Performance Assessment of OFDM-based and OWDM-based Radio-over-Fiber Systems in the Presence of Phase Noise

  • 249 views
Uploaded on

For more projects please contact us @ www.nsrcnano.com

For more projects please contact us @ www.nsrcnano.com

More in: Technology
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
    Be the first to like this
No Downloads

Views

Total Views
249
On Slideshare
0
From Embeds
0
Number of Embeds
0

Actions

Shares
Downloads
0
Comments
0
Likes
0

Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide

Transcript

  • 1. Performance Assessment of OFDM-based and OWDM-based Radio-over-Fiber Systems in the Presence of Phase Noise Gholamreza Baghersalimi Department of Electrical Engineering, University of Guilan, Rasht, Iran Abstract: In this paper the impact of a Radio-over- Fiber (RoF) optical subsystem on the sensitivity to the phase noise of an Orthogonal Frequency Division Multiplexing (OFDM) system using Discrete Fourier Transform (DFT) and Discrete Wavelet Transform (DWT) are evaluated and compared by computer simulation. The study investigates the effect of phase jitter on the system Bit Error Rate (BER) of the DFT/DWT-based OFDM for different modulation schemes in the presence of optical sub-system's nonlinearities in AWGN channel. The results of this research show that at low phase noise levels, some DWT-based schemes outperform DFT-based scheme, however they achieve the same performance at high phase noise levels. It is understood that RoF systems are less robust against phase noise compared to their linear counterparts. Also, increasing modulation order leads to an increase in sensitivity to phase noise regardless of the Multi-Carrier (MC) scheme used. Furthermore, the performance of some wavelets are slightly different, especially at low ࣌ (the variance of the phase noise) values. Finally, QAM signals are more robust against phase noise compared to PSK signals in RoF systems. KEYWORDS: RoF, OFDM, OWDM, Wavelet, AWGN, Phase Noise. I-INTRODUCTION Radio-over-Fiber (RoF) is a technology by which information bearing signals using RF carries are delivered by means of optical components and techniques. Better coverage and increased capacity, centralized upgrading and adaptation, higher reliability and lower maintenance costs, support for future broadband applications, and economic access to mobile broadband are among the most important advantages of RoF [1], [2]. However, RoF systems are vulnerable to nonlinearities in the optical subsystem that cause degradation of the system BER performance. Normally, these effects are expressed as AM-AM and AM-PM characteristics; the former is an amplitude transfer function while the latter is a phase transfer function. One area of interest in modern communications is OFDM which is becoming widely used in wireless communication systems due to its high data rate transmission capability with high bandwidth efficiency and also its robustness to multi-path fading without requiring complex equalization techniques [3], [4], [5]. OFDM has been adopted in a number of wireless applications including Digital Audio Broadcast (DAB), Digital Video Broadcast (DVB), Wireless Local Area Network (WLAN) standards such as IEEE802.11g and Long Term Evolution (LTE) [6].To mitigate Inter-Symbol Interference (ISI), a cyclic prefix (CP) is used which in turn leads to spectral inefficiency [3], [5]. In comparison to single carrier systems, more vulnerability to frequency offset and phase noise are OFDM disadvantages. In recent years, wavelet transform has been suggested to replace DFT in OFDM systems. While signals in DFT-OFDM systems overlap in the frequency domain only, DWT- OFDM signals overlap in the time domain as well, so there is no need for the CP as in the DFT-OFDM case [7], [8]. Hence, some savings in bandwidth can be achieved.Hereafter DFT- OFDM and DWT-OFDM are referred as OFDM and OWDM (Orthogonal Wavelet Division Multiplexing), respectively. In wavelet transform, a signal is decomposed into shifted and scaled versions of a particular wavelet called the mother wavelet [7], [8], [9]. The reverse operations are carried out to reconstruct the original signal. Therefore, not only can the CP be removed in OWDM but also P/Sand S/P blocks can be removed from the OWDM transmitter and receiver, respectively. In coherent communication systems, the receiver must provide carrier and symbol synchronization capabilities. Misalignment between oscillator frequencies of receiver and 978-963-8111-77-7
