Anda di halaman 1dari 5

INTERNATIONAL CONFERENCE ON COMMUNICATION, COMPUTER AND POWER (ICCCP09)

MUSCAT, FEBRUARY 15-18, 2009

Impulsive Noise Effects on DWT- and WPTOFDM versus FFT-OFDM


Khaizuran Abdullah
1

and Zahir M. Hussain1 , SMIEEE

Abstract We demonstrate a performance study on waveletbased OFDM, particularly using DWT and WPT-OFDM as substitutions for Fourier- based OFDM with the focus on impulse noise effects. We begin by constructing the models of the inverse and forward transforms. We explain in detail each model and study the BER performance in two scenarios when varying the Poisson recurrence parameter a from small to large. The waveletbased OFDM (DWT-OFDM and WPT-OFDM) are assumed to have orthonormal bases and perfect reconstruction properties. Results show that a large value of a limits the impact of impulsive noise on the system. We also found that the WPT-OFDM platform is superior in BER to others.

-Keywords: Impulse noise, Discrete Wavelet Transform, Wavelet Packet Transform, Fourier-based OFDM, waveletbased OFDM. I. I NTRODUCTION Orthogonal Frequency Division Multiplexing (OFDM) is a promising multicarrier wireless system for transmission of high-rate data stream with power efciency and fading immunity. Conventional OFDM systems use IFFT and FFT algorithms at the transmitter and receiver respectively to multiplex the signals and transmit them simultaneously over a number of subcarriers. The system employs guard intervals or cyclic prexes (CP) so that the delay spread of the channel becomes longer than the channel impulse response to minimize inter-symbol interference (ISI). However, the CP has the disadvantage of reducing the spectral containment of the channels. Alternative methods are to use wavelet transforms replacing IFFT and FFT [1], [2], [5], [6], [7]. By using these transforms, the spectral containment of the channels is better since they are not using CP [1], [2], [5], [6]. They can be considered as Discrete Wavelet Transform OFDM (DWT-OFDM) or Wavelet Packet Transform OFDM (WPT-OFDM). Both transforms employ Low Pass Filter (LPF) and High Pass Filter(HPF) operating as Quadrature Mirror Filters satisfying perfect reconstruction and orthonormal bases properties. The transforms use lter coefcients as approximate and detail in LPF and HPF respectively. The approximated coefcients is sometimes referred to as scaling coefcients, whereas, the detailed is referred to wavelet coefcients [3]. Sometimes these two lters can be called subband coding since the signals are divided into sub-signals of low and high frequencies respectively.
1 School of Electrical and Computer Engineering, RMIT University, Melbourne, Victoria 3000, Australia. E-mails: khaizuran.abdullah@student.rmit.edu.au, zmhussain@ieee.org.

The purpose of this paper is to view the effects of impulse noise on the wavelet-based OFDM particularly using DWT and WPT-OFDM as substitutions for Fourier-based OFDM. A recent work has focussed on the effect of impulse noise when wavelet packet division multiplexing (WPDM) is used [8]. Some discussions are related to the performance comparison between OFDM and Time Division multiplexing (TDM). However, there is no indication of comparative study and performance using DWT and WPT as alternative replacement of FFT. Although the studies in [9] provides strong analysis of impulsive noise and its effect on the performance of OFDM system, the discussion is mainly for the application in the power line communications (PLC). To the best of the Authors knowledge, there is no work showing the steps of how to simulate exible transformed models of DWT- and WPTOFDM as alternative replacements of FFT-OFDM under the effect of impulse noise. This paper is divided into four main sections: section II will briey explain conventional FFTOFDM, section III will describe in detail the models for DWTOFDM and WPT-OFDM, section IV will provide the impulse noise effects on the OFDM system, and section V will discuss the bit error rate (BER) performance considering two different scenarios of impulse noise effects. II. F OURIER -BASED OFDM A typical block diagram of an OFDM system is shown in Fig. 1. The inverse and forward blocks can be FFTbased, DWT-based or Wavelet Packet (WP)-based OFDM. The system model for FFT-based OFDM will not be discussed in detail as it is well known in the literature. Thus, we merely present a brief description about it. The data dk rst being processed by a constellation mapping. M-ary QAM modulator is used for this work to map the raw binary data to appropriate QAM symbols. These symbols are then input into the IFFT block. This involves taking N parallel streams of QAM symbols (N being the number of sub-carriers used in the transmission of the data) and performing an IFFT operation on this parallel stream. The output in discrete time domain is as follows: 1 Xk(n) = N
N 1

