Evaluating the effect of antenna tilt and rotation on antenna performance in an indoor environment.

Size: px

Start display at page:

Download "Evaluating the effect of antenna tilt and rotation on antenna performance in an indoor environment."

Suzan Banks
5 years ago
Views:

1 Evaluatig the effect of atea tilt ad rotatio o atea performace i a idoor eviromet. Geoffrey Hilto, Evagelos Mellios, Di Kog, David Halls ad Adrew Nix Cetre for Commuicatios Research, Uiversity of Bristol Merchat Veturer s Buildig, Woodlad Road, Bristol, UK geoff.hilto@bristol.ac.uk Abstract I this paper the relative performaces of atea elemets are assessed by combiig the measured full (3D) farfield radiatio patters of each with measured agle of arrival data for a laboratory/office space. Three elemet types (Cavitybacked Slot, Prited Iverted-F ad Dielectric Resoator) at two positios o a small mobile termial are cosidered, ad the termial is tilted ad rotated i order to determie the variatio i performace that is obtaied for three trasmitter locatios. Results are preseted as cumulative distributio fuctios of directivity relative to a bechmark Hertzia dipole for operatio at 5.GHz. A variatio i sigal stregth of greater tha 30dB was observed due to both the directivity of the elemets ad their polarisatio aligmet with the vertical trasmittig source. I. INTRODUCTION I the aalysis of atea performace, it is importat that there is a complete kowledge of the atea s far-field radiatio patter i terms of magitude, phase ad polarisatio for all agles. This is especially importat whe dealig with the way the user operates a termial. For istace, the way a mobile hadset is tilted ad rotated withi the eviromet i the course of ormal operatio. Furthermore, i order to attempt to aalyse the performace i a multipath eviromet, the sigal agles of arrival/departure (AoA/AoD) also have to be kow i terms of magitude ad phase for vertical ad horizotal polarised compoets. While it may be cosidered that phase is less importat tha the magitude, it is importat i ascertaiig whether a sigal is, say, liearly or circularly polarised ad whether icidet sigals are i phase or out of phase i both istaces these will result i cosiderable variatios i received sigal stregth. There are a variety of methods of aalysig the performace of ateas i multipath eviromets such as Mea Effective Gai (MEG), Total Radiated Power (TRP) ad Total Isotropic Sesitivity (TIS) []-[5]. While 3D patter measuremets are required for full aalysis, Over The Air testig (OTA) by subjectig the atea uder test (AUT) to simulate multipath fadig usig a umber of radiatig sources surroudig the AUT [6], [7] or iside a reverberatio chamber [8], [9] have also bee used to yield comparative performace figures for ateas. I this paper 3D atea radiatio patters ad AoA data for a typical laboratory/office eviromet are combied to evaluate the performace of various ateas at 5.GHz. Both sets of data are measured ad both cotai magitude, phase ad polarisatio iformatio for all agles. The mai aim is to show the variatio i performace of a umber of atea types mouted o a termial remiiscet of a small hadheld uit while the termial is tilted ad rotated, Figure. Three types of atea (Prited Iverted-F, Dielectric Resoator ad Prited Cavity-backed Slot) are mouted i two positios (vertical ad horizotal arragemets) for the tests - for the Slot ad Dielectric Resoator (DRA) elemets this was o the top ad o the side as idicate i Figure, while for the Prited Iverted-F elemets (IFA) these were cotaied o a sigle PCB. The patter data used here is a subset of data from measuremets by Pal et al [0] i which 4 elemets were mouted o the termial of dimesios cm high by 6cm wide by.4cm thick. While the origial work cosidered a MIMO scheme, this aalysis assumes SISO operatio. The patter data does ot iclude the effects of body shadowig, as used for istace i [], ad further measuremets will be udertake to ascertai the how shadowig affects overall system performace. All results preseted here are bechmarked agaist a theoretical (ideal) Hertzia dipole. Fig.. Atea elemet positios o termial ad elevatio tilts. For this aalysis it has bee assumed that the termial would be tilted i the x-z plae betwee 0 ad 60 i order to see the scree ad betwee 5 ad 45 i order to hold ear the head, whilst udergoig full 360 degree rotatio i azimuth. Agular resolutio is 0 for all rotatios ad equal weightig has bee assumed for all potetial orietatios. The result of this is that cumulative distributio fuctios (cdf) are produced showig comparative atea performaces for differet eviromets ad importatly the affect of polarisatio misaligmet o the atea performace //$ IEEE

II. ANGLE OF ARRIVAL MEASUREMENTS Figure shows the pla of the laboratory ad adjacet office space with trasmitter locatios idicated.

moopole o a 60cm diameter circular groud plae - the measuremets preseted here are oly for vertical trasmit polarisatio at 5.GHz. Both trasmit ad receive ateas were at a height of.

