Network-Based Wireless Location · 2018-06-21 · Network-based location technology, on the other hand, relies on some existing networks (either cellular or WLAN) to determine the

© DIGITALVISION

Wireless location refers to the geographic coordinates of a mobile subscriber incellular or wireless local area network (WLAN) environments. Wireless loca-tion finding has emerged as an essential public safety feature of cellular sys-tems in response to an order issued by the Federal CommunicationsCommission (FCC) in 1996. The order mandated all wireless service providers

to deliver accurate location information of an emergency 911 (E-911) caller to public safetyanswering points (PSAPs). The FCC mandate aims to solve a serious public safety problem causedby the fact that, at present, a large proportion of all 911 calls originate from mobile phones, thelocation of which cannot be determined with existing technology. However, many difficultiesintrinsic to the wireless environment make meeting the FCC objective challenging; these chal-lenges include channel fading, low signal-to-noise ratios (SNRs), multiuser interference, and mul-tipath conditions. In addition to emergency services, there are many other applications for wirelesslocation technology, including monitoring and tracking for security reasons, location sensitivebilling, fraud protection, asset tracking, fleet management, intelligent transportation systems,mobile yellow pages, and even cellular system design and management. This article provides anoverview of wireless location challenges and techniques with a special focus on network-basedtechnologies and applications.

IEEE SIGNAL PROCESSING MAGAZINE [24] JULY 2005 1053-5888/05/$20.00©2005IEEE

Network-Based Wireless Location

[Challenges faced in developing techniques

for accurate wireless location information]

[Ali H. Sayed, Alireza Tarighat,

and Nima Khajehnouri]

WIRELESS NETWORKSWireless networks are primarily designed for voice and datacommunications. The widespread availability of wireless nodes,however, makes it possible to utilize these networks for wirelesslocation purposes as well. It is expected that location-basedapplications will play an important role in future wireless mar-kets. While location services are now driven by emergency andsecurity requirements imposed on the wireless networks, in thefuture they will be driven by commercial demands for location-motivated products. Increasingly, application-level software willincorporate location information into its features to fully utilizesuch information once it becomes available. For example, assettracking and management software would incorporate locationinformation into a database for enhanced tracking capabilities.As such, wireless location information will add a new dimensionto future applications.

Wireless networking devices constitute the main infrastruc-ture to be utilized for wireless location finding. A location find-ing system should be able to seamlessly use both cellular andWLANs for location finding by roaming between the networks.The result would be transparent location coverage for both out-door and indoor environments. Today, the main commerciallydeployed wireless location finding system is linked to the cellu-lar network in response to requirements by the FCC for emer-gency 911 calls made through cell phones. These requirementsare collectively known as the enhanced 911 (E911) mandate.The details of the FCC requirements for E911 will be discussed.

The purpose of this article is to provide an overview of thebasic challenges facing the wireless techniques that are beingdeveloped for accurate location information. We start with anoverview of the main applications that serve as the major drivingforce behind the technology.

For ease of reference, Table 1 collects the acronyms that arecommon in this field and used extensively in subsequent sections.

APPLICATIONSFigure 1 illustrates some of the available market forecasts forwireless location technology [1], [2]. It is estimated that loca-tion-based services (LBSs) will generate annual revenues of theorder of US$15 billion worldwide. In the United States alone,about 170 million mobile subscribers are expected to becomecovered by the FCC-mandated location accuracy for emergencyservices. To illustrate the potential of LBS, we will now provide apartial list of applications that will be enhanced using wirelesslocation information [3].

■ E911: Currently, a high percentage of E911 calls originatefrom mobile phones; the percentage is estimated at onethird of all 911 calls (170,000 a day) [5], [6]. These wirelessE911 calls do not receive the same quality of emergencyassistance that fixed-network 911 calls enjoy. This is due tothe unknown location of the wireless E911 caller. To facethis problem, the FCC issued an order on 12 July 1996 [5],requiring all wireless service providers to report accuratemobile station (MS) location information to the E911 opera-tor at the PSAP. In the FCC order, it was mandated that

within five years from the effective date of the order, 1October 1996 (a deadline that is now well passed), wirelessservice providers must convey to the PSAP the location ofthe MS within 100 m of its actual location for at least 67%of all wireless E911 calls. (The original FCC requirementwas 125 m and was later tightened to 100 m.) This FCCmandate has motivated considerable research effortstowards developing accurate wireless location algorithmsfor cellular networks and has led to significant enhance-ments to the wireless location technology (see, e.g.,[12]–[25]). According to the latest FCC rules, the new man-date and accuracy requirements will be enforced in 2005.Although the FCC does not have a specific order for indoorenvironments, a location capability coverage for both indoorand outdoor emergency situations is desirable.■ Mobile advertising: Location-specific advertising and mar-keting will benefit once the location information becomesavailable. For example, stores will be able to track customerlocations and attract them by flashing customized couponson customers’ wireless devices [14]. In addition, a cellularphone or a personal digital assistant (PDA) could act as ahandy mobile yellow pages on demand.

IEEE SIGNAL PROCESSING MAGAZINE [25] JULY 2005

ACRONYMS DESCRIPTION2G SECOND GENERATION OF MOBILE SYSTEMS3G THIRD GENERATION OF MOBILE SYSTEMSAOA ANGLE OF ARRIVALAP ACCESS POINTAMPOA AMPLITUDE OF ARRIVALBS BASE STATIONCDMA CODE DIVISION MULTIPLE ACCESSE911 ENHANCED 911FCC FEDERAL COMMUNICATIONS COMMISSIONGPS GLOBAL POSITIONING SYSTEMLBS LOCATION BASED SERVICESML MAXIMUM LIKELIHOODMS MOBILE STATIONNLOS NON-LINE-OF-SIGHTPDA PERSONAL DIGITAL ASSISTANTPSAP PUBLIC SAFETY ANSWERING POINTrms ROOT MEAN SQUARESINR SIGNAL-TO-INTERFERENCE-NOISE RATIOSNR SIGNAL-TO-NOISE RATIOTDOA TIME DIFFERENCE OF ARRIVALTOA TIME OF ARRIVALUMTS UNIVERSAL MOBILE TELECOMMUNICATIONS SYSTEMWCDMA WIDEBAND CODE DIVISION MULTIPLE ACCESSWLAN WIRELESS LOCAL AREA NETWORK

[TABLE 1] LIST OF ACRONYMS.

[FIG1] Forecast revenues for location-based services [1], [2].

US Others

Location-Based Services EstimatedAnnual Revenues for 2006/2007 (US $B)

US$3.3BUS$11.7B

■ Asset tracking (indoor/outdoor): Wireless location technol-ogy can also assist in advanced public safety applications,such as locating and retrieving lost children, patients, orpets. In addition, wireless location technology can be used totrack personnel/assets in a hospital or a manufacturing siteto provide more efficient management of assets and person-nel. One could also consider applications such as smart andinteractive tour guides, smart shopping guides that directshoppers based on their location in a store, and traffic con-trols in parking structures that guide cars to free parkingslots. Department stores, enterprises, hospitals, manufactur-ing sites, malls, museums, and campuses are some of thepotential end users to benefit from the technology.■ Fleet management: Many fleet operators, such as policeforces, emergency vehicles, and other services like shuttle andtaxi companies, can make use of the wireless location technolo-gy to track and operate their vehicles in an efficient manner tominimize response times. In addition, a large number of driv-ers on roads and highways carry cellular phones while driving.The wireless location technology can help track these phones,thus transforming them into sources of real-time traffic infor-mation that can be used to enhance transportation safety.■ Location-based wireless access security: New location-based wireless security schemes can be developed to height-en wireless network security and avoid the interception ofdigital information. By using location information, only peo-ple at specific physical areas could access certain files ordatabases through a WLAN.■ Location sensitive billing: Using the location informationof wireless users, wireless service providers can offer variable-rate call plans or services that are based on the caller location.

MOBILE-BASED VERSUS NETWORK-BASED TECHNIQUESWireless location technologies fall into two main categories:mobile based and network based. In mobile-based location sys-tems, the MS determines its location from signals received from

some base stations (BSs) or from the global positioning system(GPS). In GPS-based estimations, the MS receives and measuresthe signal parameters from at least four satellites of the currentnetwork of 24 GPS satellites. The parameter measured by the MSfor each satellite is the time the satellite signal takes to reach theMS. GPS systems have a relatively high degree of accuracy, andthey also provide global location information. There is also ahybrid technique that uses both the GPS technology and the cel-lular infrastructure. In this case, the cellular network is used toaid the GPS receiver embedded in the mobile handset forimproved accuracy and/or acquisition time [15].

Still, embedding a GPS receiver into mobile devices leads toincreased cost, size, and battery consumption. It also requiresthe replacement of millions of mobile handsets that are alreadyon the market. In addition, the accuracy of GPS measurementsdegrades in urban environments as well as inside buildings. Forthese reasons, some wireless service providers may be unwillingto embrace GPS fully as the sole location technology.