  • 2. Fig.1 Block Diagram of MC RoF. transmitter or Doppler shift will result in carrier frequency offset , or equivalently, a phase error of the received signal relative to the transmitted signal [4], [10]. Carrier Frequency Offset (CFO) destroys the orthogonality between subcarriers and therefore prevents perfect alignment of FFT bins with peaks of the Sinc shaped pulses. As a result the energy of each subcarrier is spread to other subcarriers leading to Inter-Carrier Interference (ICI). This paper investigates the impact of an optical subsystem of an OFDM/OWDM based multicarrier (MC) RoF system on the BER performance in the presence of phase noise. The paper is organized as follows. Section II discusses the system model and experimental setup followed by results and discussion in Section III. Finally, Section IV concludes the paper. II-EXPERIMENTAL SETUP In order to evaluate the system performance, computer simulations were carried out based on the system model presented in Fig. 1. The data source transmits 1,000,000 bits. For the sake of simplicity and focusing on the phase distortion itself, operations such as coding and interleaving are not considered. Then, the data are mapped using a QPSK/16PSK/16QAM modulator. To produce OFDM symbols, first the resultant signal is converted from serial to parallel. Then, either an Inverse Discrete Fourier Transform (IDFT) or an Inverse Discrete Wavelet Transform (IDWT) is taken. The resultant signal in the OFDM case is converted to serial , and in order to mitigate ISI, a CP with a length of 25% of the whole OFDM symbol period is added. This signal i.e. the OFDM/ OWDM signal is Fig. 2 AM-AM (top pane) and AM-PM (bottom pane) Characteristics. pulse shaped using an RRC filter (roll-off=0.5 and number of taps=64), and then passed to the optical subsystem. The optical subsystem comprises a laser diode (LD), a 2.2 km length of single-mode fibre, and a PIN diode for photo-detection. Fig.e 2 portrays the measured AM-AM/PM characteristics of the optical subsystem which is reproduced from [11]. Also, a pictorial description of Output Back-Off ( OBO) is shown in Fig. 2 which is defined (on a logarithmic scale) as the difference between the maximum output power and the output power at the quiescent point. After passing through an RF AWGN channel, the signal is perturbed by a multiplicative noise that models the phase noise. There are two phase noise (jitter) models namely the white phase noise model and the colored phase noise model [4]. In this paper, the phase noise is modelled as a zero- mean white Gaussian stochastic process with a standard deviation of ߪ ranging from 0 to 90 degrees with steps of 10 degrees.
  • 3. TABLE I SIMULATION PARAMETERS Parameter Value Number of DFT/DWT Points 64 Modulation QPSK / 16QAM Wavelet Family Haar,Daubechies,Symlet, Biorthogonal, Reverse Biorthogonal Number of bits 1,000,000 Output Back-off[dB] 0.2, 0.5,1 Channel AWGN Cyclic Prefix (CP) 25% for OFDM No CP for OWDM Standard Deviation of Phase Noise ( ߪ) 0-90 o with steps of 10 degrees Required ‫ܧ‬௕/ܰ௢ at target ‫ܴܧܤ‬ = 10ିଷ QPSK: 6.78 dB 16PSK: 14.32 dB 16QAM: 10.5 dB Then the resultant signal is filtered by the receive RRC filter followed by the CP removal from the OFDM signal, and thereafter it is converted to parallel symbols. Subsequently, a DFT /DWT is taken followed by a conversion to serial data. For a fair comparison of the effect of carrier phase noise, no channel estimation is performed. After being demodulated, the received data are compared with those transmitted. Simulation parameters are summarized in Table 1. III-RESULTS AND DISCUSSION According to simulations, BER degradation due to phase noise can be explained using the following categories. A- The Effect of the MC scheme (i.e. the difference between wavelets and OFDM) for the linear case Primarily for a benchmark, simulations were carried out for the linear system i.e. the RoF system without an optical subsystem. Fig. 3 plots BER versus the