Performances analysis of Vehicle-to-X Communication Systems 2 l Hieu Nguyen , Van Due Nguyen , Tien Hoa Nguyen , Duyen Trung Ha , and Thomas Kaiser IUniversity of Engineering and Technology, Vietnam National University, Hanoi, Vietnam Hanoi University of Science and Technology, Hanoi, Vietnam 3Institute of Digital Signal Processing, Faculty of Engineering, University of Duisburg-Essen, Duisburg 47057, Gennany Abstract: carrier This paper investigates the influence of the inter interference on a vehicle to infrastructure (V2X) communication system, which has been defined in the IEEE 802.lIp standard The frequency shift due to high speed of cars destroys the orthogonality between subcarriers in OFDM signals and rises the inter-carrier interference (leI) The main contributions of this paper are proposed analytic formulas and simulation results of the signal to interference power ratio (SIR) to model the V2X channels based on different Doppler spread shapes This paper is organized as follows In section II, formula of and SER for four types of Doppler spread model with Rayleigh and Rician channel in the OFDM system are presented Section III provides results and discussions Finally, section IV concludes this paper SIR by considering the statistical effects of leI and evaluation of II symbol error rate (SER) for several types of fading and Doppler spread in time domain approach The results show that for the Rayleigh fading model, SIR decreases when the maximum Doppler spread increases On the other hand, in the Rician channel, SIR depends not only on maximum Doppler spread but also on the angle of arrival (AoA) and Rician factor of the LOS component Moreover, the simulation result in terms of SER is obtained for different channel models It shows an excellent agreement between the theoretical calculation of SIR and the DOPPLER SPREAD MODELS In this section, analytic formulas of evaluating two parameters for vehicles communications SIR and SER are presented for each type of Doppler spread models SIR evaluation 1) Received signal in OFDMsystem The transmitted OFDM signal is given by simulation results of SER Keyword- leI and SIR analysis for OFDM-based vehicle X n = O, ,N-I system, V2X channel models I (1) INTRODUCTION The intelligent transportation system with a support of high-speed communications has been proposed by both academic researchers and industry [7], [8], [9], [lO], [11] This system can provide the link between vehicles (V2V) or vehicle and infrastructure (V2X) The standards are based on the IEEE 802.11 wireless local area network (WLAN) in the 5.9 GHz unlicensed band with communication range of about Ikm [12], [13] In this system, two most significant impacts are the high Doppler spread due to the very high speed of the vehicle, and the existence of line-of-sight (LOS) component due to short range transmission [14] The effects of these two properties on relevant communication systems need to be analyzed Signal to interference power ratio (SIR) and symbol error rate (SER) are two very important parameters in evaluating the performance of communication systems However, these two parameters have so far not been investigated for vehicle communication system In this paper, analytic formulas and simulation results for the signal to interference power ratio and evaluation of symbol error rate are proposed with considering several types of fading and Doppler spread shape such as: Classical Jakes's model, Uniform shape model, Round shape model and Two path mode In order to evaluate SER we have to generate AoA of all paths, which are suitable to the statistical properties of AoA in real-world channel Therefore, a method is proposed, which generates these AoA We firstly generate pseudo sequence of AoA based on the probability distribution function (PDF) of AoA After that Lp-norm method (LPNM) is used where X[k] denoted the transmitted data at the klh sub-carrier, x[n] is the n ih sample