Wireless Image Transmissions over Frequency Selective Channel Using Recent OFDMA Systems

Transcription

1 American Jornal of Comptation, Commnication and Control 2018; 5(1): ISSN: Wireless Image Transmissions over Freqency Selective Channel sing Recent OFDA Systems Faisal Saif Al-kamali, Farok Abd Al-fhaidy, Khaled Abdllah Al-sofy Department of Electrical, Faclty of Engineering and Architectre, IBB niversity, IBB, Yemen address Keywords Image Transmission, DST, DCT, DFT, OFDA, PSNR, SE Received: Janary 15, 2018 Accepted: Febrary 11, 2018 Pblished: arch 23, 2018 Citation Faisal Saif Al-kamali, Farok Abd Al-fhaidy, Khaled Abdllah Al-sofy. Wireless Image Transmissions over Freqency Selective Channel sing Recent OFDA Systems. American Jornal of Comptation, Commnication and Control. Vol. 5, No. 1, 2018, pp Abstract Orthogonal freqency division mltiple access (OFDA) is crrently attracting mch attention in recent wireless commnications to meet the increasing demands arising from the explosive growth of Internet, mltimedia and broadband services. In this paper, the isse of the wireless image transmission over OFDA is investigated for different basis fnctions, different modlation schemes and different sbcarriers mapping schemes over a freqency selective channel. Simlation reslts show that efficient wireless image transmission over OFDA is possible for different basis fnctions. Reslts also show that the discrete cosine transform (DCT)-based OFDA (DCT-OFDA) and discrete sine transform (DST)-based OFDA (DST-OFDA) system provide better performance than the conventional discrete Forier transform (DFT)-based OFDA (DFT-OFDA) system. It is fond that the interleaved systems greatly enhance the clarity of the received image and their performances are better than that of the localized systems. 1. Introdction The demand for mltimedia wireless commnications is growing today at an extremely rapid pace and this trend is expected to contine in the ftre. The common featre of many crrent wireless standards for high-rate mltimedia transmission is the adoption of a mlticarrier air interface based on OFDA. The idea behind OFDA is to convert a freqency selective channel into a collection of freqency-flat sbchannels with partially overlapping spectra. Recently, OFDA has attracted vast research attention from both academia and indstry and has become part of new emerging standards for broadband wireless access. It has been adopted in 3GPP LTE as a downlink scheme and Wiax [1]. In OFDA systems, the available sbcarriers are divided into several mtally exclsive clsters (sbchannels or sbbands) that are assigned to distinct sers for simltaneos transmission. The orthogonality among sbcarriers garantees intrinsic protection against mltiple access interference (AI) while the adoption of a dynamic sbcarrier assignment strategy provides the system with high flexibility in resorce management. Frthermore, OFDA inherits from OFD the ability to compensate channel distortions in the freqency domain withot the need of comptationally demanding time domain eqalizers. De to internet growth, wireless transmission of images and video can be easily performed over mltipath channels. Image transmission over mlticarrier code division mltiple access (C-CDA), orthogonal freqency division mltiplexing (OFD)

2 American Jornal of Comptation, Commnication and Control 2018; 5(1): systems and OFDA has attracted mch attentions in the literatre [2-8]. However, to the best of the athors knowledge, the isse of image transmission over OFDA systems is only considered with the discrete Forier transform based-ofda (DFT-OFDA) and has not been adeqately reported for the discrete cosine transform based-ofda (DCT-OFDA) and the discrete sine transform based-ofda (DST-OFDA), which is the main objective of this paper. In [2, 3], the isse of image transmission over DFT-based C-CDA system was stdied. Performance