I Figure 3, it ca be see that there are a few of strog peaks with several smaller peaks, while i Figure 4 the red ad yellow areas are of mai importace with the peaks ot occurig at 0 i elevatio (hece

A peak search routie was used to process this data, ad after applyig a threshold o the data sets, up to 00 AoA paths were used for the subsequet atea performace tests.

$As idicated i Figure, there were a small umber of strog sigals for each locatio resultig from predomiatly lie of sight (LoS) or diffractio (locatios 6 ad 0) ad reflectios (locatio ).$ $For istace with locatio 0 (i blue), there is a sigal directly through the wall but the mai sigal is diffracted from the door opeigs betwee laboratory ad corridor, while for the trasmitter i locatio,$

2 II. ANGLE OF ARRIVAL MEASUREMENTS Figure shows the pla of the laboratory ad adjacet office space with trasmitter locatios idicated. Of iterest i this paper are locatio 6 (i laboratory ad i direct lie of sight), locatio 0 (i kitche area with o visible lie of sight) ad locatio (a aechoic chamber with oe door left ope for illumiatig the laboratory space). The receive atea was a dual polarised Fla DP40 hor mouted o a positioer that performed full azimuthal scas of all elevatios for vertical ad horizotal polarisatios, while the trasmitter was a vertical moopole o a 60cm diameter circular groud plae - the measuremets preseted here are oly for vertical trasmit polarisatio at 5.GHz. Both trasmit ad receive ateas were at a height of.4m above the floor ad coected to a Arisu 37397C VNA. Bechig i Laboratory of data is ormalised to the maximum received sigal at each locatio. I Figure 3, it ca be see that there are a few of strog peaks with several smaller peaks, while i Figure 4 the red ad yellow areas are of mai importace with the peaks ot occurig at 0 i elevatio (hece would have bee missed usig purely azimuth-sca data). A peak search routie was used to process this data, ad after applyig a threshold o the data sets, up to 00 AoA paths were used for the subsequet atea performace tests. The most sigificat of these paths has bee projected oto Figure for locatios 6, 0 ad ad colour coded for each trasmitter. The bold lies idicate directios of strog sigals, whilst dotted lies are of secodary importace. As idicated i Figure, there were a small umber of strog sigals for each locatio resultig from predomiatly lie of sight (LoS) or diffractio (locatios 6 ad 0) ad reflectios (locatio ). For istace with locatio 0 (i blue), there is a sigal directly through the wall but the mai sigal is diffracted from the door opeigs betwee laboratory ad corridor, while for the trasmitter i locatio, there is a cocetratio towards bechig from where the sigal is beig scattered, while there are weak sigals from diffractio aroud a door ad reflectio from a support pillar i the cetre of the laboratory. Aechoic chamber 8m Receiver Fig. Locatios of trasmitters ad receiver i the laboratory ad office. (a) Locatio 6 (b) Locatio 0 (c) Locatio (d) Locatio 0 (differece) (b) Locatio 0 (b) Locatio Fig 3. AoA data projected oto the x-y plae. Liear power plots. Figure 3 shows the AoA data i polar format for locatios 0 ad as projected oto the x-y plae, while Figure 4 shows the power spectrum of the full sca for received vertical polarisatio (trasmit vertical polarisatio). Each set Relative power levels for trasmit vertical ad receive vertical. Red is -db ad yellow is -3dB. Fig 4. Agle of arrival data for the trasmitter locatios. The maximum horizotally received sigals were -8.3dB for locatio 6, -5.3dB for locatio 0 ad -3.5dB for locatio, which idicates a higher degree of scatterig from trasmitter locatios where the domiat sigals are reflected.

