Recommendation ITU-R P (02/2012)

Transcription

1 Recommenation ITU-R P (0/01) Propagation ata an preiction methos for the planning of short-range outoor raiocommunication systems an raio local area networs in the frequency range 300 MHz to 100 GHz P Series Raiowave propagation

2 ii Rec. ITU-R P Forewor The role of the Raiocommunication Sector is to ensure the rational, equitable, efficient an economical use of the raio-frequency spectrum by all raiocommunication services, incluing satellite services, an carry out stuies without limit of frequency range on the basis of which Recommenations are aopte. The regulatory an policy functions of the Raiocommunication Sector are performe by Worl an Regional Raiocommunication Conferences an Raiocommunication Assemblies supporte by Stuy Groups. Policy on Intellectual Property Right (IPR) ITU-R policy on IPR is escribe in the Common Patent Policy for ITU-T/ITU-R/ISO/IEC reference in Annex 1 of Resolution ITU-R 1. Forms to be use for the submission of patent statements an licensing eclarations by patent holers are available from where the Guielines for Implementation of the Common Patent Policy for ITU-T/ITU-R/ISO/IEC an the ITU-R patent information atabase can also be foun. Series of ITU-R Recommenations (Also available online at Series BO BR BS BT F M P RA RS S SA SF SM SNG TF V Title Satellite elivery Recoring for prouction, archival an play-out; film for television Broacasting service (soun) Broacasting service (television) Fixe service Mobile, raioetermination, amateur an relate satellite services Raiowave propagation Raio astronomy Remote sensing systems Fixe-satellite service Space applications an meteorology Frequency sharing an coorination between fixe-satellite an fixe service systems Spectrum management Satellite news gathering Time signals an frequency stanars emissions Vocabulary an relate subjects Note: This ITU-R Recommenation was approve in English uner the proceure etaile in Resolution ITU-R 1. Electronic Publication Geneva, 01 ITU 01 All rights reserve. No part of this publication may be reprouce, by any means whatsoever, without written permission of ITU.

3 Rec. ITU-R P RECOMMENDATION ITU-R P Propagation ata an preiction methos for the planning of short-range outoor raiocommunication systems an raio local area networs in the frequency range 300 MHz to 100 GHz (Question ITU-R 11/3) ( ) Scope This Recommenation provies guiance on outoor short-range propagation over the frequency range 300 MHz to 100 GHz. Information is given on path loss moels for line-of-sight (LoS) an non-line-of-sight (NLoS) environments, builing entry loss, multipath moels for both environments of street canyon an over roof-tops, number of signal components, polarization characteristics an faing characteristics. The ITU Raiocommunication Assembly, consiering a) that many new short-range (operating range less than 1 m) mobile an personal communication applications are being evelope; b) that there is a high eman for raio local area networs (RLANs) an wireless local loop systems; c) that short-range systems using very low power have many avantages for proviing services in the mobile an wireless local loop environment; ) that nowlege of the propagation characteristics an the interference arising from multiple users in the same area is critical to the efficient esign of systems; e) that there is a nee both for general (i.e. site-inepenent) moels an avice for initial system planning an interference assessment, an for eterministic (or site-specific) moels for some etaile evaluations, noting a) that Recommenation ITU-R P.138 provies guiance on inoor propagation over the frequency range 900 MHz to 100 GHz, an shoul be consulte for those situations where both inoor an outoor conitions exist; b) that Recommenation ITU-R P.1546 provies guiance on propagation for systems that operate over istances of 1 m an greater, an over the frequency range 30 MHz to 3 GHz, recommens 1 that the information an methos in Annex 1 shoul be aopte for the assessment of the propagation characteristics of short-range outoor raio systems between 300 MHz an 100 GHz where applicable.

4 Rec. ITU-R P Annex 1 1 Introuction Propagation over paths of length less than 1 m is affecte primarily by builings an trees, rather than by variations in groun elevation. The effect of builings is preominant, since most short-path raio lins are foun in urban an suburban areas. The mobile terminal is most liely to be hel by a peestrian or locate in a vehicle. This Recommenation efines categories for short propagation paths, an provies methos for estimating path loss, elay sprea an angular sprea over these paths. Physical operating environments an efinition of cell types Environments escribe in this Recommenation are categorize solely from the raio propagation perspective. Raiowave propagation is influence by the environment, i.e. builing structures an heights, the usage of the mobile terminal (peestrian/vehicular) an the positions of the antennas. Four ifferent environments are ientifie, consiere to be the most typical. Hilly areas, for example, are not consiere, as they are less typical in metropolitan areas. Table 1 lists the four environments. Recognizing that there is a wie variety of environments within each category, it is not intene to moel every possible case but to give propagation moels that are representative of environments frequently encountere. TABLE 1 Physical operating environments Propagation impairments Environment Urban very highrise Urban high-rise Urban/suburban low-rise Description an propagation impairments of concern Busiest urban eep canyon, characterize by streets line with high-ensity builings with several tens of floors which results in an urban eep canyon High ense builings an syscrapers interleave with each other which yiels to the rich scattering propagation paths in NLoS Rows of tall builings provie the possibility of very long path elays Heavy traffic vehicles an high flowrate visitors in the area act as reflectors aing Doppler shift to the reflecte waves Trees besie the streets provie ynamic shaowing Urban canyon, characterize by streets line with tall builings of several floors each Builing height maes significant contributions from propagation over roof-tops unliely Rows of tall builings provie the possibility of long path elays Large numbers of moving vehicles in the area act as reflectors aing Doppler shift to the reflecte waves Typifie by wie streets Builing heights are generally less than three stories maing iffraction over roof-top liely Reflections an shaowing from moving vehicles can sometimes occur Primary effects are long elays an small Doppler shifts