Xm(i) exp j2
i=0

n i N

(1)

where Xk(n) |0 n N 1 is a sequence in the discretetime domain and Xm(i) |0 i N 1 are complex numbers in the discrete frequency domain. The cyclic prex (CP) is lastly added before transmission to minimize the inter-symbol interference (ISI). At the receiver, the process is reversed to
373

c SQU-2009 ISSN: 1813-419X

INTERNATIONAL CONFERENCE ON COMMUNICATION, COMPUTER AND POWER (ICCCP09)

MUSCAT, FEBRUARY 15-18, 2009

Fig. 1: A Typical model of an OFDM transceiver with inverse and forward transforms that can be substituted as FFT-, DWT- or WPTOFDM.

obtain the decoded data. The CP is removed to obtain the data in the discrete time domain and then processed to FFT for data recovery. The output of the FFT in the frequency domain is as follows: Um(i) n = Uk(n) exp j2 i N n=0
N 1

Fig. 2: An Inverse and Forward Discrete Wavelet Transform DWTOFDM model. The synthesis lters (transmitter part) are at the top and the analysis lters (receiver part) are at the bottom.

(2)

lter coefcients, the orthonormal bases can be satised via four possible ways of the dot products as follows [3]: < g.g >= 1 < h.h >= 1 < g.h >= 0 < h.g >= 0 (3) (4) (5) (6)

III. WAVELET-BASED OFDM As mentioned in the previous section, the inverse and forward block transforms are exible and can be substituted with FFT-, DWT- or Wavelet Packet (WPT)-OFDM. We have discussed briey about FFT-OFDM. Thus, this section will describe wavelet based OFDM and is divided into two parts: description of the DWT-OFDM model as well as discussing the WPT-OFDM model. A. Discrete Wavelet Transform (DWT) The model of DWT-OFDM transceiver is shown in Fig. 2. In the top part, the transmitter rst uses a digital modulator (i.e., 16 QAM ) which maps the serial bits into symbols converting dk into Xm , within N parallel data stream Xm(i) where Xm(i) |0 i N 1. The main task of the transmitter is to perform the discrete wavelet modulation by constructing orthonormal wavelets. Each Xmi is rst converted to serial representation having a vector XX which will next be transposed into CA. This mean that CA not only its imaginary part have inverting signs but also its form is changed to a parallel matrix. Then, the signal is up-sampled and ltered by the LPF coefcients or namely as approximated coefcients. Since our aim is to have low frequency signals, the modulated signals XX perform circular convolution with LPF lter whereas the HPF lter also perform the convolution with zeroes padding signals CD respectively. Note that the HPF lter contains detailed coefcients or wavelet coefcients. Different wavelet families have different lter length and values of approximated and detailed coefcients. Both of these lters have to satisfy orthonormal bases in order to operate as wavelet transform. This means that they must be orthogonal and normal to each other. By assigning g as LPF lter coefcients and h as HPF

where (3) or (4) is related to the normal property and (5) or (6) is for orthogonal property accordingly. The star superscripts indicate that the coefcients are conjugate elements. Both lters are also assumed to have perfect reconstruction property. This means that the input and output of the two lters are expected to be the same [3]. In the transmitter part, this signal is simulated using MATLAB command [Xk ] = idwt(CA, CD, wv) where wv is the type of wavelet family. On the other hand, the reverse process is simulated using [ca, cd] = dwt(Uk , wv) in the receiver. The ca signal will be processed to the QAM demodulator for data recovery. However, the cd signal is discarded because it does not contain any useful information. B. Wavelet Packet Transform (WPT) A wavelet packet transform (WPT) model is illustrated in Fig. 3. The wavelet packet forming block can be simulated using the following MATLAB command T = wptree(2, 3, xx, wv). The purpose for this command is to create the wavelet packet tree T to be processed for the reconstruction block of the WPT. The rst parameter 2 is the order number depending on the input vector signal xx. The second parameter is the number of level which is depending on the size of xx and wv is the wavelet family. Furthering process to this, the MATLAB code Xk = wprec(T, wv) is invoked. The input signals of WPT block has to be a wavelet packet tree, otherwise, the signal cannot be processed for transmission. Fig. 4 shows the wavelet packet tree corresponding
374