3 This AoA data is also a fuctio of the beamshape of the hor atea (half power beamwidths of 30 ad 33 i the pricipal plaes ad directivity of 4.5dBi) as this will gover the agular resolutio betwee sources of radiatio. Furthermore, the blue ripple i the data is also due partly to the characteristics of the hor atea away from its mai beam rather tha reflected sigals from withi the eviromet. It may be possible to remove the atea patter-depedece usig a techique described i []. Two measuremets were recorded (oe directly after the other) at all locatios i order to show repeatability i terms of phase as well as magitude. Whilst there was some movemet of people i the laboratory ad office space durig measuremets the affect o the measuremets ca oly really be see by subtractig the two measuremets. This is show i Figure 4(d) for locatio 0 where the red/yellow areas show movemet i the kitche ad corridor outside the laboratory the plots show here are relative to the maximum, ad the maximum level is actually some 6dB below the maximum level show i Figure 4(b). Away from this area, there is cosiderable phase stability for this effectively static eviromet. TABLE I MAXIMUM DIRECTIVITY LEVELS FOR THE ELEMENTS Slot DRA IFA Ideal dipoles Top Side Top Side Top Side Jz Jy Uits i dbi By usig three differet elemets ad orthogoallymoutig them, a rage of patter coverage ad polarisatios were obtaied. Eve amogst the same elemet type, the directivity varies due to the moutig arragemet o the termial with the largest differece by far beig with the IFA - these values are all cosiderably higher tha the ideal (Hertzia) dipole beig used as a bechmark. The Jz source is aliged alog the z-axis, omidirectioal i the azimuth plae, vertically polarised (o horizotal compoet) ad with the patter as show i Figure 5(d), while the Jy source is the atea rotated to alig alog the y-axis ad this splits the polarisatio ito both vertical ad horizotal compoets. III. ANTENNA RADIATION PATTERN MEASUREMENTS Far-field radiatio patters for the three types of elemets i two positios produced six uique sets of data (complex E- fields for both polarisatios for all agles). Figure 5 shows the power patters for oe of each atea type (together with the bechmarkig Hertzia dipole), while Table shows the (maximum) directivity obtaied through itegratio of the Poytig vector for each atea radiatio patter. (a) Cavity-backed Slot (b) DRA Vertical polarisatio, E, Horizotal polarisatio, E Upper to lower figures: 5 tilt from vertical, 30 tilt & 45 tilt Fig 6. Effect of tilt o ideal source polarisatio compoets. (c) IFA (d) Hertzia dipole Fig. 5. Far-field radiatio patters for oe of each type of atea (Total power patters) I Figure 6, the effect that the tiltig of the termial i the x-z plae has o the polarisatio compoets of the atea is demostrated usig the ideal (z-directed) Hertzia dipole, show i Figure 5(d), with the ull i the co-polar patter, 3

4 E, rotatig to the tilt agle ad a icrease i the level of cross-polarisatio, E. IV. COMBINATION OF AOA AND ANTENNA DATA For this assessmet of atea performace, the total received power is ot required. Istead a comparative measure of the various atea types ad moutig orietatio o the termial was performed. The total (complex) field respose of the atea, g, is therefore combied with the correspodig AoA field respose, C, ad polarisatio efficiecy (mismatch),, (based upo [3]) for all agles of arrival,. The ormalisatio is with respect to a Hertzia dipole, d, whose orietatio is set up to correspod to the polarisatio of the trasmittig source (vertical), Equatio. Sice the atea directivity, D r(orm), is ormalised with respect to a dipole it ca be expressed i dbd. (a) Locatio 6 maily LoS with D g N = r( orm) = 0log N ( θ, φ ) = * g C Γ d C Γ * dbd () = D () For each atea with its tilt ad rotatio variatios, a ew value of D r (orm) is calculated ad the results for the three scearios are summarised i Figures 7 as cumulative distributio fuctios (cdf). From these plots, it is possible to judge ot oly the average performace of a atea cofiguratio but also to ascertai how its performace may vary betwee propagatio eviromets. The mai observatios that ca be draw from these results are: There is clearly a wide variatio of performace with various combiatios of elemet type, positio o termial ad eviromets with over 30dB variatio i sigal stregth. Takig the -5dB level o the horizotal axis ad the IFA resposes, it ca be see that with the LoS sceario (a) the probability of exceedig this level is 60%, whilst with the diffracted sceario (b) this has falle to aroud 0%. Furthermore, the differece i performace betwee elemets is more oticeable with a high LoS cotet. For istace, i (c) at the 50% probability level, there is 4.5dB variatio i sigal level, which has rise to 7.5dB with (a). The Jz source geerally outperforms the practical elemets for the majority of the orietatios as it is omidirectioal i the azimuth plae ad hece does ot suffer from a large umber of orietatios where the more directive elemets are poitig i the wrog directio. Where the beam is poitig i the correct directio ad with the correct (vertical) polarisatio, for istace the Side-mouted DRA (show as black dashed lie), the there is a cosiderable sigal advatage as see i (a) ad (c). Hece the roll off i performace for the Jz source is oly a factor of the atea s elevatio beamwidth ad tilt agle the wider the beamwidth, the steeper the roll off. (b) Locatio 0 maily diffracted (c) Locatio maily reflected For the Slot, DRA & IFA the solid lie is the Top Atea ad the dashed lie is the Side Atea i Figure. Fig 7. Cumulative distributios fuctios for tilted/rotated ateas with the three trasmitter locatios. Havig the wrog polarisatio will sigificatly reduce performace. This ca be see i (a) where the Jy source ad the side-mouted slot (dashed gree lie) are sigificatly lower tha all the others (ad 0-5dB below the 4

