Rain Attenuation Effects on 2.6 GHz WiMAX Networks Deployment in Ghana

Transcription

1 World Journal of Engineering and Technolog, 015, 3, Published Online August 015 in Scies. ain Attenuation Effects on.6 GHz WiMAX Networks Deploment in Ghana Patrick Fiati Department of Electrical, Electronic Engineering, Kwame Nkrumah Universit of Science and Technolog, Kumasi, Ghana eceived 7 Ma 015; accepted 5 Jul 015; published 8 Jul 015 Copright 015 b author and Scientific esearch Publishing Inc. This work is licensed under the Creative Commons Attribution International License (CC BY). Abstract WiMAX communication sstems operating at.6 G frequencies are used for broadband multimedia and internet based services. At these frequencies, the signal will be affected b various propagation impairments such as rain attenuation, cloud attenuation, tropospheric scintillation, ionospheric scintillation, water vapour attenuation, and rain and ice depolarization. Among all the propagation impairments, rain attenuation is the most important and critical parameter. In this research, rain attenuation is calculated at KNUST, Kumasi using ITU- rain attenuation model. The preliminar results of the work will be used to calculate the attenuation experimentall and comparison can be made, which helps to develop a new rain attenuation model at.6 G bands. ain attenuation is an important aspect of signal propagation above.6 GHz frequenc. The attenuation time series generation from point rain rate measurement is crucial due to unavailabilit of actual signal measurements. In this research, a simple and realistic approach has been demonstrated for better estimation of rain attenuation using WiMAX-band signal propagation data and ground rain rate measurements in Ghana. The ITU- model of rain attenuation has been modified b incorporating an effective slant path model. The effective slant path has been estimated and modeled in terms of a power-law relationship of rain rate data of The methodolog has been validated with the measured data of 014. Comparison with ITU- and GMET clearl demonstrates the improved predictabilit of the proposed model at the present tropical location. Kewords ain Attenuation, ITU- Model, ain Fall ate, WiMAX, Alamouti 1. Introduction Atmospheric effects pla a major role in the design of satellite-to-earth links operating at frequencies above.6 How to cite this paper: Fiati, P. (015) ain Attenuation Effects on.6 GHz WiMAX Networks Deploment in Ghana. World Journal of Engineering and Technolog, 3,

2 GHz. aindrops absorb and scatter radio waves, leading to signal attenuation and reduction of the sstem availabilit and reliabilit. The severit of rain impairment increases with frequenc and varies with regional locations [1]. Hence the incidence of rainfall on radio links becomes even more important for frequencies as low as about 7 GHz particularl in the tropical and equatorial climates, where intense rainfall events are common []. It is therefore ver important when planning both microwave and terrestrial line-of-sight sstem links; to make an accurate prediction of rain induced attenuation on propagation paths [3]. Initiall, attenuation prediction attempts involved extrapolation of measurements to other locations, frequencies, and elevation angles; however, the complex nature and regional variabilit of rain make this approach highl inaccurate [4]. The method for the prediction of rain attenuation on microwave paths has been grouped into two classes: the empirical method which is based on measurement databases from stations in different climatic zones within a given region and the phsical method which make an attempt to reproduce the phsical behaviour involved in the attenuation process. However, when a phsical approach is used not all the input parameters needed for the analsis is available. Empirical method is therefore the most used methodologies [5] [6]. For the empirical methodolog, an appropriate distribution of rainfall rate at 1-minute integration time is needed for the site under studied in order to predict accurate rain attenuation for the location. This input is sometime provided b meteorological and environmental agencies, universities, and independent researchers. Stud has revealed that dail rainfall accumulations are universall recorded and hourl data are fairl available b national weather bureaus/environmental agencies [7]. There is still dearth of rainfall rate of 1-minute integration time necessar for the stud of rain induced impairment to telecommunication especiall in the tropical region (Ghana). This is because global national weather services are established to satisf more traditional requirements such as those for agriculture, hdrolog and forest management. A method for converting the available rain rate data to the equivalent 1-minute rain rate cumulative distribution is therefore necessar. The critical role of the propagation impairment on communication sstems cum lack of rain-measurement data from tropical regions for verification for modeling purposes has been the concern of man organizations like, the International Telecommunication Union (ITU), European Space Agenc (ESA), and European cooperative program (COST) among others. This has become necessar because of the peculiarit of the tropical regions, which are characterized b high intensit rainfall, enhanced frequenc of rain occurrence and the increased presence of large raindrops when compared with temperate climates [8]. Another ver important effort towards gathering more information is through Tropical ain Measurement Mission (TMM) jointl developed b the United States and Japan, and the Global Precipitation Climatolog Project (GPCP) of the World Climatic esearch Programme (WCP). As earlier stated, the data available from this mission can not directl be emploed in sstem design, due to its long integration time. The aim of this research is to give additional tools to the sstem designers, in the form of contour maps of rain intensit and rain attenuation, for the design of WiMAX sstems in the tropical countries particularl in Ghana. ain-rate and rain attenuation maps for the countr of Ghana were developed using the models purposel designed for tropical zones b Moupfouma and Martins [9] (which is a mix between a log-normal distribution for low rain rates and a gamma distribution for high rain rates) and that of J. Chebils [10] model for the estimation of point rain rate, and the ITU model for rain attenuation prediction method prediction method [11]. The climatic mapping of rain-rate and rain-attenuation has naturall attracted a great deal of attention for instance this kind of work has earlier been carried out for USA [1], Europe, Malasia, Colombia and on global scale b ITU [13], Salonen and Baptisa, Crane, Efforts has also been made b Fiati et al. to obtain 1 minute rain rate map for Ghana using ice-holmberg model however the model overestimates rain rates in the high-availabilit range (%), and underestimates in the range between 0.1% to 1%. These percentages unavailabilit of time are crucial for communication purposes, hence the need for this work.. Analsis ecent analsis suggests that the rain rate distribution is better described b a model which approximates a lognormal distribution at the low rates, and a gamma distribution at high rain rate. This kind of model was developed b Moupfouma and Martins [14]. Thus, the Moupfouma model requires three parameters; λ, γ and. The first two parameters have been provided. To estimate, the use of J. Chebils model [15] appears suitable, 84

