2018 Asia Communications and Photonics Conference (ACP)
A combinding PAPR reduction method for DCO-
                   OFDM UOWC
              Jurong Bai                                             Hang Dai                                        Yi Yang
   School of Electronic Engineering                      School of Electronic Engineering                School of Electronic Engineering
     Xi'an University of Posts and                         Xi'an University of Posts and                   Xi'an University of Posts and
         Telecommunications                                    Telecommunications                              Telecommunications
             Xi'an, China                                          Xi'an, China                                    Xi'an, China
         jurongbai@163.com                                     couragedh@163.com                               yangyi@xupt.edu.cn
    Abstract—Optical orthogonal frequency division multiplexing                Many techniques for reducing PAPR are listed in [5]. The
(OFDM) has been adopted in underwater optical wireless                     clipping method brings nonlinear distortion of OFDM signals
communication (UOWC) system, to achieve high data rate and                 that results in serious in-band noise and BER increase [6].
huge bandwidth, as well as to deal with the problem of inter               Many iteratives are required in the PTS method, which
symbol interference. However, it has the problem of high peak-
                                                                           increases its complexity [7]. Only part of the total subcarriers
to-average power ratio (PAPR) which will result in serious
performance deterioration and increase equipment costs. In this            in the TR method are used for data transmission, the rest
paper, an SLM method is combined with μ-law companding                     subcarriers are reserved to carry the peak reduction signals.
method to reduce the PAPR of UOWC system, which reduce the                 Since this part of the subcarriers does not carry useful
signal PAPR by 6.8dB and is superior than both the traditional             information, it brings low spectrum efficiency [8]. μ-law
SLM and μ-law companding methods. As for the system bit error              companding is a non-linear transformation, which amplifies
rate (BER), the proposed method has better performance than                small signals, and compresses large signals to reduce PAPR
the μ-law companding method. With reasonable tradeoff with the             [9]. In [10], selective mapping (SLM) is a non-distorting
BER, the proposed method has faster convergence than the                   method. By reducing the correlation between input sequences,
traditional SLM method.
                                                                           the modulation data have multiple forms of phases. Through
    Keywords—Orthogonal frequency division multiplexing,                   comprehensive analysis, the signal with the lowest PAPR was
underwater optical wireless communication, peak-to-average ratio,          selected for transmission.
selective mapping, companding                                                  In this paper, the SLM and μ-law companding schemes are
                                                                           combined to reduce PAPR of O-OFDM signal in UOWC.
                        I. INTRODUCTION                                    Compared with the traditional schemes, the PAPR of the O-
                                                                           OFDM signal is reduced with much faster convergence.
    Underwater wireless optical communication (UOWC)
research is a hot topic in the field of communication systems.                                    II. SYSTEM MODEL
UOWC has the advantages of low power consumption and
high security [1]. Optical orthogonal frequency division                       Blue-green light is less attenuated in seawater channels than
multiplexing (O-OFDM) is a multi-carrier modulation                        other optical bands, and has the advantages of strong
technology that can be applied in various water environments.              penetrating power, high speed, high efficiency, and strong anti-
                                                                           interference ability. Therefore, the blue-green laser
It converts a set of high-speed serial data into a low-speed
                                                                           transmission signal is most adopted in the seawater channel.
parallel data stream, so that each sub-carrier is orthogonal to
                                                                           The optical properties of absorption and scattering can cause
each other for multiplexing. O-OFDM technology is good at                  severe attenuation, as illustrated in Figure 1 [11]. The total
anti-multipath interference. Its main idea is to convert the               attenuation coefficient of seawater is the sum of seawater
frequency domain signal into the time domain signal through                absorption coefficient and scattering coefficient. Absorption is
IFFT at the transmitting end. Its orthogonality makes O-                   a process in which photons lose energy and are converted into
OFDM system have high spectrum utilization. However, due                   other forms, which causes the propagation energy of light
to the superposition of multi-carrier signals, high peaks are              continuously decrease, and limit the link distance of UOWC.
generated, resulting in excessive demands on the linear power              Scattering is caused by changes in photon propagation
amplifier and increasing the industrial cost [2]. To solve this            directions. In the UOWC system, the size of the aperture that
problem, many research have been take.                                     absorbs photons is constant, therefore, the number of photons
    The optical properties of seawater are mainly related to               received by the receiver is reduced due to the effect of
three factors: purity of seawater, dissolved substances, and               scattering. Moreover, the time at which the photons reach the
suspended particles [3]. Reference [4] studied the performance             receiving surface may be different, and multipath dispersion
of UOWC systems, describes the efficient and reliable                      may occur [12]. Impurities in seawater have an absorption
characteristics of optical channels and comprehensively                    effect on light, and laser signals form a beam spread when
considered the performance of the transmitter and receiver                 applied in a seawater channel. When transmitting at a close
parameters under different water qualities.