5 bechmark) as the majority of the sigal comes directly from a vertically-polarised source. However, oce there are more multi path compoets from higher/lower elevatios ad/or scatterig ito the horizotal polarisatio, (b) ad (c), the sidemouted slot is ot sigificatly worse tha ay other elemets. Based upo these three trasmitter locatios, it would appear that overall the IFAs give a geerally better performace over the wider rage of tilts ad rotatios of the termial (plots are to the right i Figure 7), while the DRAs give the poorer performace (plots geerally o the left). It ca be see from the radiated power patter for the IFA, Figure 5(c), that of all elemets tested, this most resembles the ideal case, Figure 5(d), alog the x-y plae. V. CONCLUSIONS This paper has take full measured far-field radiatio patter data for a umber of atea elemets ad combied this with measured AoA data i azimuth ad elevatio i order to evaluate the atea elemets whe tilted ad rotated. AoA data for a umber of trasmitter locatios withi a idoor eviromet are cosidered, ad these show that with o direct (visible) LoS, the sigals reachig the receiver are domiated by a few diffracted paths. Levels of sigal (origially vertically polarised) scattered ito the horizotal polarisatio icreased (relative) magitude from a maily LoS sceario (-8.3dB) to reflected ad o LoS (-3.5dB). Phase stability is also good i this predomiatly static eviromet. Radiatio patters (i terms of magitude, phase ad polarisatio for all agles) have bee preseted for three atea types (Slot, DRA ad IFA) ad two orthogoal moutig positios o the termial, ad has show there is sigificat variatio i polarisatio ad patter coverage that results i directivities ragig from 5.9 to 8.8dBi. While the ateas were rotated 360 i azimuth (x-y plae), the tiltig i the x-z ad y-z plaes were more restricted correspodig to the most usual operatio of a hadset. Equal weightig was applied to each set of agles, but differetial weightig scheme could be used to bias the outcomes to the more commo termial orietatios. Variatios i received sigal stregth of greater tha 30dB with respect to a Hertzia dipole bechmark were observed due to polarisatio misaligmet ad patter directivity. At best the directive ateas oly achieved greater tha 0dBd levels for 0% of the orietatios cosidered this was maily as a result of the rotatio i azimuth poitig the directive beam i the wrog directio. I the case of the sidemouted slot, the polarisatio was orthogoal to the trasmitter polarisatio ad hece was 0-5dB below the bechmark. For the trasmitter locatios, atea type ad termial moutigs cosidered, the IFA performace was the best o average, though was still lower tha the bechmark. Further measuremets usig a horizotally polarised source will be udertake to show how the choice of trasmit polarisatio affects system performace. It is clear that i order to fully uderstad the operatio of ateas i multipath eviromet, full patter ad AoA data is required that cotais magitude ad phase for both polarisatios ad for all agles. Thus i the future, polarisatio ad patter agile atea systems ca be