evalation of image transmission over C-CDA system sing two interleaving schemes was presented in [2]. In [3], an efficient wireless transmission scheme based on the DST-based C-CDA system was proposed. Image transmission over a space-time coded OFD system was sggested and discssed in [4, 5]. Chaotic interleaving for robst image transmission with LDPC coded OFD was proposed in [6]. Recently, Image transmission over OFDA and SC-FDA was reported in [7, 8]. The main objective of this paper is to investigate the isse of wireless image transmission over DFT-OFDA, DCT- OFDA and DST-OFDA systems for different modlation and sbcarriers mapping schemes. In this paper, the DFT- OFDA, the DCT-OFDA and the DST-OFDA system models are derived. In addition, the peak-signal-to-noiseratio (PSNR) and the mean sqare error (SE) performances of the received image over DFT-OFDA, DCT-OFDA and DST-OFDA systems are stdied, compared and investigated for different modlation and sbcarriers mapping schemes. This paper is organized as follows. In Section II, the basis fnctions are described. In Section III, the OFDA system models for different basis fnctions are derived and presented. In this Section, the complexity of the OFDA system with different basis fnctions is investigated and compared. Compter simlation reslts are given in Section IV for mltipath channel environments. Finally, the conclsion of this paper is presented in Section V. 2. Basis Fnctions The transform operation is mathematical operation that is applied to a given signal to convert it from one domain into another domain and vice versa. It is desired to convert the inpt signal from discrete-time data representation into a discrete-freqency representation and vice and versa. Ths once the signal is converted into freqency domain, it will have varios components that can be sed to remove specific nwanted freqency components. The most poplar transform systems related to signal processing are DFT, DST, and DST The Discrete Forier Transform The DFT is one of the most important tools that is widely sed in Digital Signal Processing and related fields. The forward Forier transform is defined as the integral: = The DFT takes N samples in the time-domain and transforms them into N vales X(k) in the freqency-domain. It means that DFT operates at a finite nmber of discrete data points. Ths eq. (1) becomes:!" =.,=0,1,, 1 #$% By sing Eler's formla than eq. (2) can be rewritten as:!" =.&cos #*! +sin#* #$%!. It is clear that DFT takes two parts, real part and imaginary part The Discrete Cosine Transform The DCT is similar to DFT except in the fact that it exploits only the real part of DFT. It expresses the samples of the inpt signal as the sm of cosinsoidal fnctions oscillating at different freqencies. There are for types of DCT, however, the type-ii DCT is the most commonly sed and often called simply the DCT. DCT is given by [9]: (1) (2) (3) =/! Where x(k) is the n th sample of the inpt signal. Β(k) can be given as what follows [9]: " 0=2,=0 1,=0,1,, The Discrete Sine Transform Like any Forier-related transform, DST expresses the inpt signal in terms of a sm of sinsoids with different freqencies and amplitdes. However, it ses only real fnctions instead of the complex fnctions sed in the!" 0 &cos*#1" #$%! (5).,=0,1,.., 1 conventional DFT. There are many types of DST and DST-I type is considered in this paper. DST can be expressed by []:! 5= sin& *#.,=1,2,,!1" #$" 3. OFDA Systems odels 3.1. DFT-OFDA System In this sbsection, DFT-OFDA system is considered (4) (6)