5 Rec. ITU-R P TABLE 1 (en) Environment Resiential Rural Description an propagation impairments of concern Single an ouble storey wellings Roas are generally two lanes wie with cars pare along sies Heavy to light foliage possible Motor traffic usually light Small houses surroune by large garens Influence of terrain height (topography) Heavy to light foliage possible Motor traffic sometimes high For each of the four ifferent environments two possible scenarios for the mobile are consiere. Therefore the users are subivie into peestrian an vehicular users. For these two applications the velocity of the mobile is quite ifferent yieling ifferent Doppler shifts. Table shows typical velocities for these scenarios. Environment TABLE Physical operating environments Typical mobile velocity Velocity for peestrian users (m/s) Velocity for vehicular users Urban high-rise 1.5 Typical owntown spees aroun 50 m/h (14 m/s) Urban/suburban low-rise 1.5 Aroun 50 m/h (14 m/s) Expressways up to 100 m/h (8 m/s) Resiential 1.5 Aroun 40 m/h (11 m/s) Rural m/h (-8 m/s) The type of propagation mechanism that ominates epens also on the height of the base station antenna relative to the surrouning builings. Table 3 lists the typical cell types relevant for outoor short-path propagation. Cell type Cell raius TABLE 3 Definition of cell types Typical position of base station antenna Micro-cell 0.05 to 1 m Outoor; mounte above average roof-top level, heights of some surrouning builings may be above base station antenna height Dense urban microcell 0.05 to 0.5 m Outoor; mounte below average roof-top level Pico-cell Up to 50 m Inoor or outoor (mounte below roof-top level) (Note that ense urban micro-cell is not explicitly efine in Raiocommunication Stuy Group 5 Recommenation.)

6 4 Rec. ITU-R P Path categories 3.1 Definition of propagation situations Four situations of base station (BS) an mobile station (MS) geometries are epicte in Fig. 1. Base station BS 1 is mounte above roof-top level. The corresponing cell is a micro-cell. Propagation from this BS is mainly over the roof-tops. Base station BS is mounte below roof-top level an efines a ense urban micro- or pico-cellular environment. In these cell types, propagation is mainly within street canyons. For mobile-to-mobile lins, both ens of the lin can be assume to be below roof-top level, an the moels relating to BS may be use Propagation over rooftops, non-line-of-sight (NLoS) The typical NLoS case (lin BS 1 -MS 1 in Fig. 1) is escribe by Fig.. In the following, this case is calle NLoS1. FIGURE 1 Typical propagation situation in urban areas BS 1 MS1 BS MS 3 MS MS 4 P

7 Rec. ITU-R P FIGURE Definition of parameters for the NLoS1 case (Distance between BS an MS antennas: ) BS ϕ MS Plan BS Δh b θ Cross-section h b Builing h r Δh m MS h m l w Dominant wave Direct wave One-time reflecte wave Two-time reflecte wave Several-time reflecte wave Diffracte wave b Three regions Direct wave ominant region Reflecte wave ominant region Diffracte wave ominant region Primary ominant wave Seconary ominant wave Non-ominant wave P The relevant parameters for this situation are: h r : average height of builings w : street with b : average builing separation ϕ : street orientation with respect to the irect path (egrees) h b : BS antenna height h m : MS antenna height l : length of the path covere by builings : istance from BS to MS. The NLoS1 case frequently occurs in resiential/rural environments for all cell-types an is preominant for micro-cells in urban/suburban low-rise environments. The parameters h r, b an l can be erive from builing ata along the line between the antennas. However, the etermination of w an ϕ requires a two-imensional analysis of the area aroun the mobile. Note that l is not necessarily normal to the builing orientation Propagation along street canyons, NLoS Figure 3 epicts the situation for a typical ense urban micro-cellular NLoS-case (lin BS -MS 3 in Fig. 1). In the following, this case is calle NLoS.

8 6 Rec. ITU-R P FIGURE 3 Definition of parameters for the NLoS case x w x 1 w 1 α MS BS P The relevant parameters for this situation are: w 1 : street with at the position of the BS w : street with at the position of the MS x 1 : istance BS to street crossing x : istance MS to street crossing α : is the corner angle (ra). NLoS is the preominant path type in urban high-rise environments for all cell-types an occurs frequently in ense urban micro- an pico-cells in urban low-rise environments. The etermination of all parameters for the NLoS case requires a two-imensional analysis of the area aroun the mobile Line-of-sight (LoS) paths The paths BS 1 -MS an BS -MS 4 in Fig. 1 are examples of LoS situations. The same moels can be applie for both types of LoS path. 3. Data requirements For site-specific calculations in urban areas, ifferent types of ata can be use. The most accurate information can be erive from high-resolution ata where information consists of: builing structures; relative an absolute builing heights; vegetation information. Data formats can be both raster an vector. The location accuracy of the vector ata shoul be of the orer of 1 to m. The recommene resolution for the raster ata is 1 to 10 m. The height accuracy for both ata formats shoul be of the orer of 1 to m. If no high-resolution ata are available, low-resolution lan-use ata (50 m resolution) are recommene. Depening on the efinition of lan-use classes (ense urban, urban, suburban, etc.) the require parameters can be assigne to these lan-use classes. These ata can be use in conjunction with street vector information in orer to extract street orientation angles.