    Scientific Research Program Funded by Shaanxi Provincial Education
 Department (16JK1702).
XXX-X-XXXX-XXXX-X/XX/$XX.00 ©20XX IEEE
                                                      978-1-5386-6158-1/18/$31.00 ©2018 IEEE
                                                   2018 Asia Communications and Photonics Conference (ACP)
range through the seawater channel, the delay of channel                              A. SLM Scheme
propagation can be ignored [13].                                                          The main idea of the SLM method [5] is to copy OFDM
        As illustrated in Fig.1, Pi (λ ) is the incident light power.                 input signal X = [ X1 , X 2 ,, X N −1 ] into U groups with a length of
 Pa (λ ) is the part that is absorbed. Ps (λ ) is the part being                      N for transmission. The U-group data are multiplied by
                                                                                      different phase factors bu = [b1u , b2u ,, bNu −1 ] , in order to find a
scattered. The transmission formula of the optical channel in
the seawater channel is [14]:                                                         suitable phase factor. The phase factor should be as diverse as
                                                                                      possible. The smaller the correlation between the phase factors,
               Pr = Pt ⋅  ar2 /( d ⋅ tan θ + at ) 2  ⋅ e − c ( λ )⋅d (1)          the better the PAPR reduction performance for the OFDM
      ar is the radius of the receiving antenna. Pt is the power                      signal obtained by multiplying the multipath signal with the
                                                                                      phase factor. Then IFFT is performed to obtain {x1 ,x 2 ,,xU } 。
emitted by the light source. Pr is the power received. θ is the
                                                                                      The one with the smallest PAPR is selected for transmission.
divergence angle of the light source beam. d is the underwater                        The PAPR reduction performance differs with the number of
optical communication distance. c represents the total                                U.
attenuation of the absorption and the scattering coefficients.
                                                                                            The phase factor is expressed as:
And λ is the laser wavelength.
     Intensity modulation direct detection (IM/DD) is adopted                                                             bun = e jϕu ,n ,0 ≤ n ≤ N − 1
in the UOWC system, which requires the modulation signal to                                                                                                                                                                      (5)
be a positive and real number [15]. It is necessary to convert                                                             ϕ u ,n∈ [0,2π ],1 ≤ u ≤ U
the modulated bipolar complex signals into unipolar real
signals. The specific steps are: first go through the Hermitian                       B. Companding Scheme
Symmetry operation [8]. Next, the unipolar signal is obtained                             The traditional companding method is to amplify the small
by adding a DC bias to remove the negative part. Then, the                            signal with the large signal remains unchanged, which
transmission signal is ready for data transmission.                                   increases the average power of the signal, thus reduces the
                                                                                      PAPR. Here we choose an improved μ-law companding
                                                                                      transform method [9], which amplifies small signals and
                                                                                      compresses large signals.At the same time, the C-transform of
                                                                                      the transmitted signal is given by:
                                                                                                                           Axn                 μ
                                                                                                     Sn = C { xn } =                  ln(1 + xn )    (6)
                                                                                                                      ln(1 + μ ) xn            A
               Fig. 1. Geometry of inherent optical property.
                                                                                      where A is the average amplitude of the OFDM symbol, and it
                                                                                      is also a turning point of the C transform. μ is a constant,
                       III. PAPR REDUCTION                                            usually less than 5.
    OFDM signal is formed by operating IFFT to the QPSK                                   The receiving end performs a C inverse transform on the
signal. The expression of the OFDM symbol in the time                                 received signal according to:
domain is:                                                                                                               A′xn  [ n A′ ] 
                                                                                                                                   r ln(1+ μ )
                                                                                                      yn = C −1 {rn } =         e             − 1  (7)
                       1   N −1       j 2π kn                                                                           μ rn                    
              xn =
                       N
                            X ke
                           k =0
                                        LN
                                                , n = 0,1,, N − 1            (2)
                                                                                      where rn is the received signal after passing through the
   where X k is the QPSK symbol expressed as {1, -1, j, -j}. L                        channel, and A′ is the average amplitude of rn .
denotes the oversampling rate. N is the number of subcarriers.                        C. Proposed Scheme
    Due to the superposition of multiple orthogonal subcarriers,                         In this section, the system block diagram of SLM-μ
peaks will be generated. PAPR is defined as the peak power                            Companding, its theoretical analysis and improved design are
divided by the average power:                                                         depicted. Figure 2 shows the process of applying the proposed
                                                                                      SLM-companding to UOWC system.
