7th Latin American Workshop On Communications - 2015 Analysis of electromagnetic propagation of 5 scenarios in Mexico City Jorge Sosa Pedroza1, Fabiola Martínez Zúñiga2, Ma. Elena Acevedo Mosqueda3 Instituto Politécnico Nacional, ESIME Zacatenco Sección de Estudios de Posgrado e Investigación Edificio Z-4, 3er piso, Col. Lindavista, C.P.07738, México, D.F. jsosa@ipn.mx1, fmartinezzu@ipn.mx2, eacevedo@ipn.mx3 Abstract. We present a study of electromagnetic Our analysis considers comparison of two models: COST attenuation over different environments of Mexico City, for 231 and Erceg, for similar environment as the scenario 2 (shown an empirical modification of Free Space Attenuation (FSA). below); for COST we use the big city model, applying all Our work uses results of in-site measurements in different communications system parameters to equation (1) places in the city. We intend to measure attenuation in different places to compare and validate modification of 𝑃𝐶𝑂𝑆𝑇 = 46.3 + 33.9 ∗ log 𝑓 − 13.82 ∗ log 𝐴 𝑇 − 𝑎(𝐴𝑅 ) FSA. Even we are working in different communications +(44.9 − 6.55 ∗ 𝑙𝑜𝑔 𝐴 𝑇 ) ∗ 𝑙𝑜𝑔 𝑑 + 𝐶 … (1) standards, we present results for WiMAX at 3.5 GHz. Where: f = 3.5 GHz. Resumen. Presentamos en este trabajo un estudio de AT = 30 m (Transmission height antenna) atenuación en propagación electromagnética en diferentes d = 100 – 1200 m escenarios en la ciudad de México, para una modificación C = 3 (correction factor) empírica de la Atenuación del Espacio Libre (AEL). Nuestro Receiver height antenna correction: trabajo usa resultados de medición en diferentes lugares de 𝑎(ℎ𝑅 ) = 3.2 ∗ [𝑙𝑜𝑔 (11.75 ∗ 𝐴𝑅 )]2 − 4.97, with AR = 1.5 m la ciudad. La intención es comparar con las mediciones de atenuación para comparar y validar la modificación de AEL. EIRP=19.5 dB Aunque hemos hecho el experimento para diferentes For Erceg Model we use the A zone, meaning hills and a medium estándares de comunicaciones, presentamos en este trabajo density of trees. Then we use equation 2: 𝑑 los resultados para WiMAX en 3.5 GHz. 𝑃𝐸𝑟𝑐𝑒𝑔 = 𝑃𝐹𝑆𝐴. + 10 ∗ 𝛾 log (𝑑 ) + 𝑠; 𝑑 ≥ 𝑑0 …(2) 0 Keywords: Modification of Free Space Attenuation, Where: Measurement of urban attenuation, WiMAX. PFSA. = Free space loss Palabras clave: Atenuación del espacio libre, Mediciones d0 = Reference distance (100 m). de atenuación en zonas urbanas, WIMAX. 𝑐 𝛾 = (𝑎 − 𝑏 ∗ ℎ𝑏 + ) + 𝑥 ∗ 𝜎𝛾 ; 10 𝑚 ≥ ℎ𝑏 ≥ 80𝑚 ℎ𝑏 I. INTRODUCTION 𝑠 = 𝑦𝜎 𝜎 = 𝜇𝜎 + 𝑧 ∗ 𝜎𝜎 Using Erceg parameters, we define: Wireless communication systems are evolving rapidly, EM attenuation in urban areas is now a most for the system design. 𝑎 = 4.6 𝑏 = 0.0075 𝑐 = 12.06 Altough models as Okumura [1], Hata [1], Cost 231 (a modification of that of Hata), Erceg [3], Shittu [4] and others, are 𝜇𝜎 = 10.6 𝜎𝜎 = 2.3 𝜎𝛾 = 0.57 used all around the world to predict propagation losses, for each Results of our comparison between measurements made in site should be tested to prove at least, the deviation between Mexico City and actual models, show important differences, as prediction and measurement. We have done measurements all seen in figure 1. over Mexico City in different frequencies, comparing them with prediction of most popular models, for different communications Figure 1 shows that Mexico City environment is different to standards as WiMAX (Worldwide Interoperability Microwave those where, the models are defined. This paper, analyze our own Access), which is a metropolitan area service used with one or scenarios and adjust the Free Space Model (FSM) trying to find more base stations at different frequencies: 2.3 GHz, 2.5 GHz, the best relationship between measurements and model 3.3 GHz, 3.5 GHz and 5.8 GHz [5,6]. modifications. Considering those differences of figure 1, we Copyright © 2015 for the individual papers by the papers’ authors. Copying permitted for private and academic purposes. This volume is published and copyrighted by its editors. Latin American Workshop On Communications' 2015 Arequipa, Peru Published on CEUR-WS: http://ceur-ws.org/Vol-1538/ propose to use a modification of FSA equation, comparing it repeat the procedure. On the other hand we want to emphasize with regression curves of measurement. that we are comparing only attenuation due the environment, this means that we take a measurement power at some distance of base station and relate it to the measurement at other distance. We are still working in the procedure, and we are now measuring in similar areas in other places in the city, to compare the original results. II. MEASUREMENT METHODOLOGY Measurements were performed using base stations located in National University (UNAM) with base station in the Humanities Building at the south of the city, in Instituto Politécnico Nacional (IPN) with station located in the Dirección de Cómputo y Comunicaciones (DCyC) at the city’s north and the Instituto de Ciencia y Tecnología (ICYT) in downtown. Figure 1. Comparison of measurement