9 Rec. ITU-R P Path loss moels For typical scenarios in urban areas some close-form algorithms can be applie. These propagation moels can be use both for site-specific an site-general calculations. The corresponing propagation situations are efine in 3.1. The type of the moel epens also on the frequency range. Different moels have to be applie for UHF propagation an for mm-wave propagation. In the UHF frequency range LoS an NLoS situations are consiere. In mm-wave propagation LoS is consiere only. Aitional attenuation by oxygen an hyrometeors has to be consiere in the latter frequency range. 4.1 LoS situations within street canyons UHF propagation In the UHF frequency range, basic transmission loss, as efine by Recommenation ITU-R P.341, can be characterize by two slopes an a single breapoint. An approximate lower boun is given by: LLoS,l = Lbp 0 log log10 Rbp Rbp for for Rbp > Rbp (1) where R bp is the breapoint istance an is given by: 4 h bh R m bp λ where λ is the wavelength. The lower boun is base on two-ray groun reflective moel. An approximate upper boun is given by: () L LoS,u = L bp 5 log log R bp R bp for for R > R bp bp (3) L bp is a value for the basic transmission loss at the brea point, efine as: L bp = λ 0 log 10 (4) 8πhb hm The upper boun has the faing margin of 0 B. In equation (3), the attenuation coefficient before breapoint is set to.5 because a short istance leas to wea shaowing effect.

10 8 Rec. ITU-R P Accoring to the free-space loss curve, a meian value is given by: L LoS,m = L bp 0 log log R bp R bp for for R > R bp bp (5) SHF propagation up to 15 GHz At SHF, for path lengths up to about 1 m, roa traffic will influence the effective roa height an will thus affect the breapoint istance. This istance, R bp, is estimate by: ( h ) ( ) = 4 b hs hm h R s bp (6) λ where h s is the effective roa height ue to such objects as vehicles on the roa an peestrians near the roaway. Hence h s epens on the traffic on the roa. The h s values given in Tables 4 an 5 are erive from aytime an night-time measurements, corresponing to heavy an light traffic conitions, respectively. Heavy traffic correspons to 10-0% of the roaway covere with vehicles, an 0.-1% of the footpath occupie by peestrians. Light traffic was % of the roaway an less than 0.001% of the footpath occupie. The roaway was 7 m wie, incluing 6 m wie footpaths on either sie. TABLE 4 The effective height of the roa, h s (heavy traffic) Frequency (GHz) h b (1) The breapoint is beyon 1 m. () No breapoint exists. h s h m =.7 h m = () () () () () 8 (1) ()

11 Rec. ITU-R P Frequency (GHz) TABLE 5 The effective height of the roa, h s (light traffic) h b (1) No measurements taen. () The breapoint is beyon 1 m. h s h m =.7 h m = (1) (1) 4 () () (1) 4 () () (1) When h m > h s, the approximate values of the upper an lower bouns of basic transmission loss for the SHF frequency ban can be calculate using equations (1) an (3), with L bp given by: L λ bp = 0log10 8π( hb hs )( hm hs ) (7) On the other han, when h m h s no breapoint exists. The area near the BS ( < R s ) has a basic propagation loss similar to that of the UHF range, but the area istant from the BS has propagation characteristics in which the attenuation coefficient is cube. Therefore, the approximate lower boun for R s is given by: LLoS, l = Ls + 30 log10 (8) Rs The approximate upper boun for R s is given by: The basic propagation loss L s is efine as:, u = Ls log (9) Rs LLoS 10 L s = λ 0log10 (10) πrs R s in equations (8) to (10) has been experimentally etermine to be 0 m. Base on measurements, a meian value is given by:, m = Ls log (11) Rs LLoS 10

12 10 Rec. ITU-R P Millimetre-wave propagation At frequencies above about 10 GHz, the breapoint istance R bp in equation () is far beyon the expecte maximum cell raius (500 m). This means that no fourth-power law is expecte in this frequency ban. Hence the power istance ecay-rate will nearly follow the free-space law with a path-loss exponent of about.. Attenuation by atmospheric gases an by rain must also be consiere. Gaseous attenuation can be calculate from Recommenation ITU-R P.676, an rain attenuation from Recommenation ITU-R P Moels for NLoS situations NLoS signals can arrive at the BS or MS by iffraction mechanisms or by multipath which may be a combination of iffraction an reflection mechanisms. This section evelops moels that relate to iffraction mechanisms. Propagation for urban area Moels are efine for the two situations escribe in 3.1. The moels are vali for: h b : 4 to 50 m h m : 1 to 3 m f: 800 to MHz to 16 GHz for h b < h r an w < 10 m (or siewal) : 0 to 5000 m. (Note that although the moel is vali up to 5 m, this Recommenation is intene for istances only up to 1 m.) Propagation for suburban area Moel is efine for the situation of h b > h r escribe in 3.1. The moel is vali for: h r : any height m Δh b : 1 to 100 m Δh m : 4 to 10 (less than h r ) m h b : h r + Δh b m h m : h r Δh m m f: 0.8 to 0 GHz w: 10 to 5 m : 10 to 5000 m (Note that although the moel is vali up to 5 m, this Recommenation is intene for istances only up to 1 m.) Millimetre-wave propagation Millimetre-wave signal coverage is consiere only for LoS situations because of the large iffraction losses experience when obstacles cause the propagation path to become NLoS. For NLoS situations, multipath reflections an scattering will be the most liely signal propagation metho.

13 Rec. ITU-R P Propagation over roof-tops for urban area The multi-screen iffraction moel given below is vali if the roof-tops are all about the same height. Assuming the roof-top heights iffer only by an amount less than the first Fresnel-zone raius over a path of length l (see Fig. ), the roof-top height to use in the moel is the average roof-top height. If the roof-top heights vary by much more than the first Fresnel-zone raius, a preferre metho is to use the highest builings along the path in a nife-ege iffraction calculation, as escribe in Recommenation ITU-R P.56, to replace the multi-screen moel. In the moel for transmission loss in the NLoS1-case (see Fig. ) for roof-tops of similar height, the loss between isotropic antennas is expresse as the sum of free-space loss, L bf, the iffraction loss from roof-top to street L rts an the reuction ue to multiple screen iffraction past rows of builings, L ms. In this moel L bf an L rts are inepenent of the BS antenna height, while L ms is epenent on whether the base station antenna is at, below or above builing heights. Lbf + Lrts + Lms for Lrts + Lms > 0 L NLoS1 = (1) Lbf for Lrts + Lms 0 The free-space loss is given by: where: : L bf = log ( / 1 000) + 0 log ( ) (13) path length f f : frequency (MHz). The term L rts escribes the coupling of the wave propagating along the multiple-screen path into the street where the mobile station is locate. It taes into account the with of the street an its orientation. where: L = ) + rts log10 ( w) + 10 log10 ( f ) + 0 log10 ( Δhm Lori (14) ϕ for 0 ϕ < 35 L ori = ( ϕ 35) for 35 ϕ < 55 (15) ( ϕ 55) for 55 ϕ 90 Δ h = h h (16) m L ori is the street orientation correction factor, which taes into account the effect of roof-top-tostreet iffraction into streets that are not perpenicular to the irection of propagation (see Fig. ). The multiple screen iffraction loss from the BS ue to propagation past rows of builings epens on the BS antenna height relative to the builing heights an on the incience angle. A criterion for grazing incience is the settle fiel istance, s : r m λ s = (17) Δh b