Network-based location technology, on the other hand, relies onsome existing networks (either cellular or WLAN) to determine theposition of a mobile user by measuring its signal parameters whenreceived at the network BSs. In this technology, the BSs measurethe signals transmitted from an MS and relay them to a central sitefor further processing and data fusion to provide an estimate of theMS location. A significant advantage of network-based techniques isthat the MS is not involved in the location-finding process; thus,the technology does not require modifications to existing handsets.However, unlike GPS location systems, many aspects of network-based location are not yet fully studied.

The rest of this article focuses on network-based wirelesslocation. For location estimation, two operations must be per-formed at the BSs. The BSs have to measure some signal param-eters (such as the time or the angle of arrival) of the received MSsignals. Then, the measured signal parameters are combined in adata fusion stage to provide the final estimate for location. Bothof these stages are discussed in the following sections. Figure 2


[FIG2] Network-based wireless location finding. (a) Outdoor environment using a cellular network. (b) Indoor environment using aWLAN.

TOA/AOAEstimator

Data FusionCenter

TOA/AOAEstimator

TOA/AOAEstimator

α3

α1

r3

r1

r2α2

BS3 (x3, y3)

BS2 (x2, y2)

BS1 (x1, y1)

MS

(a) (b)

46-147

46-148

P J

K

K

N

N

BC

DE

F

GI

J

M

MLK

B C D

D

E F

J46-

127H

46-127W

46-132

43-116

44-123

44-128

r2 r4

r3r1

46-127

46-121A

46-121

46-127

A

A

A

T

B

C

B

A A B C D

C

H FG

E

9

Data FusionCenter


illustrates this two-stage procedure (measurement and datafusion) for an outdoor environment using a cellular network andfor an indoor environment using a WLAN. Although the focus ofthe article is on network-based location systems, most of the net-work-based location algorithms presented here can be used atthe MS as well. Therefore, from now on, network-based wirelesslocation will simply be referred to as wireless location.

DATA FUSION METHODSThe data fusion step combines measurements from differentBSs to obtain an estimate of the MS location. Let (xm, ym)

denote the MS location coordinates in a Cartesian coordinatesystem. Let the coordinates of three BSs (BS1, BS2, and BS3)be denoted by (x1, y1), (x2, y2), and (x3, y3), respectively. Forsimplicity of presentation, only the x and y coordinates areconsidered in the derivations and the z coordinate is ignored.This corresponds to a case where the BSs and the mobile userare located on a relatively flat plane. Without loss of generality,the origin of the Cartesian coordinate system is set at BS1, i.e.,(x1, y1) = (0, 0). Several data fusion techniques have beenintroduced in the literature; these techniques depend on whatsignal parameters are measured at the BSs [3], [4]. (These areseveral studies in the literature that compare the performanceof different fusion algorithms, e.g. [26], [27].) The most com-mon signal parameters are the time, angle, and amplitude ofarrival of the MS signal.

TIME OF ARRIVAL DATA FUSIONThe time of arrival (TOA) data fusion method is based on com-bining estimates of the TOA of the MS signal when arriving atthree different BSs. Since the wireless signal travels at thespeed of light (c = 3 × 108 m/s), the distance between the MSand BSi is given by

ri = (ti − t o)c, (1)

where to is the time instant at which the MS begins transmis-sion and ti is the TOA of the MS signal at BSi. The distances(r1, r2, r3) can be used to estimate (xm, ym) by solving the fol-lowing set of equations (see Figure 3):

r21 =x2

m + y2m (2)

r22 = (x2 − xm)2 + (y2 − ym)2 (3)

r23 = (x3 − xm)2 + (y3 − ym)2 . (4)

Without loss of generality, it can be assumed that r1 < r2 < r3.One way to solve this overdetermined nonlinear system of

equations is as follows. First, (2) and (3) are solved for thetwo unknowns (xm, ym) to yield two solutions. As shown inFigure 3, (2) and (3) each define a locus on which the MSmust lie. Second, the distance between each of the two possi-ble solutions and the circle given by (4) is calculated. Thesolution that results in the shortest distance from the circle(4) is chosen to be an estimate of the MS location coordinates

[12]. Although this method helps resolve the ambiguitybetween the two solutions resulting from (2) and (3), it doesnot combine the third measurement r3 in an optimal man-ner. Furthermore, it is not possible in this way to combineTOA measurements from more than three BSs (which wouldbe useful when the measurements {ri} are subject to inaccu-racies and noise).

This issue can be addressed by combining all the availablemeasurements using a least-squares solution as follows (alterna-tive techniques such as maximum likelihood (ML) solution canbe found, e.g, in [49], [50]). Subtracting (2) from (3) gives

r22 − r2

1 = x22 − 2x2 xm + y2

2 − 2y2 ym.

Similarly, subtracting (2) from (4) gives

r23 − r2

1 = x23 − 2x3 xm + y2

3 − 2y3 ym.

Rearranging terms, the above two equations can be writtenin matrix form as

x2 y2

x3 y3

xm

ym

= 1

2

K22 − r2

2 + r21

K23 − r2

3 + r21

, (5)

where

K2i = x2

i + y2i . (6)

Then, (5) can be rewritten as

Hx = b, (7)

where

H =

x2 y2

x3 y3

, x =

xm

ym

, b = 1

2

K22 − r2

2 + r21

K23 − r2

3 + r21

.

If more than three TOA measurements are available, it can beverified that (7) still holds with

[FIG3] TOA data fusion using three BSs.

BS3

MS

BS2

BS1

r3

r1

r2(x3, y3)

(0,0)

(xm, ym)

(x2, y2)

H =

x2 y2

x3 y3

x4 y4

......

, b = 12

K22 − r2

2 + r21

K23 − r2

3 + r21

K24 − r2

4 + r21

...

. (8)

In this case, the least-squares solution of (7) is given by ([3], [7])

x =(

HTH)−1

HTb. (9)

It is seen that the TOA data fusion method requires accuratesynchronization between the BSs and MS clocks so that themeasurements {ri} are adequate approximations for the actualdistances. Many of the current wireless system standards onlymandate tight timing synchronization among BSs (see, e.g.,[30]). The MS clock itself might have a drift that can reach a fewmicroseconds. This drift directly generates an error in the loca-tion estimate of the TOA method. In the next subsection wepresent a data fusion technique that combines time DOA(TDOA) measurements and helps avoid MS clock synchroniza-tion errors [3], [31], [34].

TDOA DATA FUSIONThe TDOA associated with BSi is ti − t1; i.e., it is the differencebetween the TOAs of the MS signal at BSi and BS1. Now wedefine the distance differences

ri1�= ri − r1

= (ti − t o)c − (t1 − t o)c = (ti − t1)c. (10)

Note that these differences are not affected by errors in the MSclock time (t o) as it cancels out when subtracting two TOAmeasurements. (3) can be rewritten in terms of the TDOA meas-urement r21 as

(r21 + r1)2 = K2

2 − 2x2 xm − 2y2 ym + r21.

Expanding and rearranging terms gives

−x2 xm − y2 ym = r21 r1 + 12

(r2

21 − K 22

).

Similarly, (4) leads to

−x3 xm − y3 ym = r31 r1 + 12

(r2

31 − K23

).

Rewriting these equations in matrix form gives

Hx = r1c + d, (11)

where

c =

−r21

−r31

, d = 1

2

K22 − r2

21

K23 − r2

31

.

Equation (11) can be used to solve for x in terms of theunknown r1 to yield

x = r1H−1c + H−1d. (12)

Substituting this intermediate result into (2) leads to a quadraticequation in r1. Solving for r1 and substituting the positive rootback into (12) yields the final solution for x.

If more than three BSs are involved in the MS location, (11)still holds with

H =

x2 y2

x3 y3

x4 y4

......

, c =

−r21

−r31

−r41

...

, d = 12

K22 − r2

21

K23 − r2

31

K24 − r2

41

...

which yields the following least-squares intermediate solution

x =(

HTH)−1

HT(r1c + d). (13)

Combining this intermediate result with (2) again, the final esti-mate for x is obtained. A more accurate solution can be obtainedas in [32] if the second-order statistics of the TDOA measure-ment errors are known.

ANGLE OF ARRIVAL DATA FUSIONAt the BS, angle of arrival (AOA) estimates can be obtainedusing an antenna array. The direction of arrival of the MS sig-nal can be calculated by measuring the phase differencebetween the antenna array elements or by measuring thepower spectral density across the antenna array in what isknown as beamforming (see, e.g., [37] and the referencestherein). By combining the AOA estimates of two BSs, an esti-mate of the MS position can be obtained (see Figure 4). Thenumber of BSs needed for the location process is less than that


[FIG4] Combining AOA measurements.

BS BS

MS

(x2, y2)(x1, y1)

r1 cos α1

r 1 s

in α

1

r 2 s

in α

2

r1 cos α2

r1

r2

(xm,ym)

α1 α2


of the TOA and TDOA meth-ods. Another advantage ofAOA location methods isthat they do not require BSor MS clock synchroniza-tion. However, one disad-vantage of the AOA methodis that antenna array structures do not currently exist in sec-ond generation (2G) cellular systems. Still, the use of antennaarrays is planned in third generation (3G) cellular systems,such as UMTS and cdma2000 networks [38], [39].