with COST and Erceg Three scenarios were chosen due their similarities but also by The idea is to measure field attenuation in an area with their differences. Table I shows characteristics of defined characteristics, make modification of FSA equation and communications system, transmitter antenna height is almost the after that, take measurement in other different place in the city, same for all scenarios. with similar characteristics and compare the new measurement regression curves with former modification of FSA. For the TABLE I COMMUNICATION SYSTEM FEATURES experiment we choose to different places, 30 Km apart, one in Operating frecquency 3,4785 GHz the north and the other in the south of Mexico City. Experiment Bandwidth 3,5 MHz starts with choosing similar scenarios in both sites, we find 4 Transmitter power -7 dBW areas with similar characteristics, those from 1 to 4. The fifth scenario was in downtown, which is a unique place in the city Transmitter antenna height 29 – 70 m and do not have a similar site to be compared with, but we Receiver antenna height 2m presented because it was part of the measurements. Transmitter antenna gain 14,8 dBi We propose to use Free Space Attenuation, modifying the Receiverer antenna gain 3,6 dBi equation exponent. As is known, FSA supposes a free obstacle region between transmitter and receiver; it is given by [3]: Measurements were made using an analyzer Anritsu Master Spectrum MS2721B, which includes a GPS antenna for 4𝜋𝑑𝑓 2 referenced positioning. Figure 2 shows the measurement 𝑃𝐿𝐹𝑆 = ( ) (3) scheme; the equipment was mounted on top of a vehicle, saving 𝑐 the received power level and geographic location (latitude and where longitude coordinates) at each measurement point. d: distance between transmitter and receiver f: operating frequency c: speed of light Equation (3) is often expressed in dB as: 4𝜋 𝐿𝑃𝐸𝐿 = 20 log + 20 log 𝑑 + 20 log 𝑓 (4) 𝑐 The new proposed equation is: 2-2 4𝜋𝑑𝑓 𝑿 𝑃𝐿𝐹𝑆 = ( ) ± 𝒀 (5) 𝑐 Adjustment is made changing the exponent value (X) for the slope and (Y) for losses magnitude, in equation (5). Next step Fig 2. Measurement scheme was to adjust FSM finding a match for slope and attenuation as best as possible with measurement regression curves. After that Measurements were taken every 20 seconds, at that time we compare modified curves with similar scenario in other place, analyzer updates GPS position. A data base with extension and if there is concordance, we can suppose that modification of wxme is constructed, extracting measurements from spectrum equation (5) is good enough. We understand that this analysis is analyzer, becoming in computer txt files containing latitude, valid only for the used frequency, for a different one we have to longitude, distance and input power. Database is used to construct linear regression curves, to be compared with adjustments of FSA. III. IDENTIFICATION OF SCENARIOS We identified 5 scenarios accordingly of existing features of different areas in Mexico City, which are described in following paragraphs: A. Scenario 1: low buildings with low tree density This scenario considers an area with a low tree density, and low high buildings, as shown in Figure 3. Figure 5. Scenario 3 Figure 3. Scenario 1 B. Scenario 2: low buildings with medium tree density Figure 6. Scenario 4 The scenario considers an area with medium tree density and a low height of the buildings as the one shown in figure 4. Figure 7. Scenario 5 Figure 4. Scenario 2 Scenario 5 is a unique environment of Mexico City (and many Latin American cities), it is located at historic downtown; it has C. Scenario 3: area with a high tree density some very unique construction features such as: large width Figure 5 shows a scenario with a high tree density. walls, tall buildings, and narrow streets. D. Scenario 4: tall buildings with medium tree density Figures 8 and 9 show UNAM (south of the city) with yellow mark, showing position of base station at Humanidades II The scenario considers an area with medium tree density and building and the one at IPN (north of the city) with mark in high height buildings. DCyC building (Dirección de Cómputo y Comunicaciones). E. Scenario 5: colonial city Measurements were taken in both zones for scenarios 1 through 4, each one bounded for different colors. The blue frame define scenario 1, while scenario 2 is red, green is the scenario 3 and scenario 4 orange. Each scenario was visually distinguished from both base stations photos. As an example, area for scenario 1 is mostly filled by buildings with few green zones. Others can be distinguished in photographs from tree density. Figure 10. Base Station of scenario 5. IV. RESULTS Figure 8. Radio base located in the UNAM Once the scenarios were identified in maps, measurement tables were constructed and obtained linear regression curves we were ready to modify FSA using UNAM measurement and then compare them for those at