3 32 Faisal Saif Al-kamali et al.: Wireless Image Transmissions over Freqency Selective Channel sing Recent OFDA Systems with sers (terminals) commnicating at the same time with a fixed base station throgh independent mltipath Rayleigh-fading channels as shown in Figre 1. Figre 1. Strctre of the plink DFT-OFDA system. At the transmitter of the DFT-OFDA system, the data symbols are mapped, sing the interleaved or the localized sbcarriers mapping techniqe, which enables DFT-OFDA modlation. Then, an -point inverse discrete Forier transform (IDFT) is performed and a cyclic prefix (CP) is added to the reslting signal. The transmitted signal from the th ser can be formlated as follows: x 1 = PaddF Td (7) where is an N 1 vector containing the modlated symbols of the th (=1,2,., ) ser. T is an N ( =Q.N) sbcarriers mapping matrix of the th ser. Q is the bandwidth expansion factor of the symbol seqence. For perfect time and freqency synchronization, if all terminals transmit N symbols per block, the system can handle Q 1 F simltaneos transmissions withot AI interference. is the IDFT matrix. The generic N-point DFT matrix F pq 1 j2π has entries [ ] F = 1 H F p, q = e, and its inverse is. P add is an (+N C ) matrix, which adds a CP of length N C. P add can be represented as follows: where P = (8) T add [ C, I] C = (9) T [ 0 NC ( N, C) I N ] C At the receiver side, assming perfect time and freqency synchronization, the received signal at the base station can be written as follows: = r H x + n = 1 () Where H is an (+N C ) (+N C ) matrix describing the channel of the th ser. x is an (+N C ) 1 vector containing the transmitted samples of the th ser. n ~ is an (+N C ) 1 vector containing the noise. After the removal of the CP, the received signal becomes: r = P r = H ~ rem Cx + n = 1 (11) where P rem is an (+N C ) matrix, which removes the CP. = P n and ~ are the noise and the n rem x =P transmitted signal after the CP removal, respectively. H C is an circlant matrix describing the channel of the th ser. P rem is given by: rem x P rem [ ( NC) I = 0, ] (12)

4 American Jornal of Comptation, Commnication and Control 2018; 5(1): After that, the received signal is transformed into the freqency domain via an -points DFT as follows: = R Λ X + N = 1 (13) where X = F ~ x is an 1 vector representing the transmitted samples from the th ser after the mapping process. N is the DFT of n. After the demapping process, and the eqalization are applied. Finally, the demodlation and the decoding processes take place DCT-OFDA System The strctre of the DCT-OFDA system is shown in Figre 2. In matrix notation, the transmitted signal of the th ser ( = 1, 2,., ) can be formlated as follows: x 1 = PaddD Td (14) Figre 2. Strctre of the plink DCT-OFDA system. 1 where D is an IDCT matrix. At the receiver side, the CP is removed from the received signal and the received signal can be written as follows: where = r H ~ x + n = 1 ~ x = D 1 Td C (15) x is an 1 vector representing the block of the transmitted symbols of the th ser. Applying the DFT, the received signal can be given by: = R Λ F ~ x + N = 1 (16) where Λ is an diagonal matrix containing the DFT of the circlant seqence of H. After that, the FDE, the -point IDFT, and the DCT- OFDA demodlation operations are performed to provide the estimate of the modlated symbols as follows: xˆ T 1 = S F W R (17) where W is the FDE matrix of the th ser. Finally, the demodlation and the decoding processes are applied DST-OFDA System The strctre of the DST-SC-FDA system is similar to that of the DCT-SC-FDA system in previos sbsection. The difference is that the DCT and the IDCT blocks at the transmitter and receiver are replaced by the DST and the