More generally, assume n BSs estimate the AOA of the MSsignal, and the goal is to combine these measurements to esti-mate the MS location. As indicated in Figure 4, let α2 denote theAOA of the MS signal at BS2. Then

[xm

ym

]=

[r1 cos α1

r1 sin α1

]

and[

xm

ym

]=

[x2

y2

]+

[r2 cos α2

r2 sin α2

].

Likewise, for any other BSi,

[xm

ym

]=

[xi

yi

]+

[ri cos αi

ri sin αi

].

Collecting these relations into a single equation yields

Hx = b,

where

H =

1 00 1

1 00 1

......

1 00 1

, x =[

xm

ym

], b =

r1 cos α1

r1 sin α1

x2 + r2 cos α2

y2 + r2 sin α2

...

xn + rn cos αn

yn + rn sin αn

. (14)

The least-squares solution for x is then

x =(

HTH)−1

HTb. (15)

Besides the regular sources of error in AOA measurements,such as noise and interference, AOA measurements can becorrupted by non-line-of-sight (NLOS) effects and errors in

the angular orientation ofthe installed antennaarrays. The issue of NLOSis discussed in another sec-tion. For the error in theangular orientation of theantenna arrays, some test

measurements can be conducted to calibrate the orientationof the antenna array.

HYBRID DATA FUSION TECHNIQUESIn TOA, TDOA, and AOA methods, two or more BSs are involvedin the MS location process. In situations where the MS is muchcloser to one BS (serving site) than the other BSs, the accuracyof these methods can be degraded due to the relatively low SNRof the received MS signal at one or more BSs. The accuracy isfurther reduced if some type of power control is used, since thisrequires that the MS reduce its transmitted power when itapproaches a BS. In these cases, an alternate data fusion proce-dure is used to obtain AOA estimates and combine them withTOA estimates (see, e.g., [40]). In real scenarios, the accuracy ofTOA and AOA estimates is usually a function of the environ-ment. For example, in rural areas, AOA measurements can bemore accurate than TOA measurements if a large-size antennaarray is deployed. On the other hand, TOA measurements aremore accurate than AOA measurements if the BS antenna arrayis surrounded by many scatterers. The following is a simple two-step hybrid least-squares procedure. Assume n BSs estimate theAOA and TOA of the MS. From (9), the least-squares estimate of(xm, ym) using TOA measurements is given by

(xm

ym

)TOA

=(

HTTOAHTOA

)−1HT

TOAbTOA, (16)

where

HTOA =

x2 y2

x3 y3

......

xn yn

, bTOA = 12

K22 − r2

2 + r21

K23 − r2

3 + r21

...

K2n − r2

n + r21

and

K2i = x2

i + y2i .

Likewise, from (15), the least-squares estimate of (xm, ym)

using only AOA measurements is given by

(xm

ym

)AOA

=(

HTAOAHAOA

)−1HT

AOAbAOA (17)

where

A LARGE PROPORTION OF ALL 911 CALLS ORIGINATE FROM MOBILE PHONES,

THE LOCATION OF WHICH SHOULD BEDETERMINED WITH SUFFICIENT ACCURACY.

IEEE SIGNAL PROCESSING MAGAZINE [30] JULY 2005IEEE SIGNAL PROCESSING MAGAZINE [30] JULY 2005

HAOA =

1 00 1

1 00 1

......

1 00 1

, bAOA =

r1 cos α1

r1 sin α1

x2 + r2 cos α2

y2 + r2 sin α2

...

xn + rn cos αn

yn + rn sin αn

. (18)

The final location estimate could be taken as a linear combina-tion of the two estimates, say as

(xm

ym

)= η

(xm

ym

)TOA

+ (1 − η)

(xm

ym

)AOA

(19)

where the positive parameter η is chosen depending on the rela-tive accuracy of the TOA and AOA measurements.

DATA FUSION WITH NLOS CONDITIONSAn important source of error in TOA-based and AOA-baseddata fusion is the case where there is no line-of-sight from themobile station to the BSs. A geometrically constrained datafusion scheme could be used to reduce the effect of such NLOSconditions [41] (see [8] for other ways to exploit the geometryof the problem). Figure 5 shows a representation of a cellularsystem assuming three BSs. Let the θi denote the anglesinduced by the topology of the BSs. Let also rij denote the dis-tance between the i th and j th BSs. Likewise, the αi denotethe AOAs of the MS signal at the BSs. In practice, the distancemeasurements ri in (1) are generally corrupted by NLOS off-sets arising from the presence of obstacles between the MS andthe BS, as well as by measurement noise. Similarly, the AOA

measurements αi are corrupted by NLOS effects and by noise.Hence, the available measurements are

αi =αi + vαi

ri =ri + vri (20)

where vαi and vri represent the corruptions to αi and ri. Onescheme for recovering (xm, ym) from the measurements{ ri, αi} is based on formulating a constrained optimizationproblem that reduces the effect of NLOS conditions on loca-tion accuracy. The constraints will be a reflection of the topol-ogy of the cellular network. Thus, consider the cellular systemshown in Figure 5. The constraints are the distances betweenthe BSs, which are given by

r212 =r2

1 + r22 − 2r1 r2 cos γ1 (21)

r223 =r2

2 + r23 − 2r2 r3 cos γ2 (22)...

where the angles {γi} are functions of {αi, θi}. This formulationis easily extendable to the case of n BSs. Then, one could posethe problem of estimating the {αi, ri} by solving

{αi, ri}ni=1 = arg min

{αi,ri}

n∑i=1

(αi − αi

σαi

)2

+(

ri − ri

σri

)2

(23)

subject to

r212 = r2

1 + r22 − 2r1 r2 cos(π − (α1 + α2))

r223 = r2

2 + r23 − 2r2 r3 cos(π − (α3 + (θ2 − α2)))

...

where σ 2ri

is the variance of the distance error and σ 2αi

is thevariance of the angle error (both at the i th BS). There are someknown methods for estimating the variances σ 2

riand σ 2

αi(see,

e.g., [42]–[44]). These methods generally use the time history ofthe signals, or the scattering model of the environment, to esti-mate the noise variance, as in

σ 2ri

≈ 1K

K−1∑n=0

( ri(n) − µri)2, (24)

where

µri = 1K

K−1∑n=0

ri(n) (25)

for K ≈ 400 and where ri(n)is the measurement of ri at experi-ment n. This is also true for σ 2

αi. Minimizing (23) results in esti-

mates of {ri, αi}. Using the equalized values in (8) or (9) willresult in improved location accuracy.[FIG5] A schematic of a cellular network topology with three BSs.

θ2

θ1

γ2

γ1

γ3

BS1

(x3, y3)

(x2, y2)

(0,0)

Base Station

Mobile Station

BS2

r1

r2

r3

r12

r23

r31

BS3

α2

α1

α3


THE WIRELESS ENVIRONMENTFrom the previous discussion, it is clear that most wireless loca-tion methods depend on combining estimates of the TOA and/orAOA of the received signal at different BSs. Estimating the TOAand amplitude of arrival (AmpOA) of wireless signals has beenstudied in many works since itis required in many wirelesssystem designs for online signaldecoding purposes. Yet, esti-mating these same parametersfor wireless location purposes ischallenging for several reasons.

■ Low SINR conditions. Cellular systems tend to sufferfrom high multiple access interference levels thatdegrade the SNR of the received signal. Moreover, theability to detect the MS signal at multiple BSs is limitedby the use of power control algorithms, which requirethe MS to decrease its transmitted power when itapproaches the serving BS. This fact, in turn, decreasesthe received MS signal power level at other BSs. In a typi-cal CDMA IS-95 cellular environment, the received SNRat the serving BS is in the order of −15 dB. However, thereceived SNR at BSs other than the serving BS can be aslow as −40 dB, which poses a challenge for wireless loca-tion in such environments.■ Channel fading and Doppler frequency. In wireless locationapplications, the estimation period can be considerably long(it might reach several seconds). Thus, in a fading environ-ment, the channel values can change significantly over thelocation estimation period. In this way, the channel values canno longer be assumed constant during the estimation period. ■ Overlapping multipath. In wireless location systems, theaccurate estimation of the TOA, AOA, and AmpOA of thefirst arriving ray of the multipath channel is vital. In gen-eral, the first arriving (prompt) ray is assumed to corre-spond to the most direct path between the MS and BS.However, in many wireless propagation scenarios, theprompt ray is succeeded by a multipath component thatarrives at the receiver within a short time of the promptray. If this delay is smaller than the duration of the pulse-shape used in the wireless system, these two rays overlap,causing errors in the prompt ray TOA and AmpOA estima-tion. These errors degrade the performance of wirelesslocation algorithms; as such, they demand careful consider-ations (see, e.g., [9]–[11], [28], and [29]).In the sections that follow, some algorithms for TOA and

AOA estimation are described. These algorithms exploit thenature of the wireless channel and are robust to low SNR andfading conditions [25], [51].

