Performance Review of Serially Concatenated RSC-RSC Codes using

Non-iterative Viterbi Decoding Technique

Navdeep Singh
1
, Gagandeep Kaur
2


1
M.tech, Electronics and Communication Engg., Punjabi University Patiala Yadavindra College of
Engineering, Talwandi Sabo, Punjab, India, [email protected]

2
Department of Electronics and Communication Engg., Punjabi University Patiala Yadavindra College of
Engineering, Talwandi Sabo, Punjab, India, [email protected]

Abstract In this paper, review about the performance of RSC-RSC concatenated system in terms of bit
error rate (BER) is done using MATLAB software. The RSC-RSC system uses Concatenation of
Recursive systematic convolution codes (RSC) as outer and inner codes with the help of interleaver. For
decoding of RSC-RSC system, we have used a non-iterative decoding scheme which includes two Viterbi
decoders in concatenation with de-interleaver. We show that in the RSC-RSC system BER decreases, as
the signal to noise ration increases. We have also investigated the performance of RSC-RSC system for
effect of varying data rate using puncturing.
Keywords Turbo codes; serially concatenated convolutional codes; Recursive systematic
convolutional codes; Viterbi decoding; Channel coding.


I. INTRODUCTION

In Digital communication system, channel encoding improves the bits in error rate by added
redundancy (additional bits). The primary objective of error-correcting coding is to improving the
reliability of communication system. Channel coding is often used in digital communication systems to
protect the digital information from noise and interference and reduce the number of bit errors. The
history of channel coding, also referred to as forward error correction, In Shannon 1948 article, "A
mathematical theory of communication",[1] According to Shannon, it is possible to design codes with
any desired small probability of error, whenever the transmission rate is smaller than the capacity of the
channel. Unfortunately, Shannon provided no insights on how to actually design these codes.
Until the late 1940s, communication devices were equipped with error detection capabilities only.
In 1950 Hamming was the rst to propose a single-error correcting code [2], while Golay developed a
more ecient scheme able to correct up to three erroneous bits [3]. Both Hamming and Golay codes
group blocks of information bits together with parity check bits, the latter being computed using a
mathematical combination of the information bits. These types of codes are known as block codes.
Popular variations of block codes are the Reed-Muller and the cyclic redundancy codes. Sub- classes of
cyclic redundancy codes, such as the Bose-Chaudhuri-Hocquenghem and Reed-Solomon codes, are still
used in a wide variety of applications [4].
In 1955, Elias introduced the concept of convolutional coding [5]. The convolutional encoder
makes use of shift registers to generate output bits based on the present input bit as well as past inputs.
Convolutional codes are one of the powerful and widely used class of codes, These codes are having
many applications, that are used in deep-space communications, voice band modems, wireless
standards(such as 802.11) and in satellite communications. The main advantage of convolutional codes
over block codes is better error rate performance, owing to the optimal exploitation of soft channel
observations by the decoding algorithm. In particular, Viterbi proposed a maximum likelihood sequence
estimation algorithm [6] in 1967, while a more ecient but more complex algorithm, based on maximum
a-posteriori decoding, was developed by Bahl et al. [7] in 1974. Convolutional coding was initially
introduced in standards for satellite communication applications and deep space missions, but was later
also adopted in mobile communication systems.
Concatenation of codes was the next signicant step that enabled better performance of codes on
communication channels that introduced burst errors. More specically, Forney showed [8] in 1966, that
a concatenated coding system with a powerful outer code can perform reasonably well, when its inner
decoder is operated in the high error probability region. In principle, block encoders can be combined
with convolutional encoders and interleavers, in parallel or serial schemes. In 1989, Concatenation of
multiple convolutional codes [9] was introduced, in which Soft Output Viterbi Algorithm (SOVA) are
used. However, in 1993, interest shifted back to channel coding, when Berrou, et.al presented turbo codes
[10]. Particularly, a turbo encoder is the parallel concatenation of two convolutional codes separated by
an interleaver. However, the name turbo comes from the similarity in logic between the turbo engine
and the iterative decoding process between the two component soft-input soft-output (SiSo) convolutional
decoders at the receiver. It shown the performance of Turbo code, in terms of Bit Error Rate (BER), is
very close to Shannons limit.
Shortly after the introduction of the turbo codes, it turned out that serially concatenated RSC codes
with a random bit interleaver were equally suitable for iterative decoding [11]. In [12], the authors claim
that these codes may have even better performance than the turbo codes. Fig. 1(a) and (b) describe the
encoder/decoder of SCCC.