14 1 Rec. ITU-R P where (see Fig. ): Δ h = h h (18) b b r For the calculation of L ms, s is compare to the istance l over which the builings exten. The calculation for L ms uses the following proceure to remove any iscontinuity between the ifferent moels use when the length of builings is greater or less than the settle fiel istance. The overall multiple screen iffraction moel loss is given by: ( ) log( ) log bp tanh ( ( ) ) L1ms Lmi + Lmi for l > s an hbp > 0 χ log( ) log( bp ) tanh ( ( ) ) Lms Lmi + Lmi for l s an hbp > 0 χ L ms = Lms ( ) for hbp = 0 (19) log( ) log( bp ) L1ms ( ) tanh ( Lupp Lmi ) Lupp + Lmi for l > s an hbp < 0 ζ log( ) log( bp ) Lms ( ) + tanh ( Lmi Llow ) + Lmi Llow for l s an hbp < 0 ζ where: h bp = L L (0) upp low ζ = L L ) υ (1) ( upp low an L mi upp ( Lupp + Llow = ) () ms ( ) L = L1 (3) low ms bp ( ) L = L (4) bp l bp = Δhb (5) λ υ = [0.0417] χ = [0.1] where the iniviual moel losses, L1 ms () an L ms (), are efine as follows:

15 Rec. ITU-R P Calculation of L1 ms for l > s (Note this calculation becomes more accurate when l >> s.) where: L ( ) L + + log ( / 1 000) + log ( f ) 9 log ( ) 1ms bsh a 10 f b = (6) 18log10(1 + Δhb ) for hb > hr Lbsh = (7) 0 for hb hr is a loss term that epens on the BS height: 71.4 for hb > hr an f > 000 MHz Δhb for hb hr, f > 000 MHz an 500 m Δhb /1 000 for hb hr, f > 000 MHz an < 500 m a = (8) 54 for hb > hr an f 000 MHz Δhb for hb hr, f 000 MHz an 500 m Δhb /1 000 for hb hr, f 000 MHz an < 500 m 18 = Δh hr b for for h h b b > h h 8 for f > 000 MHz = ( f / 95 1) for meium size city an suburban f (30) centres with meium tree ensity an f 000 MHz ( f / 95 1) for metropolitan centres an f 000 MHz r r (9) Calculation of L ms for l < s In this case a further istinction has to be mae accoring to the relative heights of the BS an the roof-tops: where: ( ) ( Q ) L ms = 10 log 10 M (31) QM 0.9 Δh b b.35 λ b = b λ 1 1 π ρ θ π + θ for for for hb > hr +δhu hb hr +δhu hb < hr +δhl an hb hr +δhl (3)

16 14 Rec. ITU-R P an Δh θ = arctan b (33) b ρ = Δh b + b (34) an ( ) b log 10 log 10 b 10 + log10 λ δh = 10 (35) u b 0.187b δh l = b (36) ( log ( f )) Propagation over roof-tops for suburban area A propagation moel for the NLoS1-Case base on geometrical optics (GO) is shown in Fig.. This figure inicates that the composition of the arriving waves at the MS changes accoring to the BS-MS istance. A irect wave can arrive at the MS only when the BS-MS istance is very short. The several-time (one-, two-, or three-time) reflecte waves, which have a relatively strong level, can arrive at the MS when the BS-MS separation is relatively short. When the BS-MS separation is long, the several-time reflecte waves cannot arrive an only many-time reflecte waves, which have wea level besie that of iffracte waves from builing roofs, arrive at the MS. Base on these propagation mechanisms, the loss ue to the istance between isotropic antennas can be ivie into three regions in terms of the ominant arrival waves at the MS. These are the irect wave, reflecte wave, an iffracte wave ominant regions. The loss in each region is expresse as follows base on GO. L NLoS1 4π 0 log10 λ = L0n 3.1 log10 + L RD RD for for for < 0 0 < RD RD (Direct wave ominant region) (Reflecte wave ominant region) (Diffracte wave ominant region) (37) where: L 0n L = L + + L L RD RD L L ( ) ( ) when ( = 0,1,... ) when < < + 1 RD < < RD + 1 (38) 1 = B + ( hb hm ) (39) sin ϕ

17 Rec. ITU-R P πp L = 0 log10 (40) λ RD ( f ) =.65 ( 3 1) log10( f ) ( 0.8 GHz f 0 GHz) (41) LRD = L L L ( ) ( ) RD RD + 1 (4) p 1 = A + ( hb hm ) (43) sin ϕ B A w = w = ( hb hm ) ( + 1) ( h h ) r m ( hb hm ) ( + 1) ( h h ) r m w (44) (45) ϕ = tan 1 B A tan ϕ (46) 4..3 Propagation within street canyons for frequency range from 800 to 000 MHz For NLoS situations where both antennas are below roof-top level, iffracte an reflecte waves at the corners of the street crossings have to be consiere (see Fig. 3). where: L r : L /10 /10 ( 10 r L + ) L NLoS log10 10 = 10 B (47) reflection path loss efine by: where: f ( α) 4π L r = 0 log10 ( x1 + x) + x1x + 0 log10 B (48) w w λ f ( α) = B (49) 3.5 α where 0.6 < α [ra] < π. L : iffraction path loss efine by: 180 4π L = 10 log10[ x1x ( x1 + x) ] + Da α + 0 log10 B (50) π λ