IPN. To do so we adjust FSA finding a match for slope and attenuation as best as possible. Adjustment is made changing the exponent value (X) for the slope and (Y) for losses magnitude, in equation 5 considering a difference no higher than 3 dB: 4𝜋𝑑𝑓 𝑿 𝐿𝐹𝑆𝐴 = ( ) ± 𝒀 (5) 𝑐 Following curves show results of our analysis. Figure 11 shows adjustments for scenario 1; we found that X=3 and Y=29 then: 4πdf 𝟑 LS1 = − ( ) + 𝟐𝟗 c Figure 9. Base station in IPN Figure 10 shows the unique scenario 5. Although is not clear from the photo the differences with other scenarios, is possible to distinguish them from Figure 10, specially the narrow background street between the colonial buildings. After each bounded polygon was defined by its geographical position, we created tables with each point measurements and then constructed regression curves showing attenuation behavoir. To validate scenarios 1-4 we take measurements in Figure 11. Adjustment of FSM for scenario 1. UNAM to define FSA new parameters and then compare them with those taken in IPN to validate X and Y parameters. 100 m and 1200 m. Differences are greater for larger distances, meaning scenario 4 needs further analysis. Figure 12. Comparison and adjust of FSA for scenario 2. Figure 12 depicts measurements comparison for scenario 2. Figure 14. Comparison and adjust of FSA for scenario 4. giving: Although two power slopes are different, a good FSA 4𝜋𝑑𝑓 𝟐 adjustment was found as: 𝐿𝑆2 = − ( ) − 𝟐𝟖 𝑐 4𝜋𝑑𝑓 𝟑 𝐿𝑆4 = − ( ) + 𝟐𝟏 𝑐 We select a FSM adjust curve between both base station regression curves, giving a difference between FSM and measurement no higher than 1.5 dB. As seen the exponent value As scenario 5 is unique, is not possible to compare with any do not change (X=2); magnitude of attennuation factor is other curve. Figure 15 shows FSA adjustment of regression selected as Y = -28 for scenario 2. curve. As can be seen from Figure 15, there is a sharp slope fall, similar to that of scenario 3, meaning a zone of high attenuation, In the same way we compare curves for scenario 3, as shown due the tall buildings with very dense walls and narrow streets. in Figure 13; again the slope of regression curves are similar, Adjustment for FSA is: with a difference of power no greater than 5 dB. After adjustment, we select a curve for FSA between both regression 4𝜋𝑑𝑓 𝟐 curves; lost equation is expressed as: 𝐿𝑆5 = − ( ) − 𝟑𝟔 𝑐 4𝜋𝑑𝑓 𝟐 𝐿𝑆3 = − ( ) − 𝟑𝟏 𝑐 Fig. 15. Adjust of FSM for scenario 5. Figure 13. Comparison and adjust of FSA for scenario 3. Table II shows a summary of adjustments for all scenarios. Following same procedure, we compare results for scenario 4. Regression curves are shown in Figure 14. As seen slopes of both curves are different, although no more than 1.5 dB between TABLE II SUMMARY OF FSM ADJUSTMENT Exponent Loss (dB) Scenario 1 3 + 29 Scenario 2 2 - 28 Scenario 3 2 - 31 Scenario 4 3 + 21 Scenario 5 2 - 36 As seen in Table II, scenarios 1 and 4 have similar slopes, with an exponent of 3. Furthermore the loss adjustment for FSM is increased with 29 and 21 dB respectively. For scenarios 2 and 3, the slope has the same exponent of 2, and loss requires an adjustment of -28 and –31 dB respectively, only a 3 dB difference. V. CONCLUSION After identification of some common scenarios in Mexico City as: low buildings-low tree density; low buildings-medium tree density; high tree density zone; tall buildings-medium tree density; Colonial City, we compare measurements over city streets with FSA, to find a relationship between them. Considering similarities for two base stations, we validate scenarios selection, at least for 4 of them, leaving scenario 5 as unique. Comparing measurements for each scenario with adjusts of exponent and amplitude of FSA losses, we conclude that model can predict path loss for WiMAX or similar communication standard. We think that 5 dB differences is a good margin, to predict propagation of mobile communication systems over an environment as Mexico City. Probably we have to define an accepted margin, but accordingly with our experience a 10 dB could be a good number. REFERENCES [1] Y. Okumura et al. “Field strength and variability in VHF and UHF land-mobile radio service” Rev. Elec. Commun. Lab., vol. 16, 1968. [2] Masaharu Hata “Empirical Formula for Propagation Loss in Land Mobile Radio Services” IEEE Transactions on Vehicular Technology, Vol. VT-29, No. 3, August 1980. [3] Vinko Erceg et al. “An Empirical Based Path Loss Model for Wireless Channels in Suburban Environments”, IEEE, Journal on Selected Areas in Communications, Vol. 17, No. 7 July 1999. [4] W.A Shittu, “Prediction of Received Signal Power and Propagation Path Loss in Open/Rural Environments using modified Free-Space Loss and Hata Models”, 2008. [5] S. Y. Tang, “WiMAX Security and Quality of Service”, 1st ed, Ed. Wiley, 2010. [6] D. Pareek, “WiMAX Taking Wireless to the MAX”, 1st ed, Ed. Taylor & Francis Group, 2006.