c SQU-2009 ISSN: 1813-419X

INTERNATIONAL CONFERENCE ON COMMUNICATION, COMPUTER AND POWER (ICCCP09)

MUSCAT, FEBRUARY 15-18, 2009

synchronization, rk as indicated in Fig. 1 is written as rk = yk + gk k = 0, 1, 2, ...L 1 (7) where yk is the transmitted OFDM signal, gk is the noise consisting of AWGN and impulsive noise, and is given by gk = wk + ik (8)

where wk is the additive Gaussian process with mean zero and 2 variance w and ik can be expressed as ik = k zk (9) In this case, k is the Poisson process indicating the arrival of impulsive noise and z(n) is the white Gaussian process with 2 2 mean 0 and variance z . It is generally assumed that z is 2 much larger than w . Note that the impulsive noise having 2 variance z amplitude occurs during the length of L samples. The occurrence of the impulsive noise generally is the Poisson distribution of a random variable X can be expressed as [9] k = 0, 1, 2, ..., L 1 (10) where a is the Poisson parameter which is the average value of Poisson random variables. P (k) = P (X = k) = exp[a( V. E XPERIMENTAL R ESULTS AND D ISCUSSION All of the transformed models in FFT-,DWT- and WPTOFDM systems used the parameters as shown in Table I. The number of samples for the subcarriers N is 64, and the number of samples for the symbols ns is 1000. Other variables are listed according to their use as in Figs. 1, 2 and 3. The signalto-noise ratio (SNR) for all the simulations is determined as Px Px = 2 (11) SN R = Pn w + Pi where Px is the mean power of the transmitted OFDM signal, 2 w is the mean power of the Gaussian noise and Pi is the 2 mean impulsive noise power. The ratio r between Pi and w 2 is dened as r = Pi /w . Since the impulsive noise follows the Poisson distribution in timing, the equation (11) can be rewritten as Px Px SN R = = 2 (12) 1 2 Pn w + a z where a is the Poisson parameter as indicated in equation (10) or it is the average value for Poisson random variables occurring during L = N ns length of samples. In our simulation, we are interested to vary the values of a with the value of r = 10. When a is small, the received OFDM signal rk has many impulse noise samples as compared to when a is large. Examples of typical samples of the received OFDM signals having different a = 5 (small value) and a = 50 (large value) are shown in Figs. 5a and 5b. We have divided this section in two parts; scenario I: a = 5 and r = 10, and scenario II: a = 50 and r = 10. Note that the value of r, which is the ratio value of Pi over Pn , is not varied because it will not take much effect of BER results even though the three platforms are different. On the other hand, a will have signicant change of performance since it is related to the exponential function as indicated in equation (10).
375

Fig. 3: Inverse and Forward Wavelet Packet Transform WPT-OFDM


model.

ak )] k!

Fig. 4: A Wavelet Packet tree corresponding to the T block.

to the T blocks in the WPT-OFDM transceiver in Fig. 3. At each branch, the wavelet packet forms the wavelet basis function which is split into scaling and wavelet coefcients corresponding to the LPF and HPF coefcients respectively [8]. This also means that they have orthonormal bases and perfect reconstruction properties as discussed in the previous section. Examples of the wavelet basis function at level 3 are (3, 0), (3, 1), (3, 2), (3, 3), (3, 4), (3, 5), (3, 6), and (3, 7). Due to this characteristic, the wavelet packets offer a more complex and exible operation. The number of nodes depending on the number of levels. At the front end of the receiver, the signals Uk is received by the system and processed by the WP forming block. For this purpose, the same command as in the transmitter is invoked. After the WP tree is formed, the WPT reconstruction will output the wavelet basis functions at the last terminals to be further process to the QAM demodulator for the data recovery. IV. I MPULSIVE N OISE E FFECT Impulsive noise effect to the OFDM system has been discussed in many literatures [8], [9], [10]. In this section, we describe the general principles of the impulse noise when it affects an OFDM system. We also mention the recurrence parameter of Poisson distribution which will affect the system performance. Assuming that the receiver having perfect c SQU-2009 ISSN: 1813-419X 3