standard deviation of the phase noise ( ߪ) for different wavelets together with that of OFDM for the linear MC system. The figure compares OWDM using different wavelets with OFDM when QPSK modulation is used. In this study, the well-known wavelet families listed in Table 1 are used, with wavelets Haar (Sym1), Sym2, db1, db2 , bior1.1, and rbio1.1 being examined for comparison purposes [7], [8], [9]. As can be seen, except for db2 and sym2 (identified as the Group I wavelets) which are very sensitive to phase noise and exhibit an error floor for all values of ߪ, the other MC schemes achieve a monotonically increasing BER behaviour at low Fig. 3 BER vs. standard deviation of phase noise for uncoded MC QPSK in AWGN. Fig. 4 BER vs. standard deviation of phase noise for uncoded MC RoF-QPSK at OBO=1 dB in AWGN. Fig. 5 BER vs. standard deviation of phase noise for uncoded MC RoF-QPSK at OBO=0.2dB (top pane) and OBO=0.5 dB (bottom pane) in AWGN. values of ߪ. In other words, the wavelets Haar, db1, bior1.1, and rbio1.1 (identified as the GroupII wavelets) along with OFDM achieve low BER performance for ߪ < 30௢ while the Group I wavelets achieve almost the same error floor performance for all values of ߪ. When comparing OFDM and OWDM, it is observed that for low ߪ values (< 35௢ ), OWDM outperforms OFDM. On the other hand, OFDM and
  • 4. OWDM show similar BER performances for large values of ߪ. B- Effect of RoF Nonlinearity To investigate the effect of RoF nonlinearity, additional simulations were performed. Fig. 4 shows BER versus ߪ for different wavelets along with that of OFDM for the MC RoF-QPSK at OBO=1 dB. The behaviors of the BER plots follow approximately the same trend as in the linear scenario. One can observe that while wavelet sym2, from Group I, achieves poor performance, OFDM and the other wavelets have comparable BER performance for low ߪ values. In comparison between OFDM and OWDM, it is observed that Group II wavelet types outperform OFDM. At ߪ = 10௢ , for example, the BER degrades from 7 × 10ିଷ (for the Group II wavelets) to 1.2 × 10ିଶ (for OFDM). In particular, there is no noticeable difference between different wavelet types regarding phase noise. To confirm the above statements, more simulations were carried out at OBO levels of 0.2 dB and 0.5 dB. The results, which are shown in Figure 5, follow the same trends as those for the OBO= 1dB case. C-Effect of Modulation Order All the conclusions in Parts 1 and 2 were obtained for QPSK modulation. To assess the effect of modulation order, other simulations were carried out using 16- PSK at OBO=1 dB . Results, which are shown in Fig. 6, show that Group II and OFDM achieve a better performance compared to Group I. In comparison between OFDM and Group II wavelets, it is observed that there is no noticeable difference between them. Also, 16PSK compared to QPSK is more sensitive to phase noise. D-Effect of Modulation Type So far, just PSK modulation has been considered. To study the effect of modulation type, other simulations were carried out with 16QAM modulation for the RoF system at OBO=1 dB. The results are shown in Fig. 7. It is clearly observed that by changing the modulation type the aforementioned statements hold. Again the wavelets Haar, db1,bior1.1, and rbio1.1 outperform OFDM. However, the difference between different BER plots depends on OBO. As expected, 16QAM is more sensitive to phase noise than QPSK in the presence of nonlinear- Fig. 6 BER vs. standard deviation of phase noise for uncodedMCRoF-16PSK at OBO=1 dB in AWGN. Fig. 7 BER vs. standard deviation of phase noise for uncodedMCRoF-16QAM at OBO=1 dB in AWGN. ity. At ߪ = 10௢ for the Group II wavelets, for example, the BER degrades from 5 × 10ିଷ for the QPSK scheme to 2.5 × 10ିଶ for the 16QAM scheme. On the other hand, it is lesssensitive to phase noise than 16PSK as the latter achieves a BER of 0.1 for the same conditions. As an overall comparison, dimension-less parameter D -a criterion for BER degradation due to phase noise- is defined as the ratio of the target BER in the linear case to the achieved BER in the presence of phase noise i.e. ‫ܦ‬ = ܶܽ‫ݐ݁݃ݎ‬ ‫ݎ݋݂ܴܧܤ‬ ‫ݐ‬ℎ݁ ݈݅݊݁ܽ‫ݎ‬ ܿܽ‫݁ݏ‬ ‫ܿܣ‬ℎ݅݁‫݀݁ݒ‬ ‫ܴܧܤ‬ ݅݊ ‫ݐ‬ℎ݁ ‫݁ܿ݊݁ݏ݁ݎ݌‬ ‫݂݋‬ ‫݌‬ℎܽ‫݁ݏ‬ ݊‫݁ݏ݅݋‬ where D is upper-bounded by unity (the ideal value). Smaller D values correspond to more sensitivity to phase noise and vice versa. As a reference, D values in percent for ߪ = 10௢ (a low value) at OBO=1 dB are listed in Table 2. From Table 2, the following key conclusions, which are valid for low ߪ values, can be drawn. ¾ In the linear scenario using QPSK, while Group II is less sensitive to