evaluated i order to determie how best to mout ad operate elemets i order to compesate for user operatio ad adapt the system for differet multipath eviromets. ACKNOWLEDGMENT The authors wish to thak the Cetre for Commuicatio Research, Uiversity of Bristol, for the use of the aechoic chamber ad extesive laboratory facilities. REFERENCES [] T. Taga, Aalysis for mea effective gai of mobile ateas i lad mobile eviromets, IEEE Tras. Vehicular Techology, vol. 39, pp. 7-3, May 990. [] J-H. Choi, Y-S.Shi, K-B. Kog ad S-O. Park, 3D Spherical Atea Measuremet System for CTIA OTA Testig ad Characterizig of Mobile MIMO Termials, 7th Iteratioal Symposium o Ateas, Propagatio & EM Theory, ISAPE '06, pp. - 4, Oct 006 [3] J.O. Nielse, G.F. Pederse, Mobile hadset performace evaluatio usig spherical measuremets, Proc.IEEE 56th Vehicular Techology Coferece, VTC 00-Fall. 00, vol., pp , 00. [4] K. Kalliola, K. Suloe, H. Laitie, O. Kivekäs, J. Krogerus, ad P. Vaiikaie, Agular Power Distributio ad Mea Effective Gai of Mobile Atea i Differet Propagatio Eviromets, IEEE Tras. Vehicular Techology, vol. 5, o. 5, pp , Sept. 00. [5] M.B. Kudse ad G.F. Pederse, Spherical Outdoor to Idoor Power Spectrum Model at the Mobile Termial, IEEE Joural o selected areas i commuicatios, vol. 0, o. 6, pp , Aug. 00. [6] J.T. Toivae, T.A. Laitie, V-M. Kolmoe ad P. Vaiikaie Reproductio of Arbitrary Multipath Eviromets i Laboratory Coditios, IEEE Tras. Istrumetatio ad Measuremet, vol. 60, o., pp 73-8, Ja. 0. [7] T. Laitie, J. Toivae, P. Kyosti, J Nuutie, ad P. Vaiikaie, O MIMO-OTA testig based o multi-probe techology, URSI Iteratioal Symposium o Electromagetic Theory (EMTS), pp. 7-30, Nov. 00 [8] P.-S. Kildal, Overview of 6 Years R&D o Characterizig Wireless Devices i Rayleigh Fadig Usig Reverberatio Chambers, Iteratioal Workshop o Atea Techology: Small ad Smart Ateas Metamaterials ad Applicatios, IWAT '07, pp. 6-65, 007. [9] L. Garcia-Garcia, B. Lidmark, N. Jalde ad C. Orleius, MIMO capacity of atea arrays evaluated usig radio chael measuremets, reverberatio chamber ad radiatio patters, IET Microwave, Ateas ad Propagatio, vol., issue 6. pp , 007 [0] A. Pal, C. Williams, G. Hilto, ad M. Beach, Evaluatio of Diversity Atea Desigs Usig Ray Tracig, Measured Radiatio Patters, ad MIMO Chael Measuremets, EURASIP Joural o Wireless Commuicatios ad Networkig, Article ID 58769, 007. [] F. Harrysso, J. Medbo, A.F. Molisch, A.J. Johasso, ad F. Tufvesso, Efficiet Experimetal Evaluatio of a MIMO Hadset with User Ifluece, IEEE Tras. Wireless Commuicatios, vol. 9, o., pp , Feb 00. [] D. Bouhired, Atea Performace Aalysis Combiig Full Radiatio Patters ad Directioal Chael Characterisatio, PhD thesis, Uiversity of Bristol, UK, Nov. 00, Ch. 4 [3] T. A. Milliga, Moder atea desig, d ed, Joh Wiley & Sos, New Jersey, USA, 005, Ch 5

APPLICATION NOTE UNDERSTANDING EFFECTIVE BITS

APPLICATION NOTE UNDERSTANDING EFFECTIVE BITS APPLICATION NOTE AN95091 INTRODUCTION UNDERSTANDING EFFECTIVE BITS Toy Girard, Sigatec, Desig ad Applicatios Egieer Oe criteria ofte used to evaluate a Aalog to Digital Coverter (ADC) or data acquisitio