EVALUATION OF DIRECTORY-LESS WLAN POSITIONING

BY DEVICE WHISPERING

Karl-Heinz Krempels, Sebastian Patzak

Janno von St¨ulpnagel and Christoph Terwelp

Informatik 4, Intelligent Distributed Systems Group

RWTH Aachen University, Aachen, Germany

Keywords:

Indoor positioning, WLAN, Geo Tagging, Whispering, Accuracy.

Abstract:

Existing positioning systems do not provide the required positioning accuracy for navigation systems in indoor

environments. Novel system approaches are based on ﬁngerprinting and triangulation techniques. Thus, they

suffer on low positioning accuracy due to multipath propagation and different sending power of the considered

access points. Other approaches are based on tagged WLAN (Wireless Local Area Network) access points or

GSM (Global System for Mobile Communication) base stations with their corresponding position stored in a

central tag directory. This would cause high communication costs for a mobile device that queries the directory

frequently. In this paper we present a ﬁltering technique for access points to determine the closest ones to the

mobile device. The geographical position of the mobile device is calculated from the geo tags broadcasted

by the access points in the mobile device’s vicinity. Systems based on this approach will provide the same

accuracy as directory-based positioning systems at a low cost. The evaluation of the approach shows that the

positioning accuracy is limited by the technical capabilities of the radios embedded in today’s mobile devices

and the provided driver software.

1 INTRODUCTION

Wireless networks become more present at many

places and the vision of ubiquitous and pervasive

computing becomes true. The current position of a

mobile device is important for navigation and guid-

ing applications as well as for the determination of a

mobile user’s context. Since outdoor positioning ap-

proaches are based on GPS, that does not work in-

doors, due to the limited reception of GPS signals in-

side of buildings, there is a need for indoor position-

ing systems.

This paper discusses a directory-less approach for

WLAN based indoor positioning which can be used

to realize indoor navigation and guidance systems at

airports or railway stations, e.g. to guide the passen-

ger to his gate or to the next restaurant. This approach

does not need any additional server infrastructure or

additional transmitter antennas, because it uses the al-

ready existing WLAN infrastructure.

The paper is organized as follows: In section 2.4

we introduce the WLAN whispering approach and

discuss its merits and ﬂaws. Section 5 discusses a

guiding application scenario for an airport. Finally, in

section 6 we summarize the results of our work and

address open problems in this area of research.

2 DIRECTORY-LESS INDOOR

WLAN POSITIONING

Approaches for indoor positioning based on WLAN

signals are discussed in (Jan and Lee, 2003) (Wall-

baum and Spaniol, 2006) (Wallbaum, 2004) (Yeung

and Ng, 2007) (Kaemarungsi, 2006) (Zhao et al.,

2008), accuracy comparisons are given in (Lin and

Lin, 2005) (Wallbaum and Diepolder, 2005) (Liu

et al., 2007).

Directory-less indoor positioning based on geo-

tags is discussed in (Krempels and Krebs, 2008a)

(Krempels and Krebs, 2008b). In this approach the

geographical coordinates of the access points are di-

rectly provided by the access points them self. A mo-

bile device with an embedded WLAN receiver analy-

ses the signals from the adjacent access points com-

bined with the received information on their positions

and calculates its own geographical position.

139

Krempels K., Patzak S., von Stülpnagel J. and Terwelp C. (2009).

EVALUATION OF DIRECTORY-LESS WLAN POSITIONING BY DEVICE WHISPERING.

In Proceedings of the International Conference on Wireless Information Networks and Systems, pages 139-144

DOI: 10.5220/0002245501390144

 SciTePress

2.1 Service Set Identiﬁer

Service Set Identiﬁers are deﬁned by the IEEE

802.11-1999 (LAN MAN Standards Committee,

1999) standard. For the approach of discourse only

the SSID is useful:

• The SSID indicates the name of the WLAN cell

that is broadcasted in beacons. The length of the

SSID information ﬁeld is 0 to 32 octets.

• Extended Service Set Identiﬁer (ESSID): Multiple

APs with the same SSID are combined to a larger

cell on layer 2. This is called ESSID.

• The Basic Service Set Identiﬁer (BSSID) is a 48-

bit ﬁeld of the same format as an IEEE 802.11

MAC address. It uniquely identiﬁes a Basic Ser-

vice Set (BSS). Normally, the value is set to the

MAC address of the AP or a broadcast MAC ad-

dress in an infrastructure BSS.

To supply the geographical coordinates of the

wireless access points to the mobile device the fol-

lowing two interaction modes can be used:

2.2 Pull Model

Every wireless access point (AP) broadcasts the same

SSID like ’geo’. Then, the client associates with the

AP and obtains an IP address over Dynamic Host

Conﬁguration Protocol (DHCP). Finally, the client

queries a positioning service provided by the access

point to retrieve the GPS coordinate of the AP.

2.3 Push Model

Every AP broadcasts a unique SSID that encodes the

GPS coordinates of the AP. The client needs only to

scan for speciﬁc geo SSIDs and selects the SSID with

the highest signal strength. It is not necessary that

the client associates with the base station, because the

client can retrieve all information from the already re-

ceived SSID broadcast.

2.4 SSID WLAN Positioning

The position of the mobile device could