Comparative Analysis of CDMA Based Wireless Communication under Radio Propagation Environment

Transcription

1 Comparative Analysis of CDMA Based Wireless Communication under Radio Propagation Environment Md. Rezaul Hoque Kan Department of EEE, CUET, Cittagong- 4349, Banglades. A H M Razibul Islam Department of Radio Communication Engineering, Kyung Hee University, Sout Korea. Razib32@gmail.com Jainul Abedin IIUC,Cittagong, Banglades Jainul_abedin2@ya oo.com Ju Bin Song Department of Radio Communication Engineering, Kyung Hee University, Sout Korea. jsong@ku.ac.kr Abstract Knowledge of te propagation caracteristics of a mobile radio cannel is essential to te understanding and design of a cellular system [1]. An appropriate propagation model is required wen estimating te link budget or designing a Code Division Multiple Access (CDMA) system [2]. Tis paper deals wit comparative parametric analysis for propagation pat loss considering macro cell region using different models and contains comparative study wit real measurement obtained from Pacific Banglades Telecom imited (PBT), a CDMA based wireless network for city Daka, Banglades. 1. Introduction. Te pioneering of CDMA in cellular radio is due to Qualcomm Inc. Starting in te late 198s, Qualcomm embarked on a series of experiments tat culminated in demonstrating tat CDMA ad te potential to provide an efficient radio interface for use in cellular networks. By July 1993, Qualcomm proposals were adopted by te Telecommunications Industry Association as Interim Standard 95 (IS-95) [3]. Wireless cellular communication is witnessing a rapid growt in markets, tecnology, and range of services. A major current trust for cellular communication systems is improved economics troug enanced coverage early in te life cycle of a network and ig spectrum efficiency later in te life cycle. An attractive approac for economical, spectrally efficient, and ig quality digital cellular and personal communication services (PCS) is te use of spread spectrum modulation wit code division multiple access (CDMA) tecnology [4]. In a cellular system reusing eac frequency at several regions of service area increases te capacity. Also cell splitting, sectoring and micro cell zone approaces are used to expand te capacity of cellular system. Consideration must be taken into account to keep te interference at acceptable limit. Macro cells correspond to cells were te base station is placed on top of tall buildings or towers and transmits enoug power to cover several miles. Te pysical propagation environment in macro cells is caracterized by near-ground irregularities (buildings, terrains, etc). Te propagation pat lengt can be up to several miles. Macro cells can be classified into different cannel types: urban, suburban, and rural. One of te most significant issues of interference into microwave systems is a line of sigt situation [5]. Te most common occurrence of tis will be from a subscriber unit located in a ig-rise building or on a balcony. In tis case, pat loss figures approacing free space loss may be experienced between te subscriber unit and microwave antennas. It is possible for tis situation tat te subscriber unit s interfering signal will be stronger tan te aggregated powers of many base station transmissions at te microwave receiver. In urban environments, te probability of an elevated subscriber unit is greater. Tus, te impact of te subscriber unit interference sources on te microwave receiver will be more substantial tan in residential areas. Te performance of wireless systems is affected by a number of propagation penomena [6]: 1) Pat-loss variation versus distance; 2) Random slow sadowing; 3) Random multipat fading; 4) Inter-Symbol Interference (ISI), co-cannel interference as well as multiuser interference; 5) Background noise.