Information SCCC
bits Coded bits

(a)

Demodulator
O/P


Recovered Information bits

(b)
Figure 1. Structure of SCCC encoder and decoder. (a) SCCC encoder structure (b) SCCC decoder structure.

The iterative decoding yield a remarkable increase in performance than a single iteration decoding and in
some cases approximates the theoretical limit. Through iterative decoding scheme, performance in terms
of BER is enhanced, but at the expense of complexity of the system.
However, the convolutional codes suffered from the problem of burst errors [13] and Reed
Solomon codes suffered from problem of random errors [14]. To compensate this problem, a new
concatenated scheme was proposed in which a concatenation of a Reed-Solomon (RS) code and a
Recursive systematic convolutional code (RSC) codes was used & it was shown that RS-RSC
concatenated codes have good performance than RSC itself [15]. For SCCC codes, simple concatenated
Viterbi decoding was proposed with certain drawbacks [16].
In recently research a solution was provided that, RSC-RSC serially concatenated convolutional
code with non-iterative concatenated Viterbi decoding was implemented and it shown that RSC-RSC
Outer
Encoder
Interleaver Inner
Encoder

Inner
Decoder
De-
Interleaver
Outer
Decoder
Decision
Device
Interleaver
system has better BER performance than RS-RSC concatenated code [17].
In this paper, we compare the BER performance of serially concatenated RSC-RSC code using
non-iterative concatenated Viterbi decoding [17]. In Modern communication system, a low bit error rate,
signal-to-noise ratio is very important parameter. The rest of the paper is organized as follows. In section
II, system structure of concatenation scheme is presented. The simulation results and its discussion are
given in section III. Finally, the section IV concludes the paper rate, signal-to-noise ratio are very
important parameter. The rest of the paper is organized as follows. In section II, system structure of
concatenation scheme is presented. The simulation results and its discussion are given in section III.
Finally, the section IV concludes the paper.


II. SYSTEM STRUCTURE

In this section, structure of the simulated system with simulation parameter is described.

The basic scheme of two serially concatenated block codes [11] is shown in Fig. 2.

(N,k) (n,N)




C
S
(n,k,N)
Figure 2. Serially concatenated (n,k,N) block code.

It is composed of two cascaded CCs, the outer (N,k) code C
o
with rate R
o
c
=k/N and the inner (n,N)
code C
i
with rate R
i
c
=N/n,

linked by an interleaver of length N. The overall Serially concatenated block
code is then an (n,k) code, and we will refer to it

as the (n,k,N) code C
s
, including also the interleaver
length.
A. RSC-RSC system:
A serially concatenated RSC-RSC code is concatenated of two recursive systematic convolutional
encoder (i.e, outer encoder and inner encoder) through an interleaver in between them. Because
convolutional encoder improves only random error, so we have used an interleaver in serially
concatenation between outer encoder and inner encoder which provides solution to improve the brust
error correction.
B. Simulation Model:
The simulation model of RSC-RSC concatenated system is shown in Figure 3. In this setup, RSC-
RSC system is implemented with two RSC encoder of feed forward polynomial of [133 171] and
feedback polynomial of [133], with constraint length of 7. Their base code rate is 1/2 each and punctured
code rates are 1/2, 2/3, 3/4. For decoding of RSC-RSC, two Viterbi algorithm decoders in serially
concatenated with de-interleavers are used. The effect of signal-to-noise ratio on bit-error rate is
observed. Table 1 describes the simulation parameters used in simulation model.
Outer
Encoder
Interleaver
Length=N

Inner
Encoder


Figure 3. Simulation Model for RSC-RSC system [17]


Table 1. Simulation Parameters of RSC-RSC System [17]

Outer Encoder Inner Encoder



RSC ( 1,171/133) RSC ( 1,171/133)



Constraint length = 7 Constraint length= 7



Base code rate = 1/2 Base code rate = 1/2



Punctured code rate =
1/2, 2/3
Punctured code rate =
1/2, 2/3, 3/4




Viterbi Algorithm Viterbi Algorithm

(hard-decision)
Depth=140
(hard-decision)
Depth=140



Helical interleaver

BPSK modulation

AWGN channel


III. RESULTS & DISCUSSION
The serially concatenated RSC-RSC system is described in section II is implemented using
MATLAB (R2012a) and BER is observed for different values of E
b
/N
o
(signal-to-noise ratio).
A. RSC-RSC concatenated system:
After simulation using MATLAB (R2012a) software results observation have been obtained as
shown in Figure 4 and noted down in Table 2. It shows that for RSC 1/2RSC 1/2 system, BER value of
0 at E
b
/N
o
of 1 dB. It is quite better than other three code rate, where BER is quite high at same value of
E
b
/N
o
i.e. 1 dB.