5 34 Faisal Saif Al-kamali et al.: Wireless Image Transmissions over Freqency Selective Channel sing Recent OFDA Systems IDST blocks, respectively. In this paper, the interleaved DFT-OFDA is denoted by DFT-IOFDA, the localized DFT-OFDA is denoted by DFT-LOFDA, the interleaved DCT-OFDA is denoted by DCT-IOFDA, the localized DCT-OFDA is denoted by DCT-LOFDA, the interleaved DST-OFDA is denoted by DST-IOFDA, and the localized DST-OFDA is denoted by DST-LOFDA. transmitted over the coded OFDA systems with different basis fnctions, the DFT, the DCT and the DST and as well as with different sbcarriers mapping schemes and different modlation schemes. The PSNR vales of the received image are calclated for different SNR vales from 0 throgh 35 db in 5 db steps. The wireless channel sed in this paper is the vehiclar A channel. 4. Complexity Evalation At the transmitter of all systems, which is the mobile nit, the complexities of the two transmitter schemes are comparable. At the receiver, which is the base station, the complexities of the DCT-OFDA and DST-OFDA systems are slightly higher than that of the DFT-OFDA system, becase the receivers still ses the DFT and the IDFT for the one-tap freqency domain eqalizer. However, the increase in the receiver complexity in the plink is tolerable considering the advantages of the DCT-OFDA and DST-OFDA systems. Figre 3. Lena image. 5. Simlation Reslts 5.1. Simlation Parameters ATLAB Simlator is sed to examine and evalate the isse of wireless images transmission over OFDA system with different basis fnctions. In the simlated OFDA system, each ser occpies 64 sbcarriers. The total nmber of sbcarriers = 256 and the nmber of sers = 4. In each simlation, all sbcarriers are assigned among all sers according to the sbcarriers mapping method sed. Qadratre phase shift keying (QPSK) and 16 qadratre amplitde modlation (16QA) modlation schemes are sed to generate a transmitted block for each ser. SE eqalization is assmed. The channel model sed for simlations is the vehiclar A channel [11]. A convoltional code with memory length seven and octal generator polynomials (133, 171) is chosen as the channel code. The transmitted Lena image for all sers of size is shown in Figre 3. As mentioned earlier, while evalating the performance of image transmission over OFDA system with different basis fnctions, PSNR and SE metrics are considered PSNR Performance In this section, the Lena image of size has been The experimental reslts are listed in Table 1 and 2 for QPSK and 16QA modlation schemes and plotted in Figres 4 and 5, respectively. Figre 4 illstrates the relationship between PSNR and SNR when Lena image is transmitted throgh the DFT- OFDA, DST-OFDA and DCT-OFDA systems for different sbcarriers mapping schemes and the QPSK is sed. From the figre, it is observed that as SNR is increased, the PSNR increases. On the other hand Figre 5 shows the efficiency of image transmission over DFT-OFDA, DST-OFDA and DCT- OFDA with different sbcarriers mapping schemes when 16QA modlation scheme is sed. As shown in the figre, the PSNR for the three systems increases with increasing of SNR. From these two figres, it is clear that the interleaved systems give the best PSNR performance when compared with the localized systems. Frthermore, it is also noted that DST-IOFDA and DCT-IOFDA systems provide better PSNR performance than DFT-IOFDA system. ore precisely, PSNR performance of DST-IOFDA is better than DCT-IOFDA bt beyond 20 db both systems provide the same PSNR performance. Table 1. PSNR vales of the received Lena image over the DFT-OFDA, the DCT-OFDA and the DST-OFDA systems when QPSK is sed. DFT-OFDA DST-OFDA DCT-OFDA LOFDA IOFDA LOFDA IOFDA LOFDA IOFDA Inf Inf Inf Inf Inf Inf Inf Inf Inf Inf

6 American Jornal of Comptation, Commnication and Control 2018; 5(1): Table 2. PSNR vales of the received Lena image over the DFT-OFDA, the DCT-OFDA and the DST-OFDA systems when 16QA is sed. DFT-OFDA DST-OFDA DCT-OFDA LOFDA IOFDA LOFDA IOFDA LOFDA IOFDA DFT-LOFDA DFT-IOFDA DST-LOFDA DST-IOFDA DCT-LOFDA DCT-IOFDA QPSK P Figre 4. PSNR against SNR of the Lena image transmission over the DFT-OFDA, the DCT-OFDA and the DST-OFDA systems when the QPSK is sed QA DFT-LOFDA DFT-IOFDA DST-LOFDA DST-IOFDA DCT-LOFDA DCT-IOFDA P Figre 5. PSNR against SNR of the Lena image transmission over the DFT-OFDA, the DCT-OFDA and the DST-OFDA systems when the 16QA is sed.

