Comparison of Measurement and Prediction of ITU-R Recommendation P.1546

Size: px

Start display at page:

Download "Comparison of Measurement and Prediction of ITU-R Recommendation P.1546"

Posy Bridges
5 years ago
Views:

1 Comparso of Measuremet ad Predcto of ITU-R Recommedato P.546 Chag-Hoo Lee *, Nam-Ryul Jeo *, Seog-Cheol Km *, Jug-m Lm * Isttute of New Meda ad Commucatos, Seoul Natoal Uversty, Cosoldated Mateace Depot, Isttute of Techology for Mateace E-mal : (lchjsa, jeo35, sckm@maxwell.su.ac.kr, teemyteemy@hamal.et Tel : , Fax : Abstract Ths paper compares measuremet wth predcto of ITU-R Recommedato P.546. Ths recommedato maly depeds o propagato curves whch are based o the measuremet result. The basc feld stregth s derved from the curve correspodg to trasmttg atea heght, frequecy ad requred dstace ad the correcto factor s added to ths. There are two mportat correcto values ths recommedato. Frst factor s the correcto for recevg/moble atea heght ad secod factor s the terra clearace agle (TCA correcto. Ths paper specally focuses o frst correcto factor wth regard to R whch s represetatve of the heght of the groud cover surroudg recevg/moble atea. We propose several cosderatos to ehace the predcto accuracy of ITU-R Recommedato P INTRODUCTION I order to set up a effcet broadcastg etwork ad to predct the exact feld stregth, we have to cosder a lot of factors such as dstace, obstructo, frequecy, terra profle ad trasmttg/recevg atea heght, etc. Because broadcastg servce has to predct the feld stregth over expaded area, t s dffcult to calculate the exact feld stregth. So we have to choose the most sutable propagato chael model the broadcastg frequecy bad. Covetoal propagato models are Hata, Okumura ad ITU-R Recommedato P.546. Amog them, ITU-R Recommedato P.546 s developed recetly ad ths recommedato serves as a stadard areas of spectrum maagemet ad rado techology ad s costatly uder developmet. I ths paper, we compare measuremet wth predcto of ITU- R Recommedato P over four ctes of Korea. I secto 2, The ITU-R Recommedato P s shortly troduced. I secto 3, we llustrate measuremet procedure ad several statstcal aalyss methods. Ad the predcto results are compared wth measuremet results ad the result of statstcal aalyss s preseted. Fally, the coclusos are preseted secto ITU-R RECOMMENDATION P A. Overvew Ths s the recommedato of ITU for pot-to-area predcto of feld stregth for the broadcastg, lad moble, martme moble ad certa fxed servces (e.g. those employg pot-to-multpot systems the frequecy rage 30MHz to 3000MHz ad for the dstace rage km to,000km. The propagato curves represet feld-stregth values for kw effectve radated power (e.r.p at omal frequeces of 00, 600 ad 2000 MHz, respectvely, as fucto of varous parameters whch are frequecy, trasmttg atea heght, tme varablty ad dstace; these curves refer to are based o measuremet data maly relatg to mea clmatc codtos temperate regos cotag cold ad warm sea. The feld stregth ca be expressed as ( f sup f (, dfh, log ( df, ff, hf 2 ( μ / ( dsup, fsup, hsup ( E= E + E E + C + C db V m log ( df, ff, hf where C, C 2 are correctos for recevg atea heght ad terra clearace agle, whch are based o ITU-R Recommedato P.546.

2 6.03 J( v db for h2 R' (2 C = Kh ( log h 2 2 log R' db for h2> R' C = J ( v' J( v db (3 2 for recevg atea heght for the omal dstaces (d=0km, 20km, 30km. It shows that the chage of R have a great effect o the feld stregth. If R s 30m, correcto value ca be about -23dB. B. Correctos B-. Terra clearace agle The terra clearace agle correcto mproves the predcto accuracy by takg to accout terra ear the recevg atea. The terra clearace agle θ tca s gve by: θtca = θ θr degrees (4 where θ s the elevato agle of the le from the recevg/moble atea whch just clears all terra obstructos the drecto of the trasmtter/base atea over a dstace of up to 6 km but ot gog beyod the trasmttg/base atea. The referece agle, θ r, s gve by: Fgure. Defto of R' θ h h 000d s 2s r = degrees (5 where h s ad h 2s are the heght of trasmttg/base ad recevg ateas above sea level respectvely. TCA correcto s based o the heght of Tx/Rx atea ad the terra betwee them. B-2. Recevg/moble atea heght The feld stregth values gve by the curves are for a referece recevg/moble atea at heght, R, represetatve of the heght of the groud cover surroudg the recevg atea, subject to a mmum heght value of 0m. Examples of referece heghts are 0 m for suburba areas, 20 m for urba areas ad 30 m for dese urba areas. Where the recevg/moble atea s o lad accout should frst be take of the elevato agle of the arrvg ray by calculatg a modfed represetatve clutter heght R' (m, gve by: 000dR 5h R = 000d 5 m ' ( (6 Fgure 2. Correcto versus R 3. MEASUREMENT PROCEDURE AND STATISTICAL ANALYSIS Measuremet was performed four ctes whch are Pusa, Daegu, Daejeo, ad Kwagju Korea. The locato of the trasmttg ad recevg atea s measured wth GPS system. The heght of recevg atea s 9m. Path profle s extracted usg self-made GIS Korea. Fgure 3 s a example of Korea path profles extracted usg the GIS other to predct the feld stregth whose resoluto dstace s 50m. where h ad R (m ad dstace d (km. A correcto depedg o the groud cover should be appled f the recevg/moble atea heght s dfferet from the surroudg groud cover R. Fgure presets the relato R wth R' []. Fgure 2 llustrates correcto