18 16 Rec. ITU-R P x π x + 1 D a = arctan arctan B (51) π w w Propagation within street canyons for frequency range from to 16 GHz The propagation moel for the NLoS situations as escribe in 3.1. with the corner angle α = π/ ra is erive base on measurements at a frequency range from to 16 GHz, where h b < h r an w is up to 10 m (or siewal). The path loss characteristics can be ivie into two parts: the corner loss region an the NLoS region. The corner loss region extens for corner from the point which is 1 m own the ege of the LoS street into the NLoS street. The corner loss (L corner ) is expresse as the aitional attenuation over the istance corner. The NLoS region lies beyon the corner loss region, where a coefficient parameter (β) applies. This is illustrate by the typical curve shown in Fig. 4. Using x 1, x, an w 1, as shown in Fig. 3, the overall path loss (L NLoS ) beyon the corner region (x > w 1 / + 1) is foun using: L = L + L + L (5) NLoS Los c att L c Lcorner log10( x w1 ) w1 + 1 < x w corner = log ( 1+ corner ) (53) 10 Lcorner x > w corner L att x 1 + x 10β log > + + = 10 x w1 1 corner x + w + corner (54) x > w corner where L LoS is the path loss in the LoS street for x 1 (> 0 m), as calculate in 4.1. In equation (53), L corner is given as 0 B in an urban environment an 30 B in a resiential environment. An corner is 30 m in both environments. In equation (54), β is given by 6 in both environments.

19 Rec. ITU-R P FIGURE 4 Typical tren of propagation along street canyons with low base station height for frequency range from to 16 GHz LoS region Relative signal level Corner region NLoS region L LoS Relative signal level L NLoS L corner L att w corner BS x 1 x Distance of travel from base station MS P In a resiential environment, the path loss oes not increase monotonically with istance, an thus the coefficient parameter may be lower than the value in an urban environment, owing to the presence of alleys an gaps between the houses. With a high base station antenna in the small macro-cell, the effects of iffraction over roof-tops are more significant. Consequently, the propagation characteristics o not epen on the corner loss. 4.3 Propagation between terminals locate below roof-top height at UHF The moel escribe below is intene for calculating the basic transmission loss between two terminals of low height in urban environments. It inclues both LoS an NLoS regions, an moels the rapi ecrease in signal level note at the corner between the LoS an NLoS regions. The moel inclues the statistics of location variability in the LoS an NLoS regions, an provies a statistical moel for the corner istance between the LoS an NLoS regions. Figure 5 illustrates the LoS, NLoS an corner regions, an the statistical variability preicte by the moel. This moel is recommene for propagation between low-height terminals where both terminal antenna heights are near street level well below roof-top height, but are otherwise unspecifie. It is reciprocal with respect to transmitter an receiver an is vali for frequencies in the range MHz. The moel is base on measurements mae in the UHF ban with antenna heights between 1.9 an 3.0 m above groun, an transmitter-receiver istances up to m.

20 18 Rec. ITU-R P FIGURE 5 Curves of basic transmission loss not exceee for 1, 10, 50, 90 an 99% of locations (frequency = 400 MHz, suburban) 0 0 Basic transmission loss (B) % 10% 50% 90% 99% Distance P The parameters require are the frequency f (MHz) an the istance between the terminals. Step 1: Calculate the meian value of the line-of-sight loss: L meian LoS ( ) = log f 0 log ( /1 000) (55) Step : For the require location percentage, p (%), calculate the LoS location correction: ( ln(1 p /100) ) ΔL LoS ( p) = 1.564σ with σ = 7 B (56) Alternatively, values of the LoS correction for p = 1, 10, 50, 90 an 99% are given in Table 6. Step 3: A the LoS location correction to the meian value of LoS loss: L LoS meian LoS Step 4: Calculate the meian value of the NLoS loss: (, p) = L ( ) + ΔL ( p) (57) LoS meian NLoS ( ) log f + 40 log10 L = 10 ( /1 000) + L (58) L urban epens on the urban category an is 0 B for suburban, 6.8 B for urban an.3 B for ense urban/high-rise. Step 5: For the require location percentage, p (%), a the NLoS location correction: 1 p urban Δ ( p) = σn ( /100) with σ = 7 B (59) L NLoS N 1 (.) is the inverse normal cumulative istribution function. An approximation to this function, goo for p between 1 an 99% is given by the location variability function Q i (x)

21 Rec. ITU-R P of Recommenation ITU-R P Alternatively, values of the NLoS location correction for p = 1, 10, 50, 90 an 99% are given in Table 6. TABLE 6 Table of LoS an NLoS location variability corrections p (%) ΔL LoS (B) ΔL NLoS (B) LoS Step 6: A the NLoS location correction to the meian value of NLoS loss: L NLoS meian NLoS (, p) = L ( ) + ΔL ( p) (60) Step 7: For the require location percentage, p (%), calculate the istance LoS for which the LoS fraction F LoS equals p: NLoS LoS [ log ( p /100)] ( p) = log10( p /100) LoS ( p) = ( p /100) if p < 45 otherwise (61) Values of LoS for p = 1, 10, 50, 90 an 99% are given in Table 6. This moel has not been teste for p < 0.1%. The statistics were obtaine from two cities in the Unite Kingom an may be ifferent in other countries. Alternatively, if the corner istance is nown in a particular case, set LoS (p) to this istance. Step 8: The path loss at the istance is then given as: a) If < LoS, then L(, p) = L LoS (, p) b) If > LoS + w, then L(, p) = L NLoS (, p) c) Otherwise linearly interpolate between the values L LoS ( LoS, p) an L NLoS ( LoS + w, p): L L LoS NLoS = L LoS = L L(, p) = L ( NLoS LoS LoS ( + ( L, p) LoS NLoS + w, p) L LoS )( The with w is introuce to provie a transition region between the LoS an NLoS regions. This transition region is seen in the ata an typically has a with of w = 0 m. 4.4 Default parameters for site-general calculations If the ata on the structure of builings an roas are unnown (site-general situations), the following efault values are recommene: h r = 3 (number of floors) + roof-height LoS )/ w