of the OFDM system, and N is the number of sub-carrier of each OFDM symbol This signal is transmitted over multipath fading channels The multipath fading channel can be expressed as: L h [n ] = � > [n ,l ] o [IT , �O - 1', ] (2) where h[n, I] is the tap gain of the l 'h tap at time n, and delay of l 'h tap The received signal sampled at time index written by n 1' , is can therefore L r [n ] = � > [n ,l ] x [n -I ] + w [n ] (3) ,�O where w[n] is the complex white Gaussian noise with zero mean and variance No After Fourier transformation, the Eq (3) can be rewritten as r[n] = N-J "liN k�O 21rnk f7? L H n k X k / ;V [ , ] [ ] n ],n + w[ =O, ,N -1 (4) Let X[k] denote the demodulated data of klh sub-carrier, than 978-1-4673-5604-6/12/$3l.00 ©2012 IEEE 000326 N-I _/7rNnk ,k=O, X [k]= h7 k�OLr[n]e A .,N-l "N (5) Plugging Eq (1) and (4) to (5) we have that X[k]= ��[(�H[n,m{7r(I�-k)n j X[m]]+W[k] =S[k]+I[k]+W[k] where S[k]=� III��OH[n, k]x [k] 27r(m-k)n N-I N-I i[k]=- m=O,Lm::t;.k n=OLH[n,m]/-N-X [m] W[k]= JN�N-I w[n]e_/lfNnk N N (6) (13) Replacing the autocorrelation function of several Doppler spread spectral shapes to (12) and (13) we can determine the Signal to Interference Ratio (SIR) for each case Following will only present the expressions for relevant and in two-path model as theorem without roof The roof will be presented in another pUblication PS Pi a) Rayleigh fading Ps (7) PI (8) (9) [ ( -lpl)RY (pT) =-4 N p�I-N3 N 2lf(m-k) N-I N-I D (N -lpl)RY (pT�J N P =-40 N m�O p�I-N (14) mOk Thereby: Let denote PSI Pi the power of desired signal and the interference signal at klh sub-carrier, we have that P , =E t BealE' � = E[II[kfIJ· To obtain PI' we use following equation: Suppose that the Fourier transform of the channel impulse response at different frequencies is independent of each other, i.e., where EFf!(n,.,ml)H*(n2,m2)FYR(n,.-m2)O(ml-m2) R((n, - m2)T) Rician fading b) is the autocorrelation of the Doppler spread of the channel at time instants n,Tand n2Trespectively Suppose also that the signal at different sub-carriers is uncorrelated and has unit power i.e., E[¥(nl'ml)X(n2, Denote Y RI(r) and R (r) the autocorrelation of the Doppler spreads for the Rayleigh channel and the Rician channel, respectively Then the power of the interference can be rewritten as: m2)FY 5(m, - m2) Thereby: N-I I(N-lpl)RI(pT) SIR P, _ '-p=I_-N , , , ,z(m-k) P N-I N-I N p I II(N-lpl)RI(pT�Jm=O m# p=l-N = = l KxI I (N_lpl�j2Z(J,pTW'O"+��k} I I (N_lpI)RY(PT/tk p -�� m:tk + -�� m:tk By the same analysis as in the previous part, the powers of signal and interference are reduced to: Eq (11) can be rewritten as: PI -k)p ( =� - l p l ) R Y (pT�j lf ; (N [3 [3 N m�O p�I-N mOk (12) Repeating the same procedure as in deriving Eq (12), we have We see that the SIR depends on the autocorrelation of Doppler spread We now obtain the autocorrelation of Doppler spread for each type of channel model 000327 2) Autocorrelation function for four models a) Classical Jakes's model (J)= � r7f"Y ' If I Jr d 1v (fJ ;; < 1) Generation of AoA based on LPNMmethod In this section, Lp - norm method (LPNM) which generates an will be discussed A stochastic continuous-time simulation h (19) (20) b) Uniform shape model {� (21) (22) Round shape model P,(f)� {l-(;']' ��� (23) ;" r R� (r)= J1 ( Jrh ) 2Jrhr Two -path model Pr(J)= p.t5(J + h)+(I-p)t5(J -h) R!; (r) = p.e- j27l!d' + (1- p).ej27l!d' (24) (25) (26) SER evaluation In this section, the rice fading process for four Doppler spread models will be presented SER results will be obtained for each model by simulating with QAM-16 OFDM system The baseband Rician fading process is given by [2]: Doppler frequency The initial values for an of four models are proposed in the next section The stochastic time process v(t) can be expressed as: N Where OJn = In, cos(an ) is the diffuse Doppler with random ()n � v(t)= D cn cos(2Jrfd cos(an)t +()n) n=1 (29) R(JJr)=E {v(r)V(r-HJr)} (30) an is finally obtained such that the Lp-Norm reach the local D E ;P) o.