                   PAPR(dB)=10log10
                                                   { }
                                                max xn
                                                             2
                                                                              (3)                          b1
                                                 E{ x }
                                                                                                                   IFFT
                                                         2                                                     2
                                                                                                           b
                                                                                                                                                                    Companding
                                                     n                                                             IFFT
                                                                                                                                           min PAPR
                                                                                                                                                                                         Symmetry
                                                                                                                                                                                         Hermitian
                                                                                                                                                                                                                   P/S DAC
                                                                                                 mapping
                                                                                                                                                                                                        +DC CP
                                                                                                                                            Seclect
                                                                                                  QPSK
                                                                                       Data in
                                                                                                                                                                                                                                 LD
    Set a peak-to-average ratio threshold PAPR0 , and use                                                  b   U
                                                                                                                   IFFT                                                                                                        UOWC
CCDF to indicate the probability that the PAPR of the OFDM                                                                                                                                                                     Channel
signal exceeds this threshold. The mathematical expression is:
                                                                                                                                                                            Companding
                                                                                                                                                      Data× phase
                                                                                                                              QPSK De-
                                                                                                                                                                                                                   Amplifier
                                                                                                                                                                                                         ADC S/P
                                                                                                                               mapping
                                                                                                                   Data out
                                                                                                                                                                                               -CP DC
                                                                                                                                                                              Inverse
                                                                                                                                                                                                                                Photo-
                                                                                                                                         FFT
                                                                                                                                                                                                                               detector
                  CCDF =Pr (PAPR ≥ PAPR0 )                                    (4)
                                                                                        Fig. 2. Schematic diagram of the UOWC system with SLM-companding.
                                                                 978-1-5386-6158-1/18/$31.00 ©2018 IEEE
                                                                                    2018 Asia Communications and Photonics Conference (ACP)
                                                          IV. SIMULATION                                             companding methods. With greater μ, the PAPR reduction
    In the simulation work, the number QPSK modulation is                                                            performance is better, when μ=3, the proposed method can
adopted and the number of subcarriers is 1024, of OFDM                                                               reduce signal PAPR by 6.8dB compared with the original
signal is 1000, μ=3. And blue-green laser is selected for                                                            signal.
transmission in the seawater channel.
                                                                                                                                                ACKNOWLEDGMENT
    Figure 3 shows the PAPR produced by the SLM-
companding scheme. It can be seen that the proposed algorithm                                                            The authors would like to thank every anonymous
can significantly reduce the PAPR compared with the                                                                  reviewer for his/her detailed comments. The authors would
traditional SLM and μ-Companding transform methods.                                                                  also like to thank the support by the Scientific Research
Compared with the SLM algorithm, the PAPR is reduced by                                                              Program Funded by Shaanxi Provincial Education
3.9 dB, which has 1.1 dB more reduction than the companding                                                          Department (16JK1702).
method. The PAPR is reduced by 6.8dB compared with the
original OFDM signal. In addition, it can be seen that as the μ                                                                                     REFERENCES
value increases, the PAPR reduction performance is better.
                                                                                                                     [1]    J. Armstrong, "OFDM for optical communications," in Special Issue on
                               10
                                    0                                                                                       OFC/NFOEC 2008 Conference, pp. 189-204, Institute of Electrical and
                                                                                              original
                                                                                              SLM
                                                                                                                            Electronics Engineers Inc. (2009).
                                                                                              SLM-C µ=1
                                                                                                                     [2]    Y. Rahmatallah and S. Mohan, "Peak-To-Average Power Ratio
       CCDF (Pr[PAPR>PAPR0])
                                                                                              Companding
                                    -1
                                                                                              SLM-C µ=2
                                                                                              SLM-C µ=3
                                                                                                                            Reduction in OFDM Systems: A Survey And Taxonomy," in IEEE
                               10
                                                                                                                            Communications Surveys & Tutorials, vol. 15, no. 4, pp. 1567-1592,
                                                                                                                            Fourth Quarter 2013.
                                                                                                                     [3]    R. C. Smith and K. S. Baker, "OPTICAL PROPERTIES OF THE
                               10
                                    -2
                                                                                                                            CLEAREST NATURAL WATERS (200-800 nm)," Applied Optics
                                                                                                                            20(2), 177-184 (1981)
                                                                                                                     [4]    M. A. Khalighi, C. Gabriel, T. Hamza, S. Bourennane, P. Léon and V.