22 0 Rec. ITU-R P roof-height = 3 m for pitche roofs = 0 m for flat roofs w = b/ b = ϕ = to 50 m 4.5 Influence of vegetation The effects of propagation through vegetation (primarily trees) are important for outoor short-path preictions. Two major propagation mechanisms can be ientifie: propagation through (not aroun or over) trees; propagation over trees. The first mechanism preominates for geometries in which both antennas are below the tree tops an the istance through the trees is small, while the latter preominates for geometries in which one antenna is elevate above the tree tops. The attenuation is strongly affecte by multipath scattering initiate by iffraction of the signal energy both over an through the tree structures. For propagation through trees, the specific attenuation in vegetation can be foun in Recommenation ITU-R P.833. In situations where the propagation is over trees, iffraction is the major propagation moe over the eges of the trees closest to the low antenna. This propagation moe can be moelle most simply by using an ieal nife-ege iffraction moel (see Recommenation ITU-R P.56), although the nife-ege moel may unerestimate the fiel strength, because it neglects multiple scattering by tree-tops, a mechanism that may be moelle by raiative transfer theory. 5 Builing entry loss Builing entry loss is the excess loss ue to the presence of a builing wall (incluing winows an other features). It is efine as the ifference between the signal levels outsie an insie the builing at the same height. Account must also be taen of the incient angle. (When the path length is less than about 10 m, the ifference in free space loss ue to the change in path length for the two measurements shoul be taen into account in etermining the builing entry loss. For antenna locations close to the wall, it may also be necessary to consier near-fiel effects.) Aitional losses will occur for penetration within the builing; avice is given in Recommenation ITU-R P.138. It is believe that, typically, the ominant propagation moe is one in which signals enter a builing approximately horizontally through the wall surface (incluing winows), an that for a builing of uniform construction the builing entry loss is inepenent of height. Builing entry loss shoul be consiere when evaluating the raio coverage from an outoor system to an inoor terminal. It is also important for consiering interference problems between outoor systems an inoor systems. The experimental results shown in Table 7 were obtaine at 5. GHz through an external builing wall mae of bric an concrete with glass winows. The wall thicness was 60 cm an the winow-to-wall ratio was about :1.

23 Rec. ITU-R P TABLE 7 Example of builing entry loss Frequency Resiential Office Commercial Mean Stanar eviation Mean Stanar eviation 5. GHz 1 B 5 B Mean Stanar eviation Table 8 shows the results of measurements at 5. GHz through an external wall mae of stone blocs, at incient angles from 0 to 75. The wall was 400 mm thic, with two layers of 100 mm thic blocs an loose fill between. Particularly at larger incient angles, the loss ue to the wall was extremely sensitive to the position of the receiver, as evience by the large stanar eviation. TABLE 8 Loss ue to stone bloc wall at various incient angles Incient angle (egrees) Loss ue to wall (B) Stanar eviation (B) Aitional information on builing entry loss, intene primarily for satellite systems, can be foun in Recommenation ITU-R P.679 an may be appropriate for the evaluation of builing entry for terrestrial systems. 6 Multipath moels A escription of multipath propagation an efinition of terms are provie in Recommenation ITU-R P Multipath moels for street canyon environments Omniirectional antenna case Characteristics of multipath elay sprea for the LoS omniirectional antenna case in an urban high-rise environment for ense urban micro-cells an pico-cells (as efine in Table 3) have been evelope base on measure ata at frequencies from.5 to GHz at istances from 50 to 400 m. The r.m.s. elay sprea S at istance of m follows a normal istribution with the mean value given by: a s = C a ns (6) a γ an the stanar eviation given by: σ = C γ σ s σ ns (63) where C a, γ a, C σ an γ σ epen on the antenna height an propagation environment. Table 9 lists some typical values of the coefficients for istances of m base on measurements mae in urban an resiential areas.

24 Rec. ITU-R P TABLE 9 Typical coefficients for the istance characteristics of r.m.s. elay sprea for omniirectional antenna case Measurement conitions a s σ s (1) () Area f (GHz) h b h m C a γ a C σ γ σ Urban (1) Urban () Resiential () Threshol value of 0 B is use for r.m.s. elay sprea calculation. Threshol value of 30 B is use for r.m.s. elay sprea calculation. From the measure ata at.5 GHz, the average shape of the elay profile was foun to be: where: an t is in ns. P 0 : τ : pea power (B) ecay factor ( τ 1) P ( t) = P + 50 e t/ B (64) From the measure ata, for an r.m.s. elay sprea S, τ can be estimate as: 0 τ = 4 S + 66 ns (65) A linear relationship between τ an S is only vali for the LoS case. From the same measurement set, the instantaneous properties of the elay profile have also been characterize. The energy arriving in the first 40 ns has a Rician istribution with a K-factor of about 6 to 9 B, while the energy arriving later has a Rayleigh or Rician istribution with a K-factor of up to about 3 B. (See Recommenation ITU-R P.1057 for efinitions of probability istributions.) 6.1. Directional antenna case In fixe wireless access systems an communications between the access points of wireless mesh networ systems, irectional antennas are employe as transmitter an receiver antennas. A typical effect of the use of irectional antennas is given hereafter. Arriving elaye waves are suppresse by the antenna pattern using irectional antennas as the transmitter an receiver antennas. Therefore, the elay sprea becomes small. In aition, the receive power increases with the antenna gain, when irectional antennas are employe as the transmitter an receiver antennas. Base on these facts, the irectional antenna is use in wireless systems. Therefore, it is important to unerstan the effect of antenna irectivity in multipath moels.