3 dcato of the qualty of the predcto model. Gve a path loss threshold, f both predcted ad measured path loss s greater or less tha the threshold, the predcto s regarded as correct rrespectve to the devato of the predcted from the measured value. For further detals o the ht rate metrcs, refer to [3]. To compare the dfferet models usg THR, we troduced [2] the average total ht rate error (AHRE. The AHRE s the mea devato from 00% THR ad s expressed as Fgure 3. Path profle I ths paper, frst order statstcs (mea error, stadard devato, correlato coeffcet ad ht rate metrcs [2] [3] have bee used to evaluate the result. A. Frst Order Statstcs The mea of the predcto error, e, postve max error, e p, egatve max error, e, ad ts stadard devato, σ e, are frst order statstcs tradtoally used for evaluatg the accuracy of predcto models. Here, e s defed as the dfferece betwee the predcto curve, p, ad measured values, m, o a logarthmc [db] scale,.e. assumg that the measured values are exactly correct. B. Correlato Coeffcet The correlato coeffcet provdes a measure of the degree of lear relatoshp betwee two radom varables ad s calculated as max( NL T AHRE = 00% THR( LT N LT m( NL T where L T s the path loss threshold ad N LT s the umber of THR pots. A small value of AHRE dcates a good ft betwee predcted ad estmated. Fgures 4-8 preset the THR correspodg to the measuremet. Predcto results are obtaed by usg ITU-R Recommedato P- 546 wth R=20 or R=30. The statstcal aalyss summary the Table shows how the dfferece the applcato of dfferet R affects the fal predcto. Dfferece of mea error, stadard devato, maxmum error, correlato coeffcet ad AHRE are about 4dB, 0.05dB, 3dB, 0.04 ad 0.3%. (8 r e = ( m m( p p 2 2 ( m m ( p p (7 where m ad p are the meas of the measured ad predcted values, respectvely, ad N s s the umber of samples. A correlato coeffcet close to oe dcates a strog lear relatoshp. C. Ht Rate Metrcs To complemet the frst order statstcs, ht rate metrcs are troduced by Owadally, Motel, ad Sauders [3]. The ht rate s a measure of goodess whe predctg the coverage by a sgle path loss threshold. I ths paper, the locato specfc total ht rate (THR s used as a drect Fgure 4. Total ht rate for total areas

4 Fgure 5. Total ht rate for Pusa Fgure 8. Total ht rate for Kwagju Table. Statstcal aalyss summary Measuremet N s e σ e e p e r e AHRE Fgure 6. Total ht rate for Daegu Total Pu sa Dae gu Dae jeo Kwag Ju R= R= R= R= R= R= R= R= R= R= CONCULSION AND FUTURE WORK Fgure 7. Total ht rate for Daejeo Ths paper shows that ITU-R Recommedato P.546 predcto ca be chaged as the surroudg groud cover R whch s the ste specfc value. To predct exact the feld stregth usg ITU-R Recommedato P.546, the exact R value should be determed. To develop relable predcto model, more cosderato should be take about the determato of correcto wth regard to R. The applcato of varable R a urba evromet ca ehace the accuracy of the predcto. Though measuremets are performed same area ad each atea heght s same, the applcato of R should be dfferet. I cocluso, R value LOS evromet must be smaller tha that LOS evromet. To

5 develop relable correcto model wth regard to R, more aalyss about the groud cover surroudg the recevg atea should be acheved usg the dffracto model ad addtoally a lot of measuremet data may be obtaed dfferet type of evromet. ACKNOWLEDGMENT Ths work was supported part by ITRC ad Seoul R&BD Program (0544. REFERENCES [] ITU-R Recommedato P.546-2, Method for pot-toarea predctos for terrestral servces the frequecy rage 30 MHz to 3000 MHz, September [2] E. Ostl, H-J. Zeperck, ad H. Suzuk, Evaluato of the ew sem-terra based propagato model Recommedato ITU-R P.546, IEEE Semaual Vehcular Techology Coferece, vol., pp. 4-8, Orlado, USA, October [3] A. S. Owadally, E. Motel, ad S.R. Sauders, A comparso of the accuracy of propagato models usg ht rate aalyss, IEEE Sem-aual Vehcular Techology Coferece, vol. 4, pp. 7-, Atlatc Cty, USA, October 200. [4] A.J.M. Soares, C.G. Rego, E.V. Melo, F.J.S. Morera, G.W.O. Costa, J.P.D.M. Pto, M.A.S. Mayrk, N. Pasqual, P.H.P. Carvalho, W.K. Gumaraes, Measuremets of a CW sgal Brazl ad Comparso wth Predcto usg ITU- R P.546, IEEE Wreless Commucatos ad Appled Computatoal Electromagetcs, page , Aprl 2005.

Short Note: Merging Secondary Variables for Geophysical Data Integration

Short Note: Merging Secondary Variables for Geophysical Data Integration Short Note: Mergg Secodary Varables for Geophyscal Data Itegrato Steve Lyster ad Clayto V. Deutsch Departmet of Cvl & Evrometal Egeerg Uversty of Alberta Abstract Multple secodary data from geophyscal