Investigating the optimal SCL decoding algorithm for
polar codes in ultraviolet communication
Lin Ai, Min Zhang, Dahai Han, Wenxiu Hu, Xuetong
Liang
State Key Lab of Info. Photon & Opt. Comm
Bejing University of Posts and Telecommunications
Beijing, China
Emile: mzhang@bupt.edu.cn
Zabih Ghassemlooy
Optical Research Group, NCRLab
Northumbria University
Newcastle England
Emile: z.ghassemlooy@northumbria.ac.uk
Abstract—The polar codes (PCs) has been shown to effectively
enhance the transmission distance of ultraviolet communication
(UVC) systems, however, an optimized decoding algorithm for
PCs is required to reduce the interference and the path-loss
propagation attenuation. In this paper, a successive cancellation
list (SCL) decoder scheme is proposed and implemented over the
classic successive cancellation (SC) decoding algorithm by
extending decoding path using the UV path loss particle model.
Numerical simulations are performed to show the performance
improvement of SCL decoding scheme for PCs in UVC that
combined with cyclic redundancy check (CRC) joint detection
decoding.
Keywords—Polar codes; SC decoding algorithm; SCL decoder;
CRC joint detection decoding; UV path loss particle model.
I.
I
NTRODUCTION
This paper is based on the recent progress reported on
channel coding schemes for ultraviolet communications (UVC)
within the solar-blind spectrum of 200-280 nm. The concept
relies upon two basic related phenomena. (i) Solar radiation
within the 200-280 nm band reaching the ground is negligible
due to strong absorption of atmospheric gases (i.e., ozone and
oxygen). (ii) UV generated at the earth’s surface is strongly
scattered, rendering a non-line-of-sight (NLOS) transmission
paradigm, which offers more flexible deployment [1], [2]. Such
properties have attracted wide interest in UVC. However, due
to the high propagation path loss
deriving from scattering,
refraction and absorption, the intensity of UV beam may
degrade severely, thus ultimately leading to degradation of the
link performance.
To advance UVC with improved performance, a number of
options are available including coding schemes such as forward
error correction (FEC) coding, which is widely used in wireless
and optical communications. In [3] a 480
×
960 low-density
parity-check (LDPC) code scheme was adopted in UVC, where
the average transmission distance was increased by 78% at
target bit error rate (BER) of
3
10
−
. In [4], PC as the channel
code was used in UVC, the effective transmission distance was
increased by 15% compared with LDPC scheme. PCs are the
first codes known to have an explicit construction, while also
being demonstrated to achieve Shannon's capacity over the
binary-input symmetric memoryless channels [5]. However,
PCs performance over short to moderate block lengths is not
very satisfactory. This is due to (i
) the channel polarization
effect, which cannot be fully performed when having a limited
code length; and (
ii
) the properties of classic successive
cancellation (SC) decoder employed to decode PCs. Therefore,
the performance of UVC system can be further improved by
means of optimizing the decoding algorithm (DA) of PCs.
In this paper, an optimization of DA for PCs is
implemented over the classic SC based DA through extending
the decoding path, which is named as the successive
cancellation list (SCL) decoder. To our best knowledge, there
is no related researches introducing such polar decoding
scheme into UV communication system. The reason for
adopting algorithm optimization instead of using other DAs is
that the selection of polarization channel is actually based on
SC decoding performance [5]. So for PCs, the most suitable
DA should be based on SC decoding [6], which takes full
advantage of its code structure. Numerical simulations carried
out validate the implementation of SCL decoding scheme in
UVC. We verify optimization of the link performance by
adopting SCL and cyclic redundancy check (CRC) joint
detection decoding schemes.
The rest of the paper is organized as follows. In section II
the functional block diagram of the single input single output
(SISO)-based UVC system and the channel model are
proposed. In section III encoding and decoding algorithms for
PCs are introduced. In decoding part, the difference between
SCL and SC in the path search process are described in detail,