                               10
                                    -3                                                                                      Rigaud, "Underwater wireless optical communication; recent advances
                                        3     4       5       6    7      8
                                                                  PAPR0 [dB]
                                                                                9        10        11      12
                                                                                                                            and remaining challenges," 2014 16th International Conference on
                                                                                                                            Transparent Optical Networks (ICTON), Graz, 2014, pp. 1-4.
  Fig. 3. Comparison of PAPR between different schemes and different μ                                               [5]    S. H. Han and J. H. Lee, "An overview of peak-to-average power ratio
                   values in UOWC DCO-OFDM system                                                                           reduction techniques for multicarrier transmission," IEEE Wireless
                                                                                                                            Communications 12(2), 56-65 (2005)
      Figure 4 shows BERs of different methods. It can be seen                                                       [6]    S. Takebuchi, G. Osada and F. Maehara, "A Deterministic Transmit
                                                                                                                            Power Control for OFDM Clipping and Filtering," 2013 IEEE 78th
that, the error rate decreases as μ decreases. Considering the                                                              Vehicular Technology Conference (VTC Fall), Las Vegas, NV, 2013, pp.
combination of bit error rate and PAPR comprehensively, it is                                                               1-5.
feasible to increase the bit error rate slightly under the best                                                      [7]    K.-S. Lee, H. Kang and J.-S. No, "New PTS Schemes With Adaptive
condition of PAPR reduction. Moreover, compared to the                                                                      Selection Methods of Dominant Time-Domain Samples in OFDM
companding method, the proposed algorithm has better BER                                                                    Systems," (2018)
performance.                                                                                                         [8]    J. Bai, Y. Li, Y. Yi, W. Cheng and H. Du, "PAPR reduction based on
                                                                                                                            tone reservation scheme for DCO-OFDM indoor visible light
                               10
                                     0                                                                                      communications," Optics Express 25(20), 24630-24638 (2017)
                                                                                              Companding
                                                                                              SLM-C µ=3              [9]    N. Ali, R. Almahainy, A. Al-Shabili, N. Almoosa and R. Abd-
                                                                                                                            Alhameed, "Analysis of improved μ-law companding technique for
                                     -1
                               10                                                             SLM-C µ=2
                                                                                              SLM-C µ=1
                                                                                              SLM                           OFDM systems," in IEEE Transactions on Consumer Electronics, vol.
                               10
                                     -2
                                                                                                                            63, no. 2, pp. 126-134, May 2017.
                                                                                                                     [10]   T. Kojima, S. Iwamoto, Y. Shida and T. Fujino, "A novel SLM PAPR
                 BER
                                     -3
                               10
                                                                                                                            reduction of OFDM signals without side information," The 10th IEEE
                                     -4
                                                                                                                            International Symposium on Signal Processing and Information
                               10
                                                                                                                            Technology, Luxor, 2010, pp. 321-325.
                               10
                                     -5                                                                              [11]   H. Kaushal and G. Kaddoum, "Underwater Optical Wireless
                                                                                                                            Communication," IEEE Access 4(1518-1547 (2016)
                               10
                                     -6
                                                                                                                     [12]   Z. Zeng, S. Fu, H. Zhang, Y. Dong and J. Cheng, "A Survey of
                                          0       2       4        6       8        10        12           14
                                                                  E b/N0 (dB)                                               Underwater       Optical     Wireless   Communications,"     in IEEE
                                                                                                                            Communications Surveys & Tutorials, vol. 19, no. 1, pp. 204-238,
 Fig. 4. Comparison of BER between different schemes and different μ values                                                 Firstquarter 2017.
                       in UOWC DCO-OFDM system                                                                       [13]   J. Alkhasraji and C. Tsimenidis, "Coded OFDM over short range
                                                                                                                            underwater optical wireless channels using LED," OCEANS 2017 -
                                                                                                                            Aberdeen, Aberdeen, 2017, pp. 1-7.
                                                          V. CONCLUSION
                                                                                                                     [14]   Fei Wang, Yi Yang, Zuoliang Duan,ect, "Characteristic analysis of
   Through the combination of SLM and μ-law companding                                                                      underwater laser signal transmission channel based on          visible
methods, the PAPR of UOWC DCO-OFDM system is greatly                                                                        light,"optical communication technology,2016,40(03):26-28.
reduced with faster convergence than both the SLM and μ-law                                                          [15]   J. Armstrong, “OFDM for optical communications,” Journal of
                                                                                                                            Lightwave Technology 27(3), 189–204 (2009).
                                                                                                978-1-5386-6158-1/18/$31.00 ©2018 IEEE