3 it allows the usage of long-time mean annual accumulation, M, at the location of interest. The power law relationship of the model is given b = αmβ (4) where α and β are regression coefficients. Chebil has made a comparison between some models based on measured values of and M in Malasia, Indonesia, Brazil, Singapore and Vietnam. He showed that his model is the best estimate of the measured data [16]. The regression coefficient α and β are defined as α = and β = (5). Thus, using the refined Moupfouma model and Chebil model, the 1 min rain-rate cumulative distribution is full determined from the longterm mean annual rainfall data.the input parameters needed for the model are: point rainfall rate for the location for % of an average ear (mm/h), height above sea level of the Earth station (km), elevation angle, latitude of the Earth station (degree), frequenc (GHz) and effective radius of the Earth (8500 km) [17]..1. Formulations Two Tx Antenna Schemes One example of the Alamouti code encodes the Q = complex smbols x 1, x to be transmitted during T = smbol periods, in the form x1 x X Ala( x1, x) = x x 1 Tx Antennas: STTD For Nr = 1 receive antenna, the optimal space time block code for two transmit antennas is STTD, with smbol rate s = QT= 1. Optimalit is seen in man different was. In flat fading, the received signal is = Xh noise = noise 1 1 Here, the X is normalized b 1 so that the transmit power is Tr X X = x x. 1 Conjugating the received signal where the equivalent channel matrix is noise where 1 x1 H noise, = 1 x 1 h1 h H = h h 1 ( 1 ) 1 1 x1 H h h = x x1 = x x1 x X = x x1 h1 h = h 1 1 x1 x h1 = noise x x1 h 85

4 Conjugating the received signal, From where, But, = Hx noise, = noise noise = h1 h x1 noise = h h 1 x H 1 x1 H noise = x 1 h1 h H = h h 1 x = noise 1 1 H H x 1 h1 h H = h h1 1 1 h1 h h1 h H H = h h1 h h1 1 hh 1 1 hh hh 1 hh 1 = hh1 hh 1 hh hh h1 h 0 = 0 h1 h ( ) 1 Tr H H = h h 1 0 ( 1 ) ( 1 ) 1 1 x1 1 x1 H h h noise h h noise = 0 1 x = x Xˆ = Hx n Z = H Hx noise ml = arg max x AΩ ( ) e ( ) ( x) Ω x = X Z X H Hx ˆ X = arg min Z H Hx, ml Q V Q = V QV, Q = 1 = H H 86