TOA ESTIMATIONThe aim of a TOA estimation scheme is to estimate an unknowndelay, τ o, of a known sequence {s(n)}. (At the serving site, theMS signal can be decoded with reasonably high accuracy; thus,it can be assumed to be known almost perfectly.) The signal is

assumed initially to be transmitted over a single path fadingchannel. A total of K measurements r(n) are collected, whichare related to s(n) via

r(n) = A h(n) s(n − τ o) + v(n), n = {1, . . . , K }, (26)

where v(n) is additive whiteGaussian noise with varianceσ 2

v , {h(n)} is the fading channelgain, and A is an unknownamplitude (real value) thataccounts for the gain of the

static channel if fading were not present. The autocorrelationfunction of h(n) is defined as

Rh(i ) = Eh(n)h∗(n − i ). (27)

Without loss of generality, we will assume that the sequenceh(n) has unit variance, i.e., Rh(0) = 1. The ML estimates of{τ o, h(n)} are defined by

{τ , h(n)} = arg maxτ,h(n)

[P(r(1) · · · r(K)|τ, h(n))], (28)

where the likelihood function P(r(1) · · · r(K)|τ, h(n)) is of theform

C1 exp

{−C2

1K

K∑n=1

‖r(n) − Ah(n)s(n − τ)‖2

}(29)

for some positive constants C1 and C2 that are independent ofthe unknowns {τ, h(n)}. Thus, the ML estimates of {τ, h(n)} aregiven by [25]

{τ , h(n)} = arg maxτ,h(n)

JML(τ, h(n)), (30)

where the cost function JML is given by

JML(τ, h(n)) = 2AK

K∑n=1

Re[r (n)h∗(n)s∗(n − τ)]

− A2

K

K∑n=1

|h(n)|2|s(n − τ)|2.

This construction requires an infinite dimensional searchover {τ, h(n)} and is not feasible in practice even when τ andh(n) are evaluated over a dense grid. To arrive at a feasiblealgorithm, we assume the channel variations are sufficientlyslow, namely, that h(n) is piecewise constant over intervals ofN samples. The value of N depends on the environmental con-ditions; an optimal choice for N is discussed later in this


WIRELESS LOCATION TECHNOLOGIES FALL INTO TWO MAIN CATEGORIES:

MOBILE BASED AND NETWORK BASED.


article. Under the slow channel variation assumption, it canbe argued that the maximization problem (30) over τ andh(n) can be reduced to maximizing the following cost func-tion instead over τ [25]:

J(τ) = 1M

M∑m=1

∣∣∣∣∣1N

mN∑n=(m−1)N+1

r(n)s∗(n − τ)

∣∣∣∣∣2

, (31)

where the integers N and M satisfy NM = K. A practical schemefor maximizing (31) over τ is shown in Figure 6 and was derivedin [25]. The received sequence {r(n)} is correlated with adelayed replica of the transmitted sequence, {s(n − τ)}, for dif-ferent values of τ . The resulting sequence is averaged coherently

over an interval of N samples and further averaged noncoher-ently for M samples after squaring to build the power delay pro-file, J(τ). The branch that results in the largest value for τprovides the desired estimate τ . Moreover, the SNR at the out-put of the searcher is [25]

SNR = A2

σ 2v

(Rh(0) +

N−1∑i=1

2(N − i)Rh(i)N

).

The optimal value of the coherent averaging period (Nopt) isobtained by maximizing the SNR with respect to N, which leadsto the following expression for finding Nopt:

Nopt−1∑i=1

iRh(i) = 0. (32)

For a Rayleigh fading channel, Rh(i) is given by

Rh(|i|) = Jo(2π fDTs i

),

where Jo(·) is the first-order Bessel function, Ts is the samplingperiod of the received sequence {r(n)}, and fD is the maximumDoppler frequency. Therefore, (32) shows that the coherentaveraging interval N should be adapted according to the channelautocorrelation function.

TOA ESTIMATION WITH ANTENNA ARRAYFurther improvement in the TOA estimation can be accomplishedby deploying an antenna array at the BS [51]. Thus assume thatthe BS uses an Na-element antenna array. Then, in contrast to(26), the received signal at time n is now an Na × 1 vector:

r(n) = ah(n)s(n − τ o) + v(n), (33)

(Note that we are assuming a scat-tered MS model. Since in a typical cel-lular system the mobile station isusually far from the BS, the reflectedrays from the scatterers around theMS reach the BS at close angles.These reflected rays cause a fadingeffect in h(n) with almost the sameAOA. Moreover, it is assumed that anybias in the direction or the angle ofthe arrays can be ignored from thederivations through some calibrationprocedures.) where now v(n) is anadditive white Gaussian noise vectorwith covariance matrix σ 2

v I, and a isthe array response defined by

a = col{

1, e j2π dλ

sin α, . . . , e j2π(Na−1)d

λsin α

}, (34)

where α is the AOA of the signal measured with respect to thearray, d is the antenna spacing, and λ is the wavelength corre-sponding to the carrier frequency. Note that the array responsedefined by (34) is valid only for a uniform linear array (ULA).This array response could be modified if other array structures,such as circular array, are used.

The ML estimates of τ and α are given by

{τ , α} = arg maxτ,α

[P(r(1) · · · r(K))|τ, α], (35)

where the likelihood function is now proportional to

exp

{−C2

1K

K∑n=1

‖r(n) − ah(n)s(n − τ)‖2

}, (36)

in which ‖.‖ is the Euclidean norm of the vector and C2 is somepositive constant. Therefore, the ML estimates of {τ, α} can befound by solving

[FIG6] A scheme for TOA estimation over fading channels.

DopplerEstimator

1N 1

N

N

N

N

Σ

1N

r (n)

1

NΣ

1N 1

NΣ

1M

1

MΣ

1M 1

MΣ

1M 1

MΣ|.|

2

|.|2

|.|2

max ˆ

s*(n–τ2)

s*(n–τ1)

s*(n–τF)

×

×

×


{τ , α} = arg maxτ,α

JML(τ, α), (37)

where

JML(τ, α) = 1K

K∑n=1

‖r(n) − ah(n)s(n − τ)‖2.

This cost function can be simplified by noting that ‖a‖2 = Na

and that the entries of a have unit norm, so that

JML(τ, α) = 1K

K∑n=1

|a∗r(n) − h(n)s(n − τ)|2. (38)

Rather than perform a two-dimensional (2-D) search, the esti-mator can perform the search for τ and α separately as follows.Assume first that α is known.Then the term a∗r(n) in (38) canbe interpreted as the output of abeamformer (antenna combiner)steered to direction α, as shown inFigure 7. The optimization prob-lem for τ , given α, then reducesto the single-antenna case ofFigure 6 using a beamformer atdirection α.

AOA ESTIMATION WITHANTENNA ARRAYNow assume τ in (38) is knownand evaluate the correlation

zm = 1N

(mN∑

n=(m−1)N+1

r(n)s∗(n − τ o)

), m = 1, . . . ,M

(39)

over intervals of length N, during which h(n) is essentiallyinvariant. Then, using (33),

zm = ah(mN)1N

(mN∑

n=(m−1)N+1

|s(n − τ o)|2)

+ 1N

(mN∑

n=(m−1)N+1

v(n)s∗(n − τ o)

). (40)

In other words,

zm = p(m)h(mN)a + um, (41)

where p(m) denotes the constant known power term

p(m) = 1N

(mN∑

n=(m−1)N+1

|s(n − τ o)|2)

,

and um refers to the noise term in (40). Collecting (41) into vec-tor form yields

z1

z2

...

zM

︸︷︷︸z

=

p(1)h(N)INa×Na

p(2)h(2N)INa×Na

...

p(M)h(MN)INa×Na

︸︷︷︸A

a +

u1

u2

...

uM

. (42)

The channel gains {h(N), h(2N), . . . , h(MN)} can be esti-mated roughly from (41) by noting that the top entry of a isunity, so that

h(mN) = zm(1)/p(m).

The LS estimate can then be obtained as

a = (A∗A)−1A∗z. (43)

The AOA information can be extracted from the estimatedarray response a based only on the phase rotation between theentries of a (see Figure 8).

MULTIPATH, MULTIUSER ENVIRONMENTAs mentioned previously, wireless propagation suffers frommultipath conditions, in which case the prompt ray may besucceeded by multipath components that arrive at the receiverwith short delays (e.g., [52]). In this section, we describe oneway to perform TOA and AOA estimation under multipath


[FIG7] TOA estimation over fading channels using an antenna array.