2 2. Propagation pat oss A measure of interest in radio propagation is te pat loss, wic is defined as te ratio between te received power Pr and te transmitted power Pt d Pr = P t Propagation models are used to determine ow many cell sites are required to provide te coverage requirement for te networks. Initial network design typically is engineered for coverage. ater on network growt is based on capacity. Te propagation model elps to determine were te cell sites sould be located to acieve an optimal position in te network. If te propagation model used is not effective in placing cell sites correctly, te probability of incorrectly deploying a cell sites into te network is ig. Te propagation model is also used in oter system performance aspects including andover optimization, power level adjustments and antenna placements [6]. Predictions of signal strengt and propagation coverage area are vital aspects in te design of wireless communication systems. Tere are several metods for finding te propagation loss as follows: Hata-Okumura Walfisc-Ikegami Bullington Elgi Epstien-Peterson ongley-rice Propagation models can de described to two distant classes: deterministic and stocastic. Te deterministic model is useful wen multipat is caused by a large number of pats between te transmitter and receiver. Penomena like multipat propagation, reflection, diffraction and sadowing ave a significant influence on te received power. So te propagation models sould consider tese penomena to obtain accurate results. In te present study we ave considered following models for Daka city, Banglades. 2.1 Hata-okumura propagation model Among te many tecnical reports tat are concerned wit propagation prediction metods for mobile radio, Okumura s [7] report is believed to be te most compreensive one. In is report, many useful curves to predict a median value of te received signal strengt are presented based on te data collected in te Tokyo area. Te Tokyo urban area was ten used as a basic predictor for urban areas. Te correction factors for suburban and open areas are determined based on te transmit frequency. Based on Okumura s prediction curves, empirical formulae for te median pat loss, p, between two isotropic antennae were obtained by Hata and are known as te Hata Empirical Formulae for Pat oss [8]. Te Hata propagation formulae are used wit te link budget calculation to translate a pat loss value to a forward link cell radius and a reverse link cell radius. Hata model illustrate a sligtly more complicated pat loss model tat s a function of parameters suc as frequency, frequency range, eigts of transmitter and receiver, and building density. Te Hata model is based on extensive empirical measurements taken in urban environments. In its decibel form, te generalized model can be written as = K K log( f ) log( ) + a( p 1 2 b m [ og( b )]log( d) K (1.1) Were f is te carrier frequency (in megaertz), b is te antenna eigt (in meters) of te base station, m is te mobile antenna eigt (in meters), d is te distance (in kilometers) between te base station and te mobile user. For tese parameters, tere are only certain ranges in wic te model is valid; tat is, b sould only be between 3m to 2m, m sould be between 1m to 1m, and d sould be between 1 km to 2 km. Note tat te slope of equation (1.1) is og( )] db/decade. [ b Te terms a( m ) and K o are used to account for weter te propagation takes place in an urban or a dense urban environment. In particular, a( m ) = [1.1 log (f).7] m [1.56 log (f).8] for urban or a( m ) = 3.2[log (11.75 m )] for dense urban and K = for urban, or K =3dB for denseurban Te term K1 and te factor K 2 are used to account for te frequency ranges. Specifically, K 1 =69.55 for frequency range 15 f 1 MHz, or K 1 =46.3 for frequency range 15 f 2 MHz And K 2 =26.16 for frequency range 15 f 1 MHz, or K 2 =33.9 for frequency range 15 f 2 MHz According to Hata model te pat loss is expressed as, P = log (f)-13.82log b -(1.1 log f-.7) m +(1.56 log( f)-.8)+( log b )logd (1.2) Modification for suburban city P M = P (Urban)-2[log (f/28)] (1.3) )