5 From this,the complexit of the ML problem is in general exponential in the model dimension. Here, the problem dimension is essentiall dictated b the dimensions of the modulation matrix. With orthogonal designs, the off-diagonal elements of the equivalent correlation matrix vanish, and optimal ML detection is performed for each smbol independentl of each other. However, the equivalent correlation matrix for non-orthogonal schemes entertains non-zero off-diagonal elements, and this leads to a combinatorial optimization problem. In space time coding problems, the ML solution is feasible if the dimensionalit of is sufficientl low, or it has some special structure that can be exploited. Unfortunatel, a low model dimension tends to require either a low smbol rate or a small number of antennas. The rain drop size distribution is exponentiall expressed mathematicall as where D is the median drop diameter and ( ) ( ) m diameters between D and D dd mm. D 1 3 ( ) e D N D N m mm m = 0 (1) N D d D is the number of drops per cubic meter with The rainfall is related to N( D ) and also to the terminal velocit of ( ) second with diameter D b Step 1. Calculate the rain height ( ) ( ) ( ) V D the falling drops in meters per 3 3 = π DV D N D d D mm hr () h (km) as h = h km (3) where h 0 is the 0 C isotherm height above mean sea level at the desired location. Step. Determine the slant-path length L s, below the rain height from ( h hs) km if θ 5 C Ls = (4) sinθ where θ is Elation angle in degrees, Step 3. Obtain the horizontal projection, Slant path through rain source: ITU [18] h s is the rain height in km. L G, of the slant path length from L = L cos θ km (5) G s Step 4. Determine the rainfall rate,, exceeded for % of an average ear, with 1-min integration time. It can be calculated with the help of statistical data available in various meteorological databases. Step 5. Calculate the specific attenuation, γ, b using the frequenc dependent regression coefficients and using, 87

6 ( ) γ = k α db km (6) where K and α depend on frequenc, polarization, raidrop size distribution and temperature and obtained using, ( ) cos θ cos( ) kh kv kh kv t k = (7) ( ) cos cos( ) khαh kvαv khαh kvαv θ t α = (8) k where t is the polarization tilt angle relative to horizontal. Step 6. Determine the horizontal path adjustment factor, r for % the time using 1 r = LGγ e L G f Step 7. Calculate the adjusted rain path length, L (km), through rain using Lr G L = for ξ > θ (10) cosθ where ( hg hs) for LS = ξ θ (11) sinθ ξ h h 1 g s = tan Lr G Step 8. Obtain the vertical redution factor v, for %, of the time b using 1 v = θ 1 χ Lγ 1 sinθ 31 1 e 0.45 f where Step 9. Determine the effective path length through rain, χ = 36 φ, for φ < 36 χ = 0, for φ 36 L E L E (Km), given b (9) (1) (13) (14) (15) = Lv (16) Step 10. Calculate the predicted attenuation exceeded for % of an average ear b using A = γ L db (17) Step 11. The estimated attenuation to be exceeded for the other percentages of an average ear, in the range 0.001% to 10% ma then be estimated using A as E A Ln( p) 0.045Ln( A) βsinθ( 1 p) p p = A (18) 88

7 where p is the percentage probabilit of interest and β is given b for p 1%, β = 0 (19) for p < 1%, β = 0 if φ 36 ( ) β = φ 36 for θ 5 and φ < 36 ( ) β = φ sin θ, for θ < 5 and φ < 36 (0) (1) ().. Effects of ain The most well known effect of rain is that it attenuates the signal. The attenuation is caused b the scattering and absorption of electromagnetic waves b drops of liquid water. The scattering diffuses the signal, while absorption involves the resonance of the waves with individual molecules of water. Absorption increases the molecular energ, corresponding to a slight increase in temperature and results in an equivalent loss of signal energ. Attenuation is negligible for snow or ice crstals in which the molecules are tightl bound and do not interact with the waves. 3. Conclusions ain rate and rain attenuation contour maps have been developed for 0.1% and % of the time using the refined Moupfouma model for rain rate maps and ITU- 618 for the rain attenuation maps over Ghana. The 0.1% of time of rain attenuation is needed for WiMAX network service-availabilit. The information from these maps will be useful in the preliminar design for both terrestrial and earthsatellite microwave links, and to provide a broad idea of rain attenuation to microwave engineers for the proposed launching of.6 GHz WiMAX Network. Majorit of the studies on Earth-space propagation have been conducted in Europe, the United States and Asia. But it will be more crucial for studies to be conducted in a tropical location like Kumasi, Accra,Tema-Ghana because of its high rainfall intensit. In order to compute reliable rain attenuation for a given location, an appropriate distribution of rainfall rate for the site is required. ainfall rate statistics specified on a percent of time basis, that is the percent of time in a ear or a month that the rain rate equals or exceeds a specific value is used in the rain attenuation prediction model. The ITU rain attenuation prediction method is based on % of a ear rain rate parameter. Data for the distribution must be based on long-term (tpicall more than 10 rs) measured data with 1- minute integration time. But a large amount of rainfall data, tpicall collected b meteorological agencies in man countries is available for longer integration time such as 30 min., 60 min., etc. This is the case with Ghana, hence the need to rel on the data provided b ITU. In this research, ITU- model is used to predict the rainfall rate and attenuation due to rain, at KNUST, Ghana. The attenuation is calculated, for different rainfall rates and exceedence percentages of an average ear. The preliminar results indicate that the attenuation increases with frequenc and rainfall rate. These predicted values can be compared with the measured experimental data after installation of the setup in the location. As for future works, we will consider (1) constructing a test bed to measure rain attenuation experimentall from measured data per 1min of integration time and comparing results with simulated equations of rain attenuation and () looking at other forms of atmospheric conditions that affect the signal strength of WiMAX networks such as dust particles and developing equations to calculate attenuation. eferences [1] Choi, Y.S., Lee, J.H. and Kim, J.M. (1997) ain Attenuation Measurements of the Koreasat Beacon Signal on 1 GHz. CLIMPAA98, Ottawa, Vol. 08, No. 11. [] Characteristics of Precipitation for Propagation Modelling, ecommendation ITU- P.837-4, ITU- P Sers., ITU-, Int. Telecomm. Union, Geneva, 003. [3] Salonen, E.T. and Poiares-Baptista, J.P.V. (1997) A New Global ainfall ate Model. Proceedings of the 10th International Conference on Antennas and Propagation (Pub N ),, [4] Segal, B. (1986) The Influence of aingauge Integration Time on Measured ainfall-intensit Distribution Functions. 89