210

240270

120

150

180

90

s*(n–τ2)

s*(n–τ1)

e– j2πfct

e– j2πfct

e– j2πfct e– j2π(Na–1)s*(n–τF)d

λ

1N 1

NΣ |.|

2

|.|2

|.|2

DopplerEstimator

maxˆ

sin α

e– j2π dλ

sin α+

××

×

× ×

×

×

×

×

1N 1

NΣ

1N 1

NΣ 1

M 1

MΣ

1M 1

MΣ

1M 1

MΣ

1N

N

N


conditions. First, we modify the channel model (26) to accom-modate a more general multiuser, multipath environment.Assuming the maximum number of channel taps to be L andthe number of mobile users to be Nu, the received signal r(n)

of size Na × 1 is now given by

r(n) =Nu∑

k=1

L∑l=1

ak,lhk,l(n)sk(n − τ ok,l) + v(n), (44)

where sk(n) is the transmitted sequence by the kth user andv(n) is an Na × 1 additive white Gaussian noise vector.Moreover, hk,l(n) and τk,l are the channel gain and delay,respectively, for user i, and ak,l is the array response correspon-ding to the l th channel tap from user k to the BS, namely

ak,l = col{

1, e j2π dλ

sin αk,l , . . . , e j2π(Na−1)d

λsin αk,l

}, (45)

where αk,l is the AOA for the l th tap and kth user.As in the single path case discussed earlier [see (39)], we

define the correlation vectors:

zk,l,m = 1N

mN∑n=(m−1)N+1

r(n)s∗k(

n − τ ok,l

),

where N is the coherent correlation length. Then [comparewith (41)]

zk,l,m = pk,l(m)hk,l(mN)ak,l + ik,l,m + uk,l,m, (46)

where

pk,l(m) = 1N

(mN∑

n=(m−1)N+1

|sk

(n − τ o

k,l

)|2

)

and

uk,l,m = 1N

(mN∑

n=(m−1)N+1

v(n)s∗k(

n − τ ok,l

))(47)

ik,l,m =Nu∑

k′=1

L∑l′=1

k′ �=k,l′ �=l

ρk,l,k′,l′(m)hk′,l′(mN)ak′,l′ (48)

where

ρk,l,k ′,l ′(m) = 1N

mN∑n=(m−1)N+1

sk′(

n − τ ok ′,l ′

)s∗k

(n − τ o

k,l

)

(49)

represents the correlation between the sequences of user k andall other users. Collecting M such realizations into a vector zk,l

yields

zk,l,1zk,l,2

...

zk,l,M

︸︷︷︸zk,l

=

pk,l(1)hk,l(N)INa×Na

pk,l(2)hk,l(2N)INa×Na

...

pk,l(M)hk,l(MN)INa×Na

︸︷︷︸Ak,l

ak,l

+

ik,l,1ik,l,2

...

ik,l,M

︸︷︷︸ik,l

+

uk,l,1uk,l,2

...

uk,l,M

. (50)

The least-squares estimation of ak,l can be obtained as

ak,l =(

A∗k,lAk,l

)−1A∗

k,lzk,l. (51)

The AOA information is finally extracted from the estimatedarray response ak,l based only on the phase rotation between theentries of ak,l according to (45). The above least-squares estima-tion is repeated for all users and multipaths, k = 1, . . . , Nu,l = 1, . . . , L.

As the number of users and multipaths increases, multipleaccess interference (MAI) and intersymbol interference (ISI)in (46) become stronger. For practical scenarios with a largenumber of active users in a cell, the accuracy of the least-squares estimation of AOA is limited by MAI and ISI. Onesolution for reducing the effect of MAI in (46) is to increasethe coherent correlation length N, as well as the realizationlength M. However, the correlation or estimation length can-not be increased indefinitely. The correlation length shouldbe short enough such that the channel taps can be assumedconstant during the estimation process.

We may use joint least-squares estimation followed by multi-user interference cancellation to provide an accurate AOA

[FIG8] AOA estimation using an antenna array over a single pathfading channel.

Arr

ay R

espo

nse

Leas

t-S

quar

es E

stim

atio

n

Col

lect

M R

ealiz

atio

ns (

z m)r(n)

â αN∑1

N∑1

N∑1

s*(n– )

×

×

×

s*(n– )

s*(n– )


estimation in the presence of interfering users. This joint tech-nique takes advantage of the following two facts:

■ A BS (in normal operatingmode) detects and decodesthe signals from all userssimultaneously. Therefore,the BS knows the trainingsequences used by the users,i.e., sk(n), k = 1, . . . , Nu.■ The BS performs the channel and array response estima-tion for all users and multipaths. Therefore, the estimation ofthe channel taps (ak,l and hk,l, k = 1, . . . Nu, l = 1, . . . , L)are available at the BS.The above information is used in a secondary stage to fur-

ther enhance the accuracy of the estimation process by cancel-ing the MAI in (46). The interfering signal is regenerated usingthe known sk(n), the estimated channel gain hk,l, and the arrayresponse ak,l. It is then subtracted from the previous correla-tion results zk,l, and a new estimate is obtained. The followingsteps are performed in the proposed architecture:

1) Use (46)–(51) to calculate {αk,l, hk,l(mN)}.2) Use the estimated channel taps to regenerate (estimate) theMAI term ik,l,m in (48), i.e.,

ik,l,m =Nu∑

k ′=1

L∑l ′=1

k ′ �=k,l ′ �=l

ρk,l,k ′,l ′(m)hk ′,l ′(mN)ak ′,l ′ . (52)

3) Subtract the estimated interference (52) from the correla-tion vector in (46). The new zk,l,m becomes

zk,l,m = zk,l,m − ik,l,m

with the interference term ik,l,m reduced.4) Use the zk,l,m in (51) to find an improved {αk,l, hk,l(mN)}.5) Repeat steps 2–4 for all users and multipaths as necessary.

The above procedure can be repeated until an AOA estimatewithin the desirable range is achieved.

WLANSSimilar technical challenges also arise for wireless locationdetermination in indoor environments, where WLAN is cur-rently the most widely deployed wireless network. WLANstandards such as IEEE802.11b, and more recentlyIEEE802.11g, have been widely adopted in offices, homes,hospitals, restaurants, and schools. WLAN connectivity hasalso become a standard feature for laptop computers andPDAs as well as the new generation of smart cellular phones.As such, there is an increasing interest in utilizing these net-works for location purposes to help provide good coverage forindoor scenarios.

THE INDOOR ENVIRONMENTYet, the indoor channel environment is challenging for a num-

ber of reasons.■ Channel fading. The channelvariation as a function of posi-tion due to the scatterer-richnature of indoor environmentscan be significant. In a scatterer-rich environment, the channel

can be considered correlated only over a distance of λ/2,where λ is the wavelength corresponding to the carrier fre-quency. At 2.4 GHz, which is the band of operation forIEEE802.11b and IEEE802.11g, λ/2 is less than 10 cm. Inother words, a movement of about 10 cm in an indoor envi-ronment can result in significant change in the channel gain.■ Path loss and shadow fading. The distance between theaccess point (AP) and mobile users causes path loss in thesignal strength. The path loss in a typical office area is pro-portional to d−3.5 , where d is the distance between themobile user and the AP. In addition to the path loss, theshadow fading caused by walls further contributes to atten-uation of the signal strength.■ Interference. The 2.4-GHz band is an unlicensed bandwhere devices such as Bluetooth, cordless phones, and evenmicrowave ovens operate. The interference from these otheractive devices can limit the achievable location accuracy.

AMPOA ESTIMATION Some of the older proposed location-aware systems for indoorenvironments [53]–[57] require specialized hardware such asultrasound transmitters, camera, and infrared transmitters.But since the IEEE 802.11b and IEEE 802.11g MAC layer soft-ware provides the signal strength and the SNR, a software-levellocation technique could be developed for WLAN networksbased on the AmpOA at different access points [58]–[66].Specifically, when an IEEE802.11 wireless network operates inthe infrastructure mode, there are several APs and many endusers within the network. RF-based systems that use the signalstrength for location purposes can monitor the received signalstrength from different APs and use the obtained statistics tobuild a conditional probability distribution network to esti-mate the location of the mobile client. These schemes usuallywork in two phases: the first phase is the offline training anddata gathering phase, and the second phase is the locationdetermination phase using the online signal strength meas-urements. In the training phase, signal strength measure-ments are used to build an a priori probability distribution ofthe received signal strength at the mobile user from all APs.Assume there are N APs in the system and the radio map iscreated based on measurements from M user locations. Theradio map model is described by [55]–[58]

p(Ai|xj, yj)�= the PDF of the received signal strength,

where


LOCATION-BASED APPLICATIONS WILL PLAY AN IMPORTANT ROLE IN FUTURE

WIRELESS MARKETS.


Ai = received signal strength from the i th AP

(xj, yj) = coordinates of the jth measurement point

i = 1, 2, . . . , N, j = 1, 2, . . . , M.

After constructing a Bayesian network, the online determi-nation phase uses ML estimation to locate the mobile user.Thus, assume that the mobile user measures the received signalstrength from all APs, as in

A ′i

�= measured signal from the ith AP.