3 2.2 Walfisc-Ikegami propagation model Tis empirical model is a combination of te models from J. Walfisc and F. Ikegami. It was furter developed by te COST 231 project. It is now called Empirical COST-Walfisc-Ikegami Model [9]. Te model considers te buildings in te vertical plane between te transmitter and te receiver. Te accuracy of tis empirical model is quite ig because in urban environments especially te propagation over te tops (multiple diffractions) is te most dominant part. If te scenario is analyzed individually for eac receiver pixel (parameters of building data are determined depending on te actual buildings between Tx and Rx) te accuracy is ig - only wave guiding effects due to multiple reflections are not considered. Te main parameters of te model are: Frequency f (8...2 MHz) Heigt of te transmitter (4...5 m) Heigt of te receiver RX (1...3 m) Distance d between transmitter and receiver (2...5 m) Parameters depending on te buildings in te vertical plane between transmitter and receiver: Mean value of building eigts ROOF Mean value of widts of streets w Mean value of building separation b Te model distinguises between two situations, te "line of sigt" (OS) and te "none line of sigt" (NOS) situation. OS situation: For te OS-case te prediction is very easy, as only one equation wit two parameters is necessary. = logf+2logd Tis OS equation is similar to te free space loss equation. It was modified after evaluating measurements in European cities. If te distance is d = 2 m, te loss is almost equal to te free space loss at te same distance. NOS situation: Te NOS equations are more complicated. Te loss in te NOS case is te sum of te free space loss, te multiple screen diffraction loss msd and te top-tostreet diffraction loss rts : p = + rts + msd wen wen rts rts + + msd msd > Te free space loss: = logf+2logd (1.4) Te top-to-street diffraction loss term l rts determines te loss tat occurs on te wave coupling into te street were te receiver is located. Te origin of tis loss comes from te Ikegami model, but COST 231 as extended tis equation (1.4). rts = logw+1logf+2log RX + ori Wit on ϕ for ϕ < 35 = ( ϕ 35) for 35 ϕ < ( ϕ 35) for 55 ϕ < 9 Te widt of te roads w, te top eigt ROOF, te receiver eigt RX and te road orientation φ are te parameters in tis equation. Te orientation loss ori is an empirical correction term obtained from te calibration wit measurements. An approximation for te multi-screen diffraction loss was publised by Walfisc and Bartoni. COST 231 modified tis approximation to be used also for base station antenna eigts below top level. Te building eigts ROOF and te building separation b are taken into account additionally: msd = bs +K a +K d logd+k f logf 9logb Wit K K bs a d K d 18(1 + = 54 = 54-.8( 54-.5( = 18 15( ) wen > < wen ) wen > d.5kmand ) wend /.5, d <.5Kmand 18 ) /( RX ) > <.7( f / 925 1) suburban centers = ( f / 925 1) for metropolit an centers

4 Te factors k d and k f control te dependence of te multi-screen diffraction loss versus te distance and te radio frequency. Te factor k a indicates te increase of te pat loss for base stations below te top. Te final expression of Pat loss using Walfisc- Ikegami model for patloss calculation becomes: p =32.4+2log(d)+2log(f) log(w)+1log(f)+2log( - rx ) (1+( tx - ))+( tx - )+54+18log(d)-[4+.7(f/925-1)]log(f)- 9log(b) 3. Simulations and results.. (1.5) Te parameters for te city Daka, Banglades for it s CDMA based system as been taken from te Pacific Banglades Telecom td (PBT). It is single cell, tree sectors, 35-cannel/sector system considering 3% andoff rate. Tese parameters were used to evaluate te pat loss for te signal [1, 11]. Te propagation model can be used to calculate te amount of signal received by a mobile. Te calculation is performed based on a BTS antenna of ANDREW (Model no CTSD D). et, BTS transmitter power BTS antenna eigt P T =5 W= 57 dbm MS P X power P T =23 dbm = b =3 m MS antenna eigt = m =1.5 m T X antenna gain G T =17 dbi MS R X sensitivity = - 16 dbm MS Antenna Gain G R = BTS R X sensitivity = dbm Down link transmit frequency= MHz Up link transmit frequency = MHz P = Pat loss D= distance from BTS to MS P R = Received power UP link (MS to BTS), received power is sown in table 1 Table 1: MS to BTS received power D(k m) P (d P T (d Bm) G T (d G R (d P R =P T + G T +G R - P (d Minimum allowable received power at BTS in CDMA -121 dbm Down link (BTS to MS), received power is sown in table 2 D( km ) Table 2: BTS to MS received power P (d Distance (Km) P T (d Bm) G T (d G R (d P R =P T +G T + G R -P (d Minimum requirement of CDMA at MS -63 to -16 dbm Table 3: Comparison of Hata s Model, Walfisc s Model wit practical value Pat loss (d Hata Modified Walfisc s Model Hata Model Practical value Model Figure 1: Pat loss Vs log of distance for Hata- Okumura model Figure 2: Pat loss comparison between Modified Hata model and Practical data