INTERNATIONAL CONFERENCE ON COMMUNICATION, COMPUTER AND POWER (ICCCP09)

MUSCAT, FEBRUARY 15-18, 2009

15

10

z > w when a=5


10

10

Bit Error Rate

10

Real (rk)

10

FFTOFDM: No Impulse FFTOFDM: Impulse DWTOFDM: No Impulse DWTOFDM: Impulse WPTOFDM: No Impulse WPTOFDM: Impulse 0 5 10 15 20 25 30 35 40

10

SNR per bit, or EbN0 in dB

Gaussian Gaussian + impulse noise


10 0 20 40 60 80 100 120 140

time

Fig. 6: BER performance of Scenario 1 for all the three transforms


when a = 5 and r = 10.

(a) Impulsive noise effect when a=5

15

10

2 > 2 when a=50


z w

Gaussian Gaussian + impulse noise


10 0 20 40 60 80 100 120 140

time

(b) Impulsive noise effect when a=50

Fig. 5: Sequence Samples of two OFDM symbols with impulsive noise effect, (a) a = 5 and (b) a = 50. TABLE I: Simulation variables and their matrix values.
Variables N ns CP wv d (N ns) Xm (N ns) xx Xk Uk uu Um (N ns) d (N ns) FFT-OFDM Matrix Values 64 1000 8 64 1000 64 1000 1 64000 64000 1 64000 1 1 64000 64 1000 64 1000 DWT-OFDM Matrix Values 64 1000 0 bior5.5 64 1000 64 1000 1 64000 128000 1 128000 1 1 64000 64 1000 64 1000 WPT-OFDM Matrix Values 64 1000 0 bior5.5 64 1000 64 1000 1 64000 64000 1 64000 1 1 64000 64 1000 64 1000

a heavily disturbed since the system will have more frequent impulse noise peaks as indicated in Fig. 5a. It is interesting to see that WPT-OFDM signal shows minimal BER when there is no impulse noise; also when the impulse noise is present. At SNR of 20 dB, the BER of WPT-OFDM is about 0.07 as compared to 0.1 and 0.18 of DWT and FFT-OFDM 2 2 respectively when there is impulse noise (z >>> w ). This also means that WPT-OFDM produces about 3 and 11 less errors in 100 of the OFDM received samples as compared to DWT- and FFT-OFDM respectively. This is due to the fact that the wavelet packet tree produces wavelet basis functions at the terminal of level 3 having about the same basis functions at the same terminal as in the transmitter satisfying the perfect reconstruction and orthonormal bases properties. The DWTOFDM also satises those properties when it deals with the LPF and HPF coefcients in the transmitter and receiver. However, the WPT-OFDM has the characteristics of forming the wavelet basis function which is the result of the spliting process of scaling and wavelet coefcients corresponding to the LPF and HPF coefcients as discussed in section III-B. B. Results: Scenario II The results of Scenario II when a = 50 and r = 10 for all platforms are shown in Fig. 7. In this scenario, the condition of impulse noise effect is less impacted since the value of a is large as illustrated in Fig. 5b. It is shown that WPT-OFDM is superior to all as discussed in the previous section. To be specic, the BER of WPT-OFDM is about 0.015 as compared to 0.03 and 0.1 of DWT and FFT-OFDM respectively at SNR of 20 dB in the presence of impulse noise. This also shows that the WPT obtain about 15 and 85 less errors in 1000 of the OFDM received samples as compared to DWT and FFT platforms accordingly. Comparing the BER of Scenarios I and II, we can point out that the performance for this section is better since it obtained the result of 0.015 as compared to 0.07 in Scenario I within the same SNR value. This shows that our results match with the values of a as discussed in section IV because the effect of the impulsive noise is less when the value
376