Then, using Bayes’ rule, the probability of having the mobileuser at location (xj, yj) given the received signal strengths fromall APs is given by

A ′ �= [A ′1, . . . , A ′

N]

p(xj, yj|A ′) = p(A ′|xj, yj)p(xj, yj)

p(A ′)

= p(xj, yj)∏N

i=1 p(A ′i|xj, yj)

p(A ′),

where ∏N

i=1 p(A′i|xj, yj) is the approximation for the conditional

probability density function of the received signal strength whenthe location of the mobile user is given. Thus, the location ofthe mobile user can be estimated as

(xm, ym) = arg maxxj,yj

p(xj, yj|A ′)

j = 1, 2, . . . , M. (53)

TOA/AOA ESTIMATIONThe TOA and AOA estimation techniques and the datafusion schemes presented in the previous sections can beused for indoor environments as well. However, the accura-cy desired for indoor applications is higher than thatrequired for outdoor environments. While an accuracy of 50m is acceptable for many outdoor applications, for indoorapplications an accuracy of few meters is desired. Therefore,the performance of the estimation algorithms should beboosted to meet the accuracy requirements. The followingfacts will improve the accuracy of location finding algo-rithms for indoor applications:

■ Higher clocking rates. The clocking rates of WLAN sys-tems are higher than the ones used in cellular systems; thisis due to the fact that the WLAN physical layers are intend-ed for higher data rates and occupy a wider bandwidth thanthe physical layer of cellular systems. The higher clockingrate (and, equivalently, the higher sampling rate at thereceiver) translates into higher accuracy in TOA measure-ments and into more accurate location estimates.Additionally, the bandwidth per channel used in 3G cellularnetworks is about 4 MHz, as opposed to the 11-MHz band-width in IEEE802.11b and 16-MHz bandwidth inIEEE802.11a and IEEE802.11g. The higher bandwidth andclocking rate effectively provide a higher resolution in esti-mating the TOA of the signals.■ Higher SNR. WLAN networks operate at higher SNR thancellular networks. The higher SNR results in more accurateestimates for TOA and AOA.■ Oversampling at the receiver. The received signal can beoversampled to further increase the resolution of TOA estimation. Since the received SNR in WLAN networks isrelatively high, an accurate TOA estimation after oversam-pling is possible.

[FIG9] An illustration of the UCLA WLAN Location Simulatorinterface. The estimated location of a mobile user using differentalgorithms would be plotted for different realizations. Moreover,the estimated accuracy of the different methods would beshown by circles surrounding the mobile location.

[FIG10] An illustration of the UCLA Cellular Location Simulatorinterface. The estimated location of a mobile user using differentalgorithms would be plotted for different realizations. Moreover,the estimated accuracy of the different methods and the FCCaccuracy requirements would be shown by circles surroundingthe mobile locations. The design engine allows the user to placeblocking objects in the simulator environment and to select thetrajectory of the mobile user such that it experiences differentsituations, fadings, and shadowings. The various parametersthat control the environment can be adjusted as well.


■ Slow-varying channel. The channel variation in indoorenvironments is slower than in outdoor environments. This isdue to the range of speeds present in indoor and outdoorenvironments. The slow-varying channel allows for longercoherent averaging periods at the receiver (the parameter Nused in the derivations), which results in a higher effectiveSNR for TOA and AOA estimation.■ Power up scheme. Due to the local nature of WLAN net-works, the mobile user can be requested by the network toraise the level of transmitted power momentarily. Thisinstant increase in the level of transmitted power will notdegrade the network performance as significantly as itwould for cellular networks.For further information on indoor location algorithms, see

[15] and [61]. Moreover, [62] describes localization algorithmsin a cooperative sensor network setting.

SIMULATION ENVIRONMENTTo test several of the techniques and algorithms described in theprevious section, a software simulator called the WirelessLocation Simulator has been developed at the UCLA AdaptiveSystems Laboratory [67]. A high degree of testability and flexi-bility, along with a user-friendly interface, are designed into thesimulator. Selected snapshots of the Wireless LocationSimulator are shown in Figures 9–12. The simulator consists ofan interface and a location-finding engine. The location engineperforms the following tasks:

■ Data fusion techniques: Different data fusion techniquesare implemented using TOA, AOA, or a combination of both.■ Channel modeling: A multipath, multiuser channel envi-ronment is created that models path loss, shadowing,Rayleigh fading, and Doppler frequency effects.■ Parameter estimation: TOA and AOA estimation algorithmsare implemented as part of the location finding engine.Different variations of the algorithms are implemented forperformance and comparison purposes.Most of the algorithms used in the location-finding engine

are generic in the sense that they could be used for both indoorand outdoor applications with minimum alteration. The loca-tion-finding engine and the capabilities listed below make thesimulator usable for both networks.

■ Configuring the physical layer for different wireless net-works. Adjusting the network physical layer parametersenables the simulator to be used for different wirelessnetworks. Among the programmable parameters are thespreading factor, packet size, training length, constellationtype, modulation technique, carrier frequency, level oftransmitted signal power, and the number of antennas atthe transmitter or receiver.■ Configuring the mobile user conditions. The wirelesschannel models depend on the Doppler frequencies presentin the environment. The Doppler frequency depends on themobile speed and the carrier frequency of the system. Thesimulator accepts different speed and carrier frequencies asinput and generates a Rayleigh fading channel with the U-

shape Doppler spectrum. Moreover, the simulator allowsthe user to define the trajectory of the mobile, and thentracks the user on the defined trajectory.■ Configuring the environmental parameters and net-work geographical structure. One of the factors affect-ing the performance of the wireless location system isthe environment type (e.g., bad urban, urban, suburban,or rural.) For example, in a bad urban area with manyblocking objects and buildings, NLOS effects play animportant role in the estimation accuracy. The simulatorenables the user to place an arbitrary number of build-ings of different sizes and shapes in the simulationenvironment. The simulator modifies the channel mod-els to capture the effect of the buildings (shadow fadingand NLOS effects).

[FIG11] Snapshot of the UCLA Location Simulator. The locationestimation accuracy and the FCC requirements are shown by twocircles in this lateral view generated by the 3-D simulationvisualizer.

BSBS

BS BS

Mobile User

[FIG12] Snapshot of the UCLA Location Simulator. The locationestimation accuracy and the FCC requirements are shown by twocircles in this top view generated by the 3-D simulation visualizer.

AttainedAccuracy

Mobile UserTrajectory

FCC Accuracy Requirement

■ Performance evaluation and monitoring tools. The simu-lator contains different tools to illustrate the results andaccuracy of the location procedure, such as

1) comparing the estimated trajectory with the true tra-jectory2) statistics of the location error, including the 67% and95% outage values (the 67% and 95% outage values are ofinterest due to the FCCrequirements)3) a three-dimensional(3-D) simulator thatcreates a 3-D virtualreality environment.The 3-D viewer hasdifferent adjustablecamera views from different angles. It provides the userwith an interactive 3-D environment that shows themovement of the mobile user on the predefined trajecto-ry. It also shows the accuracy of the algorithm and com-pares it with the FCC standard.

SOME SIMULATION RESULTSSome simulation results are shown in Figure 13(a) and (b) fora CDMA cellular network with the following parameters:CDMA chip rate of 4 MHz, a processing gain of 64, a 3-tapRayleigh fading channel with path loss exponent of two, anantenna array of size four at the BS, and a Doppler frequencycorresponding to a maximum speed of 30 mph. The location

algorithms are simulated ina multi-user environmentwith a different number ofactive users. Figure 13(a)and (b) shows the resultingaccuracy outage curves. Anoutage curve measures theprobabilities that the loca-

tion estimation errors will be below certain values.For example, in Figure 13(a), it is seen that for a receiver

employing both TOA and AOA measurements, the locationerror is below 100 m 90% of the time. The FCC-mandatedrequirement for a network-based solution is for the location


[FIG13] Outage curves for location accuracy in outdoor environment (IC denotes interference cancellation method). (a) Assuming oneuser. (b) Assuming six users.

• Number of Base Stations=3• Number of Active Users=1• Number of Antennas=4• Number of Multipaths=3• Maximum Speed=30 mph

(a) (b)

Multiple Antenna (TOA)–Section IV-AMultiple Antenna (TOA/AOA)–Section IV-D

Multiple Antenna (TOA)–Section IV-AMultiple Antenna (TOA/AOA)–Section IV-DMultiple Antenna (TOA/AOA) with IC–Section X

• 67% Outage=47 m• 95% Outage=144 m • 67% Outage=45 m

• 95% Outage=128 m

• Number of Base Stations=3• Number of Active Users=6• Number of Antennas=4• Number of Multipaths=3• Maximum Speed=30 mph

0 50 100 150 200 250 300 350 4000

10

20

30

40

50

60

70

80

90

100

Location Error (m)

Out

age

(%)

FCC Requirement For Handset-Based Location

Out

age

(%)

0 50 100 150 200 250 300 350 4000

10

20

30

40

50

60

70

80

90

100

Location Error (m)

FCC Requirement for Network-Based Location

FCC Requirement For Handset-Based Location

FCC Requirement for Network-Based Location

A LOCATION FINDING SYSTEM SHOULD BEABLE TO SEAMLESSLY USE BOTH CELLULAR

AND WLANS FOR LOCATION FINDING BYROAMING BETWEEN THE NETWORKS.


error to be below 100 m only 67% of the time. For a mobile-based solution, on the other hand, the FCC requirementmandates an outage probability of 67% for 50 m. Both ofthese requirements are met by the TOA/AOA fusion methodpresented earlier. Figure 13(b) repeats the simulation for thecase of six total users.