  • 5. TABLE II D Values in Percent for Different MC Schemes for ߪ = 10௢ at OBO=1 dB. Group I OFDM Group II [db2, sym2] [ Haar,db1 , bior1.1,rbio1.1] MC-QPSK (Linear) 0.67 12.5 25 RoF-QPSK 0.67 9 14 RoF-16PSK 0.33 1 1 RoF-16QAM 0.33 2 2.5 Phase noise compared to OFDM (D=25 vs.D=12.5), it outperforms Group I (D=12.5 vs. D=0.67). ¾ Regarding the RoF system –a typical system with both phase and amplitude nonlinearities- such systems are less robust to phase noise compared to their linear counterpart (D=14 vs. D=25 for Group II or D=9 vs. D=12.5 for OFDM). ¾ Regarding modulation order, it is observed that increasing modulation order increases sensitivity to phase noise regardless of the MC scheme (D=14 vs. D=1 for Group II or D=9 vs. D=1 for OFDM). ¾ Regarding modulation type, one can observe that phase modulated (PSK) signals are more sensitive to phase noise compared to phase- and amplitude-modulated (QAM) signals. IV-CONCLUSION In this paper the impact of a RoF optical sub-system on the BER performances of OFDM and OWDM were assessed in the presence of phase noise. It was found that RoF systems are less robust to phase noise compared to their linear counterparts. At low ߪ values, some OWDM schemes outperform OFDM, however they achieve the same performance at high ߪ values. Also, increasing modulation order leads to an increase in sensitivity to the phase noise regardless of the MC scheme. In addition, the performance of different wavelets are slightly different, especially at low ߪ values. Finally, MQAM signals are more robust to phase noise than MPSK signals in a RoF system. Performance assessment of the aforementioned modulation schemes in the presence of colored phase noise, both in AWGN and time- varying RF channels, is the subject of further investigation by the author. REFERENCES [1] H. Al-Raweshidy, and S. Komaki,” Radio over-Fiber Technologies for Mobile Communications Networks”, Artech House, 2002. [2] J. Mitchell, “Performance of OFDM at 5.8 GHz using radio over fibre link,” Electronics Letters, vol. 40, No. 21, pp. 1353 – 1354, October 2004. [3] R. Van Nee and R. Prasad, OFDM for Wireless Multimedia Communications. Artech House, 2000 [4] L. Hanzo, M. Munster, B. J. Choi, and T. Keller, OFDM and MC-CDMA for Broadband Multi-User Communications, WLANS and Broadcasting. IEEE Press, 2003. [5] A. R. Sheikh Bahai, B. R. Saltzberg, and M. Ergen, Multi- Carrier Digital Communications: Theory and Applications of OFDM. Springer, 2004 [6] IEEE, “IEEE std 802.11g; supplement to IEEE standard for information technology telecommunications and information exchange between systems-local and metropolitan area networks specific requirements-part 11: Wireless LAN medium access control(MAC) and physical layer (PHY) specifications; amendment 4: Further higher datarate extension in the 2.4 GHz band,” tech. rep., IEEE, 2003. [7] Sobia Baig, Fazal-ur-Rehman, M. junaid Mughal ,“Performance Comparison of DFT, Discrete Wavelet Packet and Wavelet Transforms, in an OFDM Transceiver for Multipath Fding Channel”, IEEE Communication Magazine,2004 [8] Deepak Gupta, Torry Harris, Vipin B Vats, Kamal K. Garg,“ Performance Analysis of DFT-OFDM,DCT-OFDM and DWT- OFDM in AWGN channel”, IEEE,The Fourth International Conference on Wireless Mobile Communications,2007. [9] F. Farrukh, S. Baig, and M. J. Mughal ,“Performance Comparison of DFT-OFDM and Wavelet-OFDM with Zero- Forcing Equalizer for FIR Channel Equalization”, IEEE Communication Magazine,2008. [10] D. Karamehmedovic, M.K. Lakshmanan and H. Nikookar,“ Performance of Wavelet Packet Modulation and OFDM in the Presence of Carrier Frequency and Phase Noise”, Proceedings of the 1st European Wireless Technology Conference, pp.166-169, 2008. [11] X. N. Fernando, and A. B. Seasay ,” Characteristics of directly modulated RoF link for wireless access”, CCECE2004, Volume 4, Pages: 2167 – 2170, 2-5 May 2004.