7 36 Faisal Saif Al-kamali et al.: Wireless Image Transmissions over Freqency Selective Channel sing Recent OFDA Systems 5.3. SE Performance In this section, several experiments are condcted to test and investigate the SE performance of the received image over DFT-OFDA, DST-OFDA and DCT-OFDA for different sbcarriers mapping schemes when the QPSK and the 16QA modlation techniqes are sed. The obtained SE vales are listed in Tables 3 and 4 and plotted in Figres 6 and 7. Table 3. SE vales of the received Lena image over the DFT-OFDA, the DCT-OFDA and the DST-OFDA systems when QPSK is sed. DFT-OFDA DST-OFDA DCT-OFDA LOFDA IOFDA LOFDA IOFDA LOFDA IOFDA e e e e e e e e e e e e e e e Figre 6 shows the relation between SE and SNR when Lena image is transmitted throgh the DFT-OFDA, DST- OFDA and DCT-OFDA systems for different sbcarriers mapping schemes and QPSK modlation scheme is sed. It is observed that with increasing SNR, the SE is decreased. Table 4. SE vales of the received Lena image over the DFT-OFDA, the DCT-OFDA and the DST-OFDA systems when 16QA is sed. DFT-OFDA DST-OFDA DCT-OFDA LOFDA IOFDA LOFDA IOFDA LOFDA IOFDA e e e e e e e e e e e e e e e-007 It is clear that the transmitted image sing the DST-OFDA and DCT-OFDA systems give lower SE vales than DFT- OFDA system. oreover, it is noted that DST-IOFDA and DCT-IOFDA have significantly smaller SE than DFT- IOFDA. Same previos scenario is done when 16QA modlation scheme is sed and the reslts are shown in Figre 7. As shown in the figre, it is observed that DST-IOFDA and DCT-IOFDA have significantly smaller SE than DFT-IOFDA system QPSK DFT-LOFDA DFT-IOFDA DST-LOFDA DST-IOFDA DCT-LOFDA DCT-IOFDA SE Figre 6. SE verss SNR of the Lena image transmission over the DFT-OFDA, the DCT-OFDA and the DST-OFDA systems when the QPSK is sed.

8 American Jornal of Comptation, Commnication and Control 2018; 5(1): QA 0 DFT-LOFDA DFT-IOFDA DST-LOFDA DST-IOFDA DCT-LOFDA DCT-IOFDA -1-2 SE Figre 7. SE verss SNR of the Lena image transmission over the DFT-OFDA, the DCT-OFDA and the DST-OFDA systems when the 16QA is sed Clarity Investigation The visal qality of the reconstrcted images at SNR vale 15 db is demonstrated and the received images of one plink ser are shown in Figre 8 for different OFDA systems. By comparing these received images with the original image in Figre 3, it can conclde the speriority of the DST-OFDA and DCT-OFDA system over the conventional DFT-OFDA system for different sbcarriers mapping schemes. 6. Conclsion In this paper, wireless transmission of gray-scale images over OFDA system is tested for different basis fnctions, different sbcarriers mapping schemes and different modlation schemes. It is shown that transmitting wireless images over OFDA systems is possible for different basis fnctions. The obtained reslts show a noticeable performance improvement for the DST-OFDA and DCTOFDA systems in terms of the SE and PSNR over the conventional DFTOFDA system. On the other hand, transmitting wireless images sing the DCT or the DST as the basis fnction in the OFDA system greatly enhances the clarity of the received images than that sing the DFT as the basis fnction in the OFDA system. oreover, the interleaved scheme provides better SE and PSNR performances than the localized schemes for all systems, especially with QPSK. References [1] Figre 8. Image for different OFDA Systems at SNR = 20 db and QPSK. H. G. yng and D. J. Goodman, Single Carrier FDA: A New Air Interface for Long Term Evalation, John Wiley & Sons, Chichester, K, 2008.