5 comparison table 3. Te results are closer to Walfisc s Model as compared to Hata Model s. So in order to maximize te spectral efficiency of a cellular system, it is advisable to use Walfisc s Model wit appropriate factor as computed in te present study. Te computation time for Walfisc s Model is a bit long as it depends upon several variable owever terrain profiles as been neglected. 4. Conclusion Figure 3: Pat loss Vs log of distance for Walfisc-Ikegami model Figure 4: Pat loss comparison between Walfisc-Ikegami model and Practical data Figure 1 sows te variation in pat loss as te distance in logaritmic form is varied. Te correction factor was also calculated using equation (1.3) for varying pat loss of (P M ) for te carrier frequency (824 MHz). Te values of te correction factor are as follows: P M = db at P = db P M = db at P = 27. db P M = db at P = db P M =19.28 db at P = db P M = dB at P = db Modified Hata value vs practical data of pat loss ave been sown in figure 2. Similarly, pat losses were calculated using Walfisc-Ikagami Model for te Daka city using parameters as obtained from Pacific Banglades Telecom imited (PBT). Table 1 and 2 sows tat te pat loss were found to be increase as te distance is increased from 1 to 3 Km. A comparison of pat loss using Modified Hata and Walfisc-Ikegami models for teir standard value wit real measurement is sown in figure 2 and figure 4 respectively. Te result for Daka city are on te line of result of above models as given in te above Different pat loss sources (statistical, deterministic, real world) may be used by te simulator to aid in defining te CDMA coverage area. Eac pat loss type as its benefits and disadvantages. Wile Hata, Power aw and Micro Cell models do not consider terrain variation, tey do allow for simulation in areas were digitized terrain or overlay databases do not exist. Deterministic propagation models incorporate terrain variation, antenna data, overlay (clutter) data, etc. in an attempt to model actual installations. Tis predictive modeling can be performed for a large area. Measured street data incorporates terrain variation and actual installed antennas, but is limited in te area covered. A ognormal-fading overlay can also be included to account for te effects of sadowing on te system being modeled. Te present study as presented te comparative parametric analysis of different pat loss models for CDMA based mobile communication. Te study as impact on te design of radio frequency planning of CDMA based system. 5. Acknowledgement Autor is tankful to all personnel s of Pacific Banglades Telecom imited (PBT) at Daka for teir valuable tecnical suggestion. 6. Reference [1] Hess, G. C., and-mobile Radio System Engineering, Norwood, MA: Artec House, [2] ee, W. C. Y., Mobile Communications Design Fundamentals, New York, NY: Jon Wiley & Sons, [3] TIA/EIA Interim standard-95, Mobile station base station compatibility standard for dual-mode wideband spread spectrum cellular systems, July [4] ee, W. C. Y., Mobile Cellular Telecommunications: Analog and Digital Systems, New York, NY: McGraw-Hill, [5] Jakes, W. C., Microwave Mobile Communications, New York, NY: Jon Wiley & Sons, 1974.

6 [6] Merotra, A., Cellular Radio Performance Engineering, Norwood, MA: Artec House, [7] Okumura, Y., Omori, E., Kawano, T., Fukada, K,"Field strengt and ITs Variability in VHF and UHF and-mobile Radio Service", Rev. Elec. Commun. ab., 16 (1968), pp [8] Hata, M., Empirical Formula for Propagation oss in and Mobile Radio Services, IEEE Trans. on Veicular Tecnology, Volume. VT-29, 198, pp [9] [1] Samir Karmaker, Planning and design of CDMA based cellular network for daka city, Msc. Tesis submitted to te department of Electrical & Electronic Engineering, Banglades University of Engineering & Tecnology, June 22. [11] Muammad Rezaul Hoque Kan, Measurement of Signal Strengt in CDMA by Simulation of Power osses Estimation, Asian Journal of Information Tecnology (AJIT), Volume 4 Number 6, 25 pp