-.!.= �rmax an p s (a) for path n, ps(a)is given in (32) and (33) U( -Jr, Jr) is the random diffuse phase G is average � fading signal power [2] K is the ratio of specula power to the diffuse power and K is positive Drmax D D' P IR (0 r) - R (D rf d (0 r)D � ,p L N + (31) R(D is given at (20), (22), (24), (26) and R(O r)is given at (30) Asswning that among interval [0, Llrmax] the R(JJ r)= R(JJ r) In this case, Llrmax is set 0.08s [1] 2) initial values of AoA With The initial values of an of Classical Jakes's model are given by the following [4]: an =Jr(n-1I2)/2n (32) For Two-path model, these are only two paths We simply seek the angle of arrivals of these paths as Error! Reference source not found al,2 =[Jr I 90, Jr I 45] (33) On the other hand, we have only the PDF of AoA listing as followings: / / AoA denotes the maximum minimwn by using numerical optimization method Where fd is the maximwn Doppler frequency, p presents the normalized power of one path, (J denotes the Dirac function, P(j) is the power spectral density, R (r) is the autocorrelation function and Jo, Jj compute the Bessel function of the order and order 1, respectively B fd The time correlation of v(t) defined as [17] 'I , (28) Where Cn denoted channel gain, frequencies In are non-zero real-valued constant quantities and the phase en are independent, identically distributed random variables, each having a uniform distribution over the interval (0,2Jr) Let denote In = h cos(an ) , where lf l Pu (J)= / < h o ,otherwise R� (r) = sinc (J/Z") d) N v(t)= D cn cos( 2Jrfnt +()n) n=l , otherwise o c) for a lognormal process can be modeled by a sum of sinusoids [1]: Uniform shape model AOA [2J: Pu (a)= sin (a) I 4, lal � Jr Round shape model AOA[2j: Pr (a)= sin2 (a) I 4, lal � Jr (34) (35) With these models which not have a , we will generate sequence an by generating corresponding pseudo sequences using the correspondingp(a) 000328 " 'l:l\"'; t\ + :: \ j ·-f - · '· ;····i ·· ·· i ·· ; ··· · · I � � _ + ·· + l! + ,··········i + j + · + · · + ··· ··ii ··· · ! i ··f·· lt i ····i·�:···i······· · i····+···"····+ ····i ········1········· n ·,,··········'/""'·· ······'".- .; · i ,"-" , , ,,'""',,,., + �i 1' :\ I , i \' i ! • i : [� - � !i r\j V \1 �f\)N rv � V " , 1\ , , �I Fig , , ,., - 110 O!il!i ·· · ;"'''''= ' Fig 1: The signal to interference ratio for the Rayleigh and Rician fading channels ,'.'f t IH t Ht Ht H I H I 11 11 1+1 11 ! " t·;· t · j ;-I h· t 'j ; -I HI- HI-· H·j " f/ ++H +H H · ; ·H Ht ,, ·tt ,,· tt ,!- -7, 7,�, c,; �-! t ! �1 •• I 3: Time Correlation of four models Fig illustrates time correlation of four models after generating parameters In the figure, the simulation model has nearly the same statistics as the reference model III SIMULAnON RESULTS AND DISCUSSION Here we consider the parameters for the WAVE IEEE 802.11P standard This system operates at center frequency of 5.9 GHz Some typical parameters are listed in Table I Table Standard 802.1lQ 802.lIa Carrier frequency 5.9 GHz 2.4 GHz Bandwidth 10MHz 20MHz Symbol duration 8f.1.s 4f.1.s Guard Time 1.6f.1.s 0.8f.1.s Number ofFFT])oint 64 64 Number ofsub-carrier 52 52 48 48 15625MHz 0.3125MHz Number ofuseflll subcarrier Frequency spacing SIR Fig illustrates the effect of maximum Doppler spread.fd on SIR for four kinds of Doppler spread The maximum Doppler spread is proportional to the speed of a vehicle If the speed of the vehicle increases, the fd increases, the SIR decreases too For example, if the velocity of the vehicle is v=300kmlh, the carrier frequency is fc = 5.9GHz and the speed of light is c=3.108m1s then the maximum Doppler frequency is determined by: K with different AoA eo for the Rician fadin� channels Jd 1: Parameters of Wave 802.llp Parameters Fig 2: The signal to interference ratio versus vfc = -;;- = 300x103 x5.9x109 3600x 3x108 The symbol duration of OFDM symbol in WAVE standard is Ts = 81ls and hence JdT "" 1.31x10-2 • When the Rician factor increases to K=200 and the AoA of LOS component is eo=nI2, the SIR of the Rician channel and Rayleigh channel are show in Figure 5, where the solid curves are for the Rayleigh and the dash curves are for the Rician fading At this AoA of the LOS, the specular component improves the SIR of the system significantly, e.g the SIR is improved about 25 dB for the Doppler spread of two-path model This result agrees with the intuition that the LOS component usually maintains the quality of transmission The different results are illustrated in detail in Fig 3, where the maximum Doppler spread is set at.fd= 1000Hz and AoA of LOS component are Bo=O; 0.457r, 0.57r respectively In figure 4, at the same value of the lowest curve is the SIR for the two-path model and the highest curve is the SIR for uniform model In the two-path model, the SIR is constant when eo=o and increases with K monotonously when eo takes other values In other three model, i.e classical Jakes, round and uniform, at Bo=O, if K increases the SIR increases to a saturation level corresponding to that of the two-path model At this situation, the LOS component destructs the quality of the transmission link If Bo=0.457r, and Bo=0.57r, the SIR increases with K monotonously There is a critical value of 000329 VI [I] M Pittzol and Van Duc Nguyen "A Spatial Simulation Model for Shadow Fading Processes In Mobile Radio Channels" Personal, indoor and Mobile Radio Communications, 2004 PIMRC 2004 15th IEEE International Symposium on1 Clerk Maxwell, A Treatise on Electricity and Magnetism, 3rd ed., vol Oxford: Clarendon, 1892, pp.68-73 [2] Jared Dulmage and Dr Michel P.Fitz "Non-Isotropic Fading Channel Model for the Highway Enviroment" Vehicular Technology Conference, 2007 VTC-2007 Fall 2007 iEEE 66thK Elissa, 'Title of paper if known," unpublished [3] R H Clarke" A statiscal theory of mobile-radio reception" Bell System Technical Journal, pp, 975-1000, Jul 1968 [4] Matthias Piltzol, Ulrich Kilat, Frank Laue and Yingchun Li "On the problems of monte carlo method based simulation models for mobile radio channels" Spread Spectrum Techniques and Applications Proceedings, 1996., IEEE 4th International Symposium on [5] G Acosta- Marum and M.A ingram"Doubly selective vehicle-to-veicle channel Measurements and modeling at 5.9 GHz" in Wireless Personal Multimedia Communication(WPMC) 2006, 17-20 September 2006 [6] M Patzol, Ulrich Kilat, Frank Laue and Yingchun Li,"On the statistical properties of deterministic simulation models for mobile fading channels" iEEE Trans, Veh Technol., vol 47, no 1, pp 254-269, Feb.199 [7] L Cheng, B E Henty, D D Stancil, F Bai, and P Mudalige, "Mobile Vehicle-to-Vehicle Narrow-Band Channel Measurement and Characterization of the 5.9 GHz Dedicated Short Range Communication (DSRC) Frequency Band" iEEE J Sel Areas Commun., vol 25., no 8, pp 1501-1516-2007 [8] A Paier, Karedal, N Czink, H Hofstetter, C Dumard, T.Zemen, F.Tufvesson, A F Molisch, and C F Mecklenbrauker, "Car-to-car radio channel measurements at 5GHz: Pathloss, power-delay-profile, and delay-Doppler spectrum" in Proc 4'h into Symp Wireless Communication Systems iSWCS 2007, 2007, pp 224-228 [9] "CAR CAR Communication Consortium Manifesto," CAR CAR Communication Consortium, Tech Rep, Aug 2007 [Online] Avaiable: http://www.car2car.org DO W if> -30� -: -7 -o -, -!: :': , :'!o62 :':41 ,':': 2: :- :';; -: -; SNRdb Fig 4: The symbol error rate for Rician fading channels ' AoA, denotes as 90 , at this point the SIR for each one of the three models of the Doppler spread is constant If AoA is ' greater than 90 , the SIR increases with K otherwise, the SIR ' decreases with K if the AoA is less than 90 This critical value ' 90 is determined by solving the equation that the fIrst derivative of SIR with respect to K is set to zero This equation is quite complex and usually solved by numerical method SER Fig shows