CONCLUSIONNetwork-based wireless location poses several interesting prob-lems from a signal processing perspective. The estimation algo-rithms must provide accurate parameter estimates underchallenging conditions such as fast fading channels, low SNR,multipath effects, and multiuser interference. Small errors inestimation can lead to large errors in location. For example, in3G CDMA, the chip duration is roughly 0.25 µs. An error in TOAestimation of the order of Tc/2 can translate into 37.5 m in loca-tion error. Likewise, for a cell with a two-mile radius, an error of1◦ in the AOA measurement can result in a location error of theorder of 55 m. For this reason, the location estimation algo-rithms and the data fusion methods must exploit any availableinformation about the environment (e.g., fading conditions,Doppler frequency, and network topology) to attain high accura-cy. In addition, the resulting location searchers need to exhibit acertain degree of adaptation to changing conditions (e.g., mobilespeed) so as to maintain reliable performance.

ACKNOWLEDGMENTSThis work was supported in part by National ScienceFoundation Grants CCR-0208573 and ECS-0401188.

AUTHORSAli H. Sayed received the Ph.D. degree in electrical engineer-ing in 1992 from Stanford University, Stanford, California. Heis currently a professor and vice chair of electrical engineer-ing at the University of California, Los Angeles. He is also theprincipal investigator of the UCLA Adaptive SystemsLaboratory (www.ee.ucla.edu/asl). He has authored over 200journal and conference publications and several books includ-ing the textbook Fundamentals of Adaptive Filtering (Wiley,2003). He has served on the program committees of severalinternational meetings, and has also consulted with industry.His research interests include adaptive and statistical signalprocessing, filtering and estimation theories, signal process-ing for communications, interplays between signal processingand control methodologies, system theory, and fast algo-rithms for large-scale problems. He has received manyawards, including the 1996 IEEE Donald G. Fink Award, the2002 Best Paper Award from the IEEE Signal ProcessingSociety (SPS), and the 2003 Kuwait Prize in Basic Sciences.He has held many roles within the IEEE Signal ProcessingSociety, including a member of the publication and awardboards. He is on the editorial board of IEEE Signal ProcessingMagazine and is editor-in-chief of IEEE Transactions onSignal Processing. He is a Fellow of the IEEE.

Alireza Tarighat received the B.Sc. degree in electricalengineering from Sharif University of Technology, Tehran,Iran, in 1998 and the M.Sc. degree in electrical engineeringfrom the University of California, Los Angeles (UCLA), in2001. He is pursuing the Ph.D. degree in electrical engineer-ing at UCLA. He received the Gold Medal of the NationalPhysics Olympiad, Iran, 1993, and the Honorable MentionDiploma of the 25th International Physics Olympiad, Beijing,China, 1994. His research focuses on signal processing tech-niques for communication systems, including MIMO OFDMreceiver design and multi-user MIMO communications. He isa student member of the IEEE.

Nima Khajehnouri received the B.Sc. degree in electricalengineering from Sharif University of Technology, Tehran,Iran, in 2001 and the M.Sc. degree in electrical engineeringfrom the University of California, Los Angeles (UCLA), in2002 with emphasis on signal processing. Since 2002, he hasbeen pursuing the Ph.D. degree in electrical engineering atUCLA. His research focuses on signal processing techniquesfor communication systems, including multiuser MIMO com-munications, relay networks, and signal processing for sen-sor networks.

REFERENCES[1] “Report on wireless location-based markets,” PELORUS Group, NJ, Tech. Rep.,2001.

[2] “Location-based services: Finding their place in the market,” In-Stat/MDR, SanJose, CA, Tech. Rep., Feb. 2003.

[3] A.H. Sayed and N.R. Yousef, “Wireless location,” in Wiley Encyclopedia ofTelecommunications, J. Proakis, Ed. New York: Wiley, 2003.

[4] F. Gustafsson and F. Gunnarsson, “Mobile positioning using wireless networks,”IEEE Signal Processing Mag., vol. 22, no. 4, pp. 41–53, July 2005.

[5] “FCC Docket No. 94–102. Revision of the commissions rules to insure compati-bility with enhanced 911 emergency calling systems,” Federal CommunicarionsCommission Tech. Rep. RM-8143, July 1996.

[6] “Report on the New Jersey wireless enhanced 911 terms: The first 100 days,”State of New Jersey, Tech. Rep., June 1997.

[7] A.H. Sayed, Fundamentals of Adaptive Filtering. New York: Wiley, 2003.

[8] J.E. Maloney and J.O. Stevenson, “Robust, efficient, localization system,” U.S.Patent 6 047 192, May 13, 1997.

[9] O. Hilsenrath and M. Wax, “Radio transmitter location finding for wirelesscommunication network services and management,” U.S. Patent 6 026 304, Jan. 8,1997.

[10] M. Wax and O. Hilsenrath. “Signature matching for location determination inwireless communication systems,” U.S. Patent 6 108 557, Feb. 18, 1998.

[11] J.P. Kennedy, C.D. Roller, and R.W. Hooper. “Passive high accuracy geoloca-tion system and method,” U.S. Patent 5 317 323, Mar. 5, 1993.

[12] J.J. Caffery and G.L. Stuber, “Overview of radiolocation in CDMA cellular sys-tems,” IEEE Commun. Mag., vol. 36, no. 4, pp. 38–45, Apr. 1998.

[13] J.J. Caffery and G.L. Stuber, “Radio location in urban CDMA microcells, inProc. IEEE Int. Symp. Personal, Indoor and Mobile Radio Communications,Toronto, Canada, Sept. 1995, vol. 2, pp. 858–862.

[14] M. Yunos, J. Zeyu Gao, and S. Shim, “Wireless advertising’s challenges andopportunities,” IEEE Computer, vol. 36, no. 5, pp. 30–37, May 2003.

[15] G. Sun, J. Chen, W. Guo, and K.J.R. Liu, “Signal processing techniques in net-work-aided positioning,” IEEE Signal Processing Mag., vol. 22, no. 4, pp. 12–23,July 2005.

[16] L.A. Stilp, “Carrier and end-user application for wireless location systems,” inProc. SPIE, Philadelphia, PA, Oct. 1996, vol. 2602, pp. 119–126.

[17] X. Li and K. Pahlavan, “Super-resolution TOA estimation with diversity forindoor geolocation,” IEEE Trans. Wireless Commun., vol. 3, pp. 224–234, Jan. 2004.

[18] L. Cong and W. Zhuang, “Hybrid TDOA/AOA mobile user location for wide-band CDMA cellular systems,” IEEE Trans. Wireless Commun., vol. 1, no. 3, pp.439–447, July 2002.

[19] A.J. Weiss, “On the accuracy of a cellular location system based on RSS meas-urements,” IEEE Trans. Veh. Technol., vol. 52, no. 6, pp. 1508–1518, Nov. 2003.

[20] I. Paton et al., “Terminal self-location in mobile radio systems,” in Proc. 6thInt. Conf. Mobile Radio and Personal Communications, Coventry, UK, Dec. 1991,vol. 1, pp. 203–207.

[21] J.J. Caffery and G.L. Stuber, “Vehicle location and tracking for IVHS in CDMAmicrocells,” in Proc. IEEE Int. Symp. Personal, Indoor and Mobile RadioCommunications, Amsterdam, The Netherlands, Sept. 1994, vol. 4, pp. 1227–1231.

[22] J.J. Caffery, Wireless location in CDMA cellular radio systems. Norwell, MA:Kluwer, 2000.

[23] K.W. Cheung, H.C. So, W.K. Ma, and Y.T. Chan, “Least squares algorithms fortime-of-arrival-based mobile location,” IEEE Trans. Signal Processing, vol. 52, no.4, pp. 1121–1130, Apr. 2004.

[24] J.J. Caffery and G.L. Stuber, “Subscriber location in CDMA cellular networks,”IEEE Trans. Veh. Technol., vol. 47, no. 2, pp. 406–416, May 1998.

[25] N.R. Yousef, A.H. Sayed, and L.M.A. Jalloul, “Robust wireless location over fad-ing channels,” IEEE Trans. Veh. Technol., vol. 52, no. 1, pp. 117–126, Jan. 2003.

[26] M.A. Spirito, “On the accuracy of cellular mobile station location estimation,”IEEE Trans. Veh. Technol., vol. 50, no. 3, pp. 674–685, May 2001.

[27] A.J. Weiss, “On the accuracy of a cellular location system based on RSS meas-urements,” IEEE Trans. Veh. Technol., vol. 52, no. 6, pp. 1508–1518, Dec. 2003.

[28] N.R. Yousef and A.H. Sayed, “Detection of fading overlapping multipath com-ponents for mobile-positioning systems,” in Proc. Int. Conf. Communications,Helsinki, Finland, June 2001, vol. 10, pp. 3102–3106.

[29] J. Friedmann, R. Raich, J. Goldberg, and H. Messer, “Bearing estimation for adistributed source of nonconstant modulus—Bounds and analysis,” IEEE Trans.Signal Processing, vol. 51, no. 12, pp. 3027–3035, Dec. 2003.