25 Rec. ITU-R P Characteristics of the multipath elay sprea for the LoS irectional antenna case in an urban high-rise environment for ense urban micro-cells an pico-cells (as efine in Table 3) were evelope base on measure ata in the 5. GHz ban at istances from 10 to 500 m. The antennas were configure such that the irection of the maximum antenna gain of one antenna face that of the other. Table 10 lists equation for eriving coefficients relative to the antenna half power beamwith for formula (58) for istances of m base on measurements in an urban area. These equations are only epening on the antenna half power beamwith an effective to any with of the roa. TABLE 10 Typical coefficients for the istance characteristics of r.m.s. elay sprea for irectional antenna case Measurement conitions a s Area f (GHz) h b h m Urban log(θ) θ 10 NOTE 1 Threshol value of 0 B is use for r.m.s. elay sprea calculation. Here, θ represents antenna half-power beamwith at both transmitting an receiving antenna an the unit is raian. Note that θ shoul be set to π when omniirectional antenna is applie to both transmitting an receiving antenna. 6. Multipath moels for over-rooftops propagation environments Characteristics of multipath elay sprea for both LoS an NLoS case in an urban high-rise environment for micro-cells (as efine in Table 3) have been evelope base on measure ata at MHz, MHz, an MHz using omniirectional antennas. The meian r.m.s. elay sprea S in this environment is given by: ( A L B) C a S u = exp + ns (66) where both A an B are coefficients of r.m.s. elay sprea an L is path loss (B). Table 11 lists the typical values of the coefficients for istances of 100 m 1 m base on measurements mae in urban areas. TABLE 11 Typical coefficients for r.m.s. elay sprea γ a Measurement conitions Coefficients of r.m.s. elay sprea Area Urban Frequency (GHz) Range MHz MHz, MHz A B

26 4 Rec. ITU-R P The istributions of the multipath elay characteristics for the 3.7 GHz ban in an urban environment with each BS antenna height of 40 m an 60 m, an MS antenna height of m were erive from measurements. The istributions of the multipath elay characteristics for the 3.7 GHz an 5. GHz ban in a suburban environment with a BS antenna height of 0 m, an MS antenna height of.0 m an.8 m were erive from measurements. Table 1 lists the measure r.m.s. elay sprea for the 3.7 GHz an 5. GHz ban for cases where the cumulative probability is 50% an 95%. TABLE 1 Typical r.m.s. elay sprea values* Measurement conitions r.m.s. elay sprea (ns) Area Urban very high-rise Scenario Frequency (GHz) Antenna height h BS h r Range 50% 95% LoS NLoS Urban 3.7 Suburban * Threshol value of 30 B was use for r.m.s. elay sprea calculation. 7 Number of signal components For the esign of high ata rate systems with multipath separation an synthesis techniques, it is important to estimate the number of signal components (that is, a ominant component plus multipath components) arriving at the receiver. The number of signal components can be represente from the elay profile as the number of peas whose amplitues are within A B of the highest pea an above the noise floor, as shown in Fig. 6.

27 Rec. ITU-R P FIGURE 6 Definition for the etermination of the number of peas Receive power level A Noise floor Time elay P Table 13 shows the results for the number of signal components from measurements in ifferent scenarios for ifferent antenna heights, environments an for ifferent frequencies. TABLE 13 Maximum number of signal components Type of environment Time elay resolution Frequency (GHz) Antenna height Range Maximum number of components h b h m 3 B 5 B 10 B 80% 95% 80% 95% 80% 95% Urban 00 ns Suburban 175 ns Urban 0 ns Resiential 0 ns Suburban 175 ns Suburban 50 ns Suburban 100 ns Urban 0 ns Urban 0 ns

28 6 Rec. ITU-R P For the measurements escribe in 6., the ifferential time elay winow for the strongest 4 components with respect to the first arriving component an their relative amplitue is given in Table Polarization characteristics Cross-polarization iscrimination (XPD), as efine in Recommenation ITU-R P.310, iffers between LoS an NLoS areas in an SHF ense urban micro-cellular environment. Measurements inicate a meian XPD of 13 B for LoS paths an 8 B for NLoS paths, an a stanar eviation of 3 B for LoS paths an B for NLoS paths at SHF. These meian values are compatible with the UHF values for open an urban areas, respectively, in Recommenation ITU-R P.1406.

29 Rec. ITU-R P Type of environment BS antenna Frequency (GHz) Antenna height TABLE 14 Range Maximum number of signal components h b h m A = 3 B A = 5 B A = 10 B 80% 95% 80% 95% 80% 95% Urban Low Urban Low Urban Low Urban High Resiential Low Suburban High TABLE 15 Differential time elay winow for the strongest 4 components with respect to the first arriving component an their relative amplitue Type of environment Time elay resolution Frequency (GHz) Antenna height Range Excess time elay (μs) h b h m 1st n 3r 4th 80% 95% 80% 95% 80% 95% 80% 95% Urban 00 ns Relative power with respect to strongest component (B)