Figure 4. BER probability analysis for RSC-RSC system


Table 2. BER v/s Signal-to-noise ratio for RSC-RSC system
Coding scheme
Signal to noise ratio (Eb/No)



0 dB 1 dB 2 dB 3 dB 4 dB





RSC 1/2- RSC 1/2 0 0
0 0 0

RSC 1/2 -RSC 2/3
10
-1.4

10
-2.9

0 0 0

RSC 1/2 -RSC 3/4
10
-0.5

10
-0.8

10
-2.3
0 0

RSC 2/3- RSC 3/4
10
-0.7

10
-0.8

10
-1.2

10
-4.3

0




IV. CONCLUSION

In this paper we have simulated and reviewed the BER performance of serially concatenated RSC-
RSC code using non-iterative Viterbi decoding. The purpose of the simulation is to analyse the
performance of serially concatenated RSC-RSC system. The simulation results shows clearly that the
performance of serially concatenated RSC-RSC system is decrease as increasing the overall code rate i.e.
low value BER occur at low overall code rate.
REFERENCES

[1] E. Shannon, A mathematical theory of communication, Bell System Technical Journal, pp. 379427, 1948.
[2] R. Hamming, Error detecting and error correcting codes, Bell System Technical Journal, pp. 147160, 1950.
[3] M. Golay, Notes on digital coding, Proc. IEEE, vol. 37, p. 657, 1949.
[4] P. Sweeney, Error Control Coding: An Introduction. New York: Prentice Hall, 1991.
[5] P. Ellias, Coding for noisy channels, IRE Nat. Conv. Record, vol. 3, no. 4, pp. 3746, 1955.
[6] J. Viterbi, Error bound for convolutional codes and asymptotically optimum decoding algorithm, IEEE
Trans. Inform. Theory, vol.IT-13, pp. 260269, Apr. 1967.
[7] L. R. Bahl, J. Cocke, F. Jelinek, and J. Raviv, Optimal decoding of linear codes for minimising symbol error
rate, IEEE Trans. Inform. Theory, vol. IT-20, pp. 284287, Mar. 1974.
[8] G. D. Forney, Concatenated Codes. Cambridge, Mass.: MIT Press, 1966.
[9] Hagenauer and P. Hoeher, Concatenated Viterbi Decoding Proceedings of the International Workshop on
Information Theory, 1989.
[10] Berrou, A. Glavieux, and P. Thitimajshima, Near Shannon limit error- correcting coding and decoding:
Turbo-codes, in Proc. IEEE International Conference on Communications (ICC93), Geneva, Switzerland,
May 1993, pp. 10641070.
[11] S. Benedetto and G. Montorsi, Serial concatenation of block and convolutional codes, Electron. Lett., Vol.
32, No. 13, May 1996, pp. 887-888.
[12] S. Benedetto, D. Divsalar, G. Montorsi, and F. Pollara, Serial concatenation of interleaved codes: Performance
analysis, design, and iterative decoding, IEEE Trans. Inf. Theory, Vol. IT- 44, May 1998, pp. 909-926.
[13] P. Elias, Coding for Noisy Channel, IRE Conv. Rec : 37-46, 1955.
[14] S. Reed and G. Solomon, Polynomial codes over certain finite fields, J. Soc. Industrial Appl. Math, vol. 8,
pp. 300-304, 1960
[15] Byun, Sunghan, Dong Joon, and Jung Sun, Performance comparison of RS-CC concatenated codes using NSC
and RSC codes, Network Infrastructure and Digital Content, 2nd IEEE International Conference, Beijing,24-
26 Sept. 2010
[16] Boutillon, C. Douillard, and G. Montorsi, Iterative Decoding concatenated codes Implementation issues,
draft, January 29, 2007.
[17] Manish, and S. Jyoti, Performance Comparison of RSC-RSC Concatenated code and RS-RSC Concatenated
code using Non- iterative Concatenated Viterbi Decoding Technique, Manuscript submitted for publication,
IEEE proceedings of ICMIRA, Dec 2013.