9 38 Faisal Saif Al-kamali et al.: Wireless Image Transmissions over Freqency Selective Channel sing Recent OFDA Systems [2] E.. El-Bakary, E. S. Hassan, O. Zahran, S. A. El-Dolil, and F. E. Abd El-Samie, Efficient image transmission with mlticarrier CDA, Wireless Personal Commnications, Springer, vol. 69, No. 2, pp , arch [3] A. F. Al-Jnaid and F. S. Al-kamali, Efficient wireless transmission scheme based on the recent DST-C-CDA, Wireless Networks, Springer, Vol. 22, No. 3, pp , April [4] Y. Sn and Z. Xiong, Progressive image transmission over space-time coded OFD-based IO systems with adaptive modlation, IEEE Trans. On obile Compting, vol. 5, No. 8, pp , [5]. Shayegannia, A. Hajshirmohammadi, S. haidat and. Torki, Space time coding IO-OFD SAR for highresoltion imaging, IET Image Processing, vol. 7, No. 1, pp , [6] N. F. Soliman, Y. Albagory,. A. Elbendary, W. Al-Hanafy, E.. El-Rabaie, S. A. Alshebeili, and F. E. Abd El-Samie, Chaotic interleaving for robst image transmission with LDPC coded OFD, Wireless Personal Commnications, Vol. 79, No. 3, pp , Dec [7] Faisal S. Al-Kamali, Abdllah A. Qasem, Wireless Images Transmission over OFDA Systems: Investigation and Evalation, IET JOE. pp. 1-12, [8] F. S. Al-Kamali, A. A. Qasem, S. A. Abasbaa and G. A. Qasem SC-FDA and OFDA: An Efficient Wireless Image Transmission Schemes, Jornal of Control and Systems Engineering, Vol. 4, No. 1, pp , [9] P. Tan, and N. C. Bealie, A Comparison of DCT-Based OFD and DFT-Based OFD in Freqency Offset and Fading Channels, IEEE Transactions on Commnications, Vol. 54, No. 11, Nov [] accessed date 15/07/2017s. [11] 3rd generation partnership project, 3GPP TS 25.1 technical specification grop radio access network; ser eqipment (E) radio transmission and reception (FDD) (Release 7), Section B.2.2, Sep Biographies Faisal Saif Al-kamali has received the B.Sc. degree in Electronics and Commnications Engineering from the Faclty of Engineering, Baghdad niversity, Baghdad, Iraq, in He has received the.sc., and PhD degrees in Commnication Engineering from the Faclty of Electronic Engineering, enofia niversity, enof, Egypt, in 2008, and 2011, respectively. He joined the teaching staff of the Department of Electrical, Faclty of Engineering and architectre, Ibb niversity, Ibb, Yemen in He has served as a head of the Electrical department, Ibb niversity from Oct to Oct He is a co-athor of more than 30 papers in national and international conferences and jornals and one textbook. His research areas of interest inclde CDA Systems, OFDA Systems, Single Carrier FDA (SC-FDA) System, IO Systems, Interference Cancellation, Synchronization, Channel Eqalization and Channel Estimation. Khaled Abdllah Al-sofy. He received his B.Sc in Compter and control engineering from Sanaa niversity, Yemen, in He received his.tech in Compter Science and Engineering from Osmania niversity, Hyderabad, India in He received his Ph.D. in 2015 from Compter Engineering, from department of compter Engineering, Z. H. College of Engineering and Technology, Aligarh slim niversity, Aligarh, India. He has joined the teaching staff of the Department of Electrical Engineering, Faclty of Engineering and Architectre, Ibb niversity, Yemen, in His research interest incldes mobile compting and Qality of service in mobile ad hoc networks, Sensor network, wireless network, image and signal processing. Farok Abd Al-fhaidy. He received the B.S.E.E. Compter Engineering and.s.e.e. in Compter Engineering respectively and Ph.D. in Commnication Engineering from the ilitary Technical College, Cairo, Egypt in 2007, 2009 and 2013 respectively. He joined the teaching staff of the Department of Electrical Engineering, Faclty of Engineering and Architectre, Ibb niversity, Yemen, in His research interests and activities are in the areas of Wireless Sensor Networks, Embedded Systems, broadband commnications and spread-spectrm commnications, advanced signal processing, and image processing.