the SER of Rician channels Where the Rician factor K=5, AoA of LOS component is 80=7[12 and Power delay profIle is given in [5] When SNR increases, SER decreases In case of SNR < 10 dB, SERs of four models have no signifIcant difference In contrast, when the SNR> 10 dB, SER of Uniform model decreases deeply so the quality of system increases considerably; with Two-path model, SER of it also decreases but slowly Hence, quality of Uniform model is much greater than Two-path model IV CONCLUSION The communication link in intelligent transportation systems has been based on WLAN IEEE 802.11P standard It uses OFDM as a basic modulation scheme for the communication systems If the very high speed causes high Doppler spread, the orthogonality among sub-carriers is destroyed On the other hand, in DSRC system, the existing of LOS and the Rician fading channel model should be considered due to the short range of transmission link In this paper, the effect of Doppler spread to OFDM signals are studied for both SIR and BER performance with four well known Doppler spread models (two-path, classical Jakes, round and uniform model) When the speed of vehicles increased, the maximum Doppler frequency is also increased, thus making the decreasing of SIR Besides, a different result illustrates that the considerable differences of symbol error rate (SER) or quality of four models when SNR of system increases The article only evaluates the performance of OFDM-based Vehicle-to-X communication systems As future works, we plan to study such communication system considering Vehicle-to-Vehicle scenarios V REFERENCES [10] Karedal, F.Tufvesson, N Czink, A Paier, C Dumard, T Zemen, C F Mecklenbrauker, and A F Molisch, "A geometry- based stochastic MIMO model for vehicle-to-vehicle communications" iEEE Trans Wireless Commun., vol 8, no 7, pp 3646-3657, 2009 [11] A G Zaijic and G L Stuber, "Three-dimensional modeling and simulation of wideband MIMO mobile-to-mobile channels" iEEE Trans Wireless Commun., vol 8, no 3, pp 1260-1275, 2009 [12] "Part 11: Wireless LAN Medium Access Control (MAC) and Physical Layer (PHy) specifications; Amendment 7: Wireless Access in Vehicular Environments," IEEE, Draft Standard, Nov 2008 [13] "Intelligent Transportation Systems (ITS); Vehicular Commnunication; Basic Set f Applications; Definition," ETSI, Draft Standard, Jan 2009 [14] G Acosta- Marum and M A Ingram, "Six Time and Frequency Selective Empirical Channel Models for Vehicular Wireless LANs" in Proc VTC-2007 Fall Vehicular Technology Con] 2007 IEEE 66th, 2007, pp 2134-2138 [15] Li and M Kavehrad, "Effects of time selective multipath fading on OFDM systems for broadband mobile applications," lEE Commun Lett, vol 3, no 12,pp 332-324, Dec 1999 [16] Y.Li and J.Cimini, LJ.,"Bounds on the interchannel interference of OFDM in time-varying impairments," lEEE Trans Commun, vo1.49, no.3, pp 401-404,2001 [17] M Paetzold, U Killat, F Laue, and Y Li, "On the statistical properties of deterministic simulation models for mobile fading channels," iEEE Trans Veh Technol., vol 47, no I, pp 254-269, Feb 199 [18] M Patzol"Mobile radio channels", second edition, John Wiley & Sons, 2011 ACKNOWLEDGEMENT This research is funded by Vietnam National Foundation for Science and Technology Development (NAFOSTED) under grant number 102.02-2011.15 000330 ... differences of symbol error rate (SER) or quality of four models when SNR of system increases The article only evaluates the performance of OFDM-based Vehicle-to-X communication systems As future works,... Two-path model IV CONCLUSION The communication link in intelligent transportation systems has been based on WLAN IEEE 802.11P standard It uses OFDM as a basic modulation scheme for the communication. .. and uniform, at Bo=O, if K increases the SIR increases to a saturation level corresponding to that of the two-path model At this situation, the LOS component destructs the quality of the transmission