCAS’90, 1756-59; also L Thiele, G.Fettweis, “-,”
Electronics & Commun. no.2, 83- 91, 4/1990.
[22] A. P. Volnei (2004) Circuit design with VHDL. United
States of America: MIT press.
[23] P. P. Chu (2006) RTL hardware design using VHDL:
coding for efficiency, portability, and scalability. USA: John
Wiley and Sons.
[24] Project seminar (2010) VHDL (VHSIC hardware
description language) [Internet] available at URL:
http://www.seminarprojects.com/Thread-vhdl-vhsichardwaredescription- language--14251#top>, [accessed on
20th November 2010]
29. S. Kato, M. Morikura, S. Kubota, H. Kazama and K.
Enomoto, “A TDMA satellite communication system for
ISDN services,” Proc. Globecom‟91, pp. 1533-1540,
December 1991.
30. K. Kawazoe, S. Honda, S. Kubota and S. Kato, “Ultrahigh-speed and universal coding-rate Viterbi decoder
VLSIC,” Proc. ICC‟93, pp. 1434-1438, May 1993.
31. K. Koora, et. al, “ From algorithms testing to ASIC input
code of SOVA algorithm for Turbo codes,” Proceedings of
Turbo-code seminar, 1996.
32. S. Kubota, K. Ohtani and S. Kato, “A high-speed and
high-coding-gain Viterbi decoder with low power
consumption employing SST (scarce state transition)
scheme,” Electron. Lett., 22 (9), pp. 491-493,1986.
33. L. Lang; C.Y. Tsui and R.S. Cheng, “Low power soft
output Viterbi decoder scheme for turbo code decoding,”
Conference-Paper, ISCAS ‘97(Cat. No97CH35987). IEEE,
New York, NY, USA, 4 vol. Lxvi+2832 pp. p. 1369-1372
vol.2, 1997.
34. H.F. Lin and D.G. Messerschmitt, “Algorihtms and
architectures for concurrent Viterbi decoding,” in Proc. ICC,
pp.836-840, 1989.
35. B.K. Min and N. Demassieux, “A versatile architecture
for VLSI implementation of the Viterbi algorithm,” in Proc.
ICASSP, pp. 1101-1104, May 1991.
36. D. Oh and S. Hwang, “Design of a Viterbi decoder with
low power using minimum transition traceback scheme,”
Electronic-Letters, IEE, Vol.32, No.22, pp. 2198-2199, Oct.
1996.
BIOGRAPHIES:
http://www.ijettjournal.org
Karri Megha Shyam completed B.Tech
Degree
in
Electronics
and
Communication Engineering from AlAmeer College of Engineering &
Information
Technology,
Visakhapatnam. currently pursuing
M.Tech in from Avanthi Institute of
Page 4655
International Journal of Engineering Trends and Technology (IJETT) – Volume 4 Issue 10 - Oct 2013
Engineering & Technology,Narsipatnam, Visakhapatnam,
Andhra Pradesh. Interesting research areas VLSI custom
design.
B.N. Srinivasa Rao received his B.Tech
degree
in
Electronics
and
Communication Engineering from JNT
University, Hyderabad, India
and
M.Tech in VLSI System Design from
JNT University, Hyderabad, India. He is
currently working as an Assistant Professor in Avanthi
Institute of Engineering and Technology, Visakhapatnam,
Andhra Pradesh, India. He has 5 years teaching and 9 years
industrial experience. He has 10 publications in various
International conferences. His area of interest VLSI Semi
and full custom design. He guided many projects for B.Tech
and M.Tech students.
R. Prasada rao received his B.Tech degree
in
Electronics
and
Comunication
Engineering from Andhra University,
Visakhapatnam, India and M.Tech in VLSI
System
Design
from
JNT
University,Hyderabad,
India.
He is
currently working as an Assistant Professor
in Avanthi Institute of Engineering and Technology,
Visakhapatnam, Andhra Pradesh, India. He has 8 years
teaching experience. He has 2 publications in various
International conferences. His area of interest is Digital
Electronics, Signals and Systems and Signal Processing. He
guided many projects for B.Tech and M.Tech students.
ISSN: 2231-5381
http://www.ijettjournal.org
Page 4656