30 8 Rec. ITU-R P Characteristics of irection of arrival The r.m.s. angular sprea as efine in Recommenation ITU-R P.1407 in the azimuthal irection in a ense urban micro-cell or picocell environment in an urban area was obtaine from the measurement mae at a frequency of 8.45 GHz. The receiving base station ha a parabolic antenna with a half-power beamwith of 4. Also measurement performe at the ense urban micro-cell environment in urban area. Angular sprea coefficients are introuce base on measurements in urban areas for istances of 10~1 000 m, uner the LoS cases at a frequency of GHz. Four elements omniirectional linear array with Bartlett beam-forming metho is use for eriving the angular profile. The coefficients for r.m.s. angular sprea were obtaine as shown in Table 16. Area TABLE 16 Typical coefficients for the istance characteristics of angular sprea Measurement conitions f (GHz) h b h m Mean (egree) s.t. (egree) Remar Urban LoS Urban LoS Urban NLoS 10 Faing characteristics The faing epth, which is efine as the ifference between the 50% value an the 1% value in the cumulative probability of receive signal levels, is expresse as a function of the prouct (ΔfΔL max MHz m) of the receive banwith Δf MHz an the maximum ifference in propagation path lengths ΔL max m as shown in Fig. 7. ΔL max is the maximum ifference in propagation path lengths between components whose level is larger than the threshol, which is 0 B lower than the highest level of the inirect waves as shown in Fig. 8. In this figure, a in ecibels is the power ratio of the irect to the sum of inirect waves, an a = B represents a NLoS situation. When ΔfΔL max is less than 10 MHz m, the receive signal levels in LoS an NLoS situations follow Rayleigh an Naagami-Rice istributions, corresponing to a narrow-ban faing region. When it is larger than 10 MHz m, it correspons to a wieban faing region, where the faing epth becomes smaller an the receive signal levels follow neither Rayleigh nor Naagami-Rice istributions.

31 Rec. ITU-R P FIGURE Relationship between faing epth an ΔfΔL max a = B 0 B 3 B Faing epth (B) B 7 B 10 B 13 B 0 B a: Power ratio ΔΔ f L max (MHz m) P Power FIGURE 8 Moel for calculating ΔL max ΔL max Direct wave Inirect highest wave 0 B Cut-off level Delay P Propagation ata an preiction methos for the path morphology approach 11.1 Classification of path morphology In the populating area except rural area, the path morphology for wireless channels can be classifie into 9 categories as shown in Table 17. The classification is fully base on real wave-propagation environment, by analysing builing height an ensity istribution for various representative locations using GIS (Geographic Information System) atabase.

32 30 Rec. ITU-R P TABLE 17 Classification of path morphologies for the MIMO channel High rise (above 5 m) Mile rise (1 m ~ 5 m) Low rise (below 1 m) Path morphology ensity High ensity (HRHD) above 35% Mile ensity (HRMD) 0 ~ 35% Low ensity (HRLD) below 0% High ensity (HRHD) above 35% Mile ensity (HRMD) 0 ~ 35% Low ensity (HRLD) below 0% High ensity (HRHD) above 35% Mile ensity (HRMD) 0 ~ 35% Low ensity (HRLD) below 0% 11. Statistical moelling metho Usually the measurement ata are very limite an not comprehensive. Therefore, for specific morphologies an specific operating frequencies, the following metho can be use to erive the parameters for the MIMO channel moel. Measurements of channel characteristics for 9 typical morphologies at GHz have shown goo statistical agreement when compare against moelling metho. Moels are efine for the situation of h b > h r. Definitions of the parameters f,, h r, h b, Δh b an h m are escribe in 3.1, an B represents builing ensity. The path morphology approach is vali for: f: 800 to MHz : 100 to 800 m h r : 3 to 60 m h b : h r + Δ h b Δ h b : up to 0 m h m : 1 to 3 m B : 10 to 45% In the statistical moelling, the builings are generate in a fully ranom fashion. It is well nown that the istribution of builing height h is well fitte statistically by Rayleigh istribution P(h) with the parameter μ. h h P ( h) = exp( ) (67) μ μ To erive the statistical parameters of Rayleigh istribution for given morphology, the use of available GIS atabase is recommene. For the horizontal positions of builings, it can be assume to be uniformly istribute. The wave-propagation calculation is performe for each realization of builing istribution using the ray tracing metho. 15 times reflection an times iffraction are recommene for simulation. Penetration though builings are also important. It is recommene to set up the receiving power threshol properly to consier the builing penetration. To obtain the moel parameters, simulations

33 Rec. ITU-R P shoul be performe for enough number of realizations for each morphology. At least 4 times realization is recommene. For each realization, enough number of receivers shoul be put in the calculation region, in orer to obtain statistically meaningful ata. It is recommene that at least 50 receivers are available at each 10 m sub-interval of istance. The transmitting antenna height an the receiving antenna shoul be set at the appropriate values. It is recommene that the values of ielectric constant an conuctivity are set at ε r = 7.0, σ = S/m for builings, an ε r =.6, σ = 0.01 S/m for grouns. The parameter values of builing height istribution for typical cases are given in Table 18. Builing sizes are 30 0 m, 5 0 m, an 0 0 m for high, mile an low rise. Builing ensities are 40%, 30%, an 0% for high, mile an low ensity. TABLE 18 Parameters of builing height istribution for statistical moelling Path morphology Rayleigh parameter μ Range of builing height istribution Average builing height HRHD 1.3~ HRMD ~ HRLD 13.~ MRHD 7.3~ MRMD 10 7.~ MRLD 7.4~ LRHD.1~ LRMD 6.5~. 9.4 LRLD.5~ Path loss moel The path loss moel in this recommenation is given by: PL = PL n log10( ) + S (B) (68) PL0 = log10( f ) (B) (69) where n is the path loss exponent. S is a ranom variable representing the ranom scatter aroun the regression line with normal istribution, an the stanar eviation of S is enote as σ s. The units of f an are MHz an metres, respectively. The path loss parameters for typical cases of 9 path morphologies from statistical moelling at GHz are summarize in Table 19. The values in the Table are fitte for all receivers at the height of m locate along the path at istances from 100 m to 800 m.