[30] “The CDMA2000 ITU-R RTT candidate submission V0.18,”Telecommunications Industry Association, July 1998.

[31] K.C. Ho and Y.T. Chan, “Solution and performance analysis of geolocation byTDOA,” IEEE Trans. Aerosp. Electron. Syst., vol. 29, no. 4, pp. 1311–1322, Oct.1993.

[32] Y.T. Chan and K.C. Ho, “A simple and efficient estimator for hyperbolic loca-tion,” IEEE Trans. Signal Processing, vol. 42, no. 8, pp. 1905–1915, Aug. 1994.

[33] R. Schmidt, “Least squares range difference location,” IEEE Trans. Aerosp.Electron. Syst., vol. 32, no. 1, pp. 234–242, Jan. 1996.

[34] B.T. Fang, “Simple solutions for hyperbolic and related position fixes,” IEEETrans. Aerosp. Electron. Syst., vol. 26, no. 5, pp. 748–753, Sept. 1990.

[35] M. Hellebrandt and R. Mattar, “Location tracking of mobiles in cellular radionets, IEEE Trans. Veh. Technol., vol. 48, no. 5, pp. 1558–1562, Sept. 1999.

[36] M. McGuire and K.N. Plataniotis, “Dynamic model based filtering for mobileterminal location,” IEEE Trans. Veh. Technol., vol. 52, no. 4, pp. 1012–1031, July2003.

[37] H. Krim and M. Viberg, “Two decades of array signal processing research: Theparametric approach,” IEEE Signal Processing Mag., vol. 13, no. 4, pp. 67–94, July1996.

[38] T. Ojanpera and R. Prasad, Wideband CDMA for Third Generation MobileCommunications. Boston, MA: Artech House, 1998.

[39] R. Prasad, W. Mohr, and W. Konhauser. Third Generation MobileCommunication Systems. Boston, MA: Artech House, 2000.

[40] K.J. Krizman, T.E. Biedka, and T.S. Rappaport. “Wireless position location:Fundamentals, implementation strategies, and sources of error,” in Proc. IEEEVehicular Technology Conf., New York, NY, May 1997, vol. 2, pp. 919–923.

[41] N. Khajehnouri and A.H. Sayed, “A non-line-of-sight equalization scheme forwireless cellular location,” in Proc. ICASSP, Hong Kong, Apr. 2003, vol. VI, pp.549–552.

[42] M.P. Wylie and J. Holtzman, “The non-line of sight problem in mobile locationestimation,” in Proc. IEEE Int. Conf. Universal Personal Communications, 1996,vol. 2, pp. 827–831.

[43] S.S. Woo, H.R. You, and J.S. Koh, “The NLOS mitigation technique for posi-tion location using IS-95 CDMA networks,” in Proc. Vehicular Technology Conf.,Boston, MA, 2000, vol. 6, pp. 2556–2560.

[44] S. Al-Jazzar, J. Caffery, and H.R. You. “A scattering model based approach toNLOS mitigation in TOA location systems,” in Proc. IEEE Vehicular TechnologyConf., Birmingham, AL, 2002, vol. 2, pp. 861–865.

[45] L.J. Greenstein, V. Erceg, Y.S. Yeh, and M.V. Clark, “A new path-gain/delay-spread propagation model for digital cellular channels,” IEEE Trans. VehicularTechnology, vol. 46, pp. 477–485, May 1997.

[46] A. Ghosh and R. Love, “Mobile station location in a DS-CDMA system,” in Proc.IEEE Vehicular Technology Conf., Ottawa, Canada, May 1998, vol. 1, pp. 254–258.

[47] K. Kuboi, S. Shirota, S. Sakagami, S. Aoyama, and A. Akeyama, “Vehicle posi-tion estimates by multibeam antennas in multipath environments,” IEEE Trans.Veh. Technol., vol. 41, no. 1, pp. 63–68, Feb. 1992.

[48] G.P. Yost and S. Panchapakesan, “Automatic location identification using ahybrid technique,” in Proc. IEEE Vehicular Technology Conf., Ottawa, Canada,May 1998, vol. 1, pp. 264–267.

[49] X. Wang, Z. Wang, and B. O’Dea, “A TOA-based location algorithm due toNLOS propagation,” IEEE Trans. Veh. Technol., vol. 52, no. 1, pp. 112–116, Jan.2003.

[50] M. McGuire, K.N. Plataniotis, and A.N. Venetsanopoulos, “Location of mobileterminals using time measurements and survey points,” IEEE Trans. Veh.Technol., vol. 52, no. 4, pp. 999–1011, July 2003.

[51] A. Tarighat, N. Khajehnouri, and A.H. Sayed, “CDMA location using multipleantennas and interference cancellation,” in Proc. IEEE Vehicular TechnologyConf., Apr. 2003, vol. 4, pp. 2711–2715.

[52] C. Botteron, A. Hst-Madsen, and M. Fattouche, “Cramer-Rao bounds for the esti-mation of multipath parameters and mobiles’ positions in asynchronous DS-CDMAsystems,” IEEE Trans. Signal Processing, vol. 52, no. 4, pp. 862–875, Apr. 2004.

[53] J. Krumm, S. Harris, B. Meyers, B. Brumitt, M. Hale, and S. Shafer, “Multi-camera multi-person tracking for EasyLiving,” in Proc. 3rd IEEE Int. WorkshopVisual Surveillance, Dublin, Ireland, July 2000, pp. 3–10.

[54] N. Priyantha, A. Miu, H. Balakrishman, and S. Teller, “The Cricket compass forcontext-aware mobile applications,” in Proc. ACM/IEEE Int. Conf. MobileComputing and Networking (MOBICOM 2001), Rome, Italy, July 2001, pp. 1–14.

[55] A. Ward, A. Jones, and A. Hopper, “A new location technique for the activeoffice,” IEEE Trans. Personal Communications, vol. 4, pp. 42–47, Oct. 1997.

[56] R. Want, A. Hopper, V. Falco, and J. Gibbons, “The Active Badge LocationSystem,” Proc. ACM Trans. Inform. Syst., vol. 10, no. 1, pp. 91–102, Jan. 1992.

[57] R.J. Orr and G.D. Abowd, “The Smart Floor: A mechanism for natural useridentification and tracking,” Proc. Conf. Human Factors in Computing Systems(CHI 2000), The Netherlands, Apr. 2000, pp. 1–6.

[58] P. Bahl and V.N. Padmanabhan, “RADAR: An in-building RF-based user loca-tion and tracking system,” in Proc. IEEE Conf. INFOCOM 2000, Tel Aviv, Israel,Mar. 2000, vol. 2, pp. 775–784.

[59] M. Youssef, A. Agrawala, and A.U. Shankar, “WLAN location determination viaclustering and probability distributions,” Proc. IEEE Conf. PerCom 2003, Mar. 2003, pp. 143–150.

[60] P. Castro, P. Chiu, T. Kremenek, and R. Muntz, “A probabilistic location servicefor wireless network environments,” in Proc. Ubiquitous Computing 2001,Portugal, Sept. 2001.

[61] S. Gezici, Z. Tian, G.B. Giannakis, H. Kobayashi, A.F. Molisch, H.V. Poor, andZ. Sahinoglu, “Localization via ultra-wideband radios,” IEEE Signal ProcessingMag., vol. 22, no. 4, pp. 70–84, July 2005.

[62] N. Patwari, A.O. Hero III, J. Ash, R.L. Moses, S. Kyperountas, and N.S. Correal,“Locating the nodes,” IEEE Signal Processing Mag., vol. 22, no. 4, pp. 54–69, July2005.

[63] T. Ross, P. Myllymaki, and H. Tirri, “A statistical modeling approach to loca-tion estimation,” IEEE Trans. Mobile Comput., vol. 1, no. 1, pp. 59–69, Jan. 2002.

[64] M. Youssef and A.K. Agrawala, “Continuous space estimation for WLAN loca-tion determination systems,” in Proc. IEEE 13th Int. Conf. ComputerCommunications and Networks, Chicago, IL, Oct. 11–13, 2004, pp. 143–150.

[65] T. Roos, P. Myllymaki, H. Tirri, P. Misikangas, and J. Sievanen, “A probabilisticapproach to WLAN user location estimation,” Proc. Int. J. Wireless Informat.Networks, vol. 9, no. 3, July 2002, pp. 155–164.

[66] A.M. Ladd, K. Bekris, A. Rudys, G. Marceau, L.E. Kavraki, and D.S. Wallach,“Robotics-based location sensing using wireless ethernet,” Proc. IEEE 8th ACMMOBICOM, Atlanta, GA, Sept. 2002, pp. 189–204.

[67] UCLA Adaptive Systmes Laboratory [Online]. Electrical Engineering Department,Univ. California, Los Angeles, CA. Available: http://www.ee.ucla.edu/asl


[SP]

Documents

Network-Based Wireless Location · 2018-06-21 · Network-based location technology, on the other hand, relies on some existing networks (either cellular or WLAN) to determine the