A. Results: Scenario I Fig. 6 shows the results of Scenario I when a = 5 and r = 10 for all platforms. This condition can be referred to as

c SQU-2009 ISSN: 1813-419X

Real (r )

INTERNATIONAL CONFERENCE ON COMMUNICATION, COMPUTER AND POWER (ICCCP09)

MUSCAT, FEBRUARY 15-18, 2009

10

10

Bit Error Rate

[9] Y. H. Ma, P. L. So, and E. Gunawan, Performance analysis of OFDM systems for broadband power line communications under impulsive noise and multipath effects, IEEE Transactions on Power Delivery, vol. 20, no. 2, pp. 674-682, 2005. [10] M. Ghosh, Analysis of the effect of impulse noise on multicarrier and single carrier QAM systems, IEEE Transactions on Communications, vol. 44, no. 2, pp. 145-147, 1996.

10

10

FFTOFDM: No Impulse FFTOFDM: Impulse DWTOFDM: No Impulse DWTOFDM: Impulse WPTOFDM: No Impulse WPTOFDM: Impulse 0 5 10 15 20 25 30 35 40

10

SNR per bit, or EbN0 in dB

Fig. 7: BER performance of Scenario 2 for all the three transforms when a = 50 and r = 10.

of a is large. Future work may include a mitigation technique to optimize the performance and obtain more less error. VI. C ONCLUSIONS We presented a performance study on DWT-OFDM and WPT-OFDM as substitutions for FFT-OFDM, with focus on the effects of impulse noise. We discussed in details the MATLAB commands regarding the DWT-OFDM and WPTOFDM platform models. Performance in terms of BER was also obtained while varying the Poisson distribution parameters. Our results showed that impulse noise has less impact on the system when its recurrence parameter a is large. The BER performance of the WPT-OFDM system is shown to be superior to others. R EFERENCES
[1] R. Mirghani, and M. Ghavami, Comparison between Wavelet-based and Fourier-based Multicarrier UWB Systems, IET Communications, Vol. 2, Issue 2, pp. 353-358, 2008. [2] R. Dilmirghani, M. Ghavami, Wavelet Vs Fourier Based UWB Systems, 18th IEEE International Symposium on Personal, Indoor and Mobile Radio Communications, pp.1-5, Sep. 2007. [3] M. Weeks, Digital Signal Processing Using Matlab and Wavelets, Innity Science Press LLC, 2007. [4] S. R. Baig, F. U. Rehman, and M. J. Mughal, Performance Comparison of DFT, Discrete Wavelet Packet and Wavelet Transforms in an OFDM Transceiver for Multipath Fading Channel,, 9th IEEE International Multitopic Conference, pp. 1-6, Dec. 2005. [5] N. Ahmed, Joint Detection Strategies for Orthogonal Frequency Division Multiplexing, Dissertation for Master of Science, Rice University, Houston, Texas. pp. 1-51, Apr. 2000. [6] S. D. Sandberg, and M. A. Tzannes, Overlapped Discrete Multitone Modulation for High Speed Copper Wire Communications, IEEE Journal on Selected Areas in Communications, vol. 13, no. 9, pp. 1571-1585, 1995. [7] A. N. Akansu, and L. Xueming, A Comparative Performance Evaluation of DMT (OFDM) and DWMT (DSBMT) Based DSL Communications Systems for Single and Multitone Interference, Proceedings of the IEEE International Conference on Acoustics, Speech and Signal Processing, vol. 6, pp. 3269 - 3272, May 1998. [8] K. M. Wong, W. Jiangfeng, T. N. Davidson, J. Qu, and P. C. Ching, Performance of wavelet packet-division multiplexing in impulsive and Gaussian noise, IEEE Transactions on Communications, vol. 48, no. 7,pp. 1083-1086, 2000.

377

c SQU-2009 ISSN: 1813-419X

Anda mungkin juga menyukai