8 Journal of Atmospheric and Oceanic Technolog, 3, [5] Crane,. and Dissanaake, A.W. (1997) ACTS Propagation Experiment: Attenuation Distribution Observations and Prediction Model Comparison. Proceedings of the IEEE, 85, [6] Dutton, E.J. (1984) Microwave Terrestrial Link ain Attenuation Prediction Parameter Analsis. US Department of Communications, National Telecommunications and Information Administration (STIA) Tech. ep., Volumes [7] Emiliani, L.D., Agudelo, J., Gutierrez, E., estrepo, J. and Fradique-Mendez, C. (004) Development of ain-attenuation and ain ate Maps for Satellite Sstem Design in the Ku and Ka bands in Colombia. IEEE Antennas and Propagation Magazine, 46, [8] Ajai, G.O., Feng, S., adicella, S.M. and edd, B.M. (1996) Handbook on adio Propagation elated to Satellite Communications in Tropical and Subtropical Countries. ICTP, Trieste, [9] Moupfouma, F. (1985) Model of ainfall-ate Distribution for adio Sstem Design. IEEE Proceedings, 13, [10] Chebil, J. and ahman, T.A. (1999) Development of 1 min ain ate Contour Maps for Microwave Applications in Malasia Peninsula. Electronics Letters, 35, [11] Gunes, M., Gunes, F. and Dimiller, K. (1994) Development of a Climatic Map of Attenuation b ainfall for Turke. Proceedings of the 7th Mediterranean Electrotechnolog Conference,, [1] Dutton, E.J. and Doughert, H.T. (1979) Year-to-Year Variabilit of ainfall for Microwave Applications in the USA. IEEE Transactions on Communications, 7, [13] Specific Attenuation Model for ain for Use in Prediction Methods. ecommendation P.838-1,, ITU- P Sers., ITU-, Int. Telecomm. Union, Geneva, [14] Crane,.K. (1996) Electromagnetic Wave Propagation through ain. Wile Interscience, New York. [15] Gallois, A.P., Hartigan, P.P. and Bock, A.M. (1989) A Comparison of Slant Path Attenuation Models Applied to the Selection of Satellite Beacon eceiver Sites. Proceedings of the 6th International Conference on Antennas and Propagation,, [16] Propagation Data and Prediction Methods equired for the Design of Earth-Space Telecommunications Sstems. ecommendation P.618-8, ITU- P Sers., ITU-, Int. Telecomm. Union, Geneva, [17] Watson, P.A., Sathiaseelan, V. and Potter, B. (1981) Development of a Climatic Map of ainfall Attenuation for Europe, No. 300, 134. Post Graduate School of Electrical and Electronic Engineering, Universit of Bradford, West Yorkshire, ep. [18] ITU- ec. P.618-8, Propagation Data and Prediction Methods equired for the Design of Earth-Space Telecommunication Sstems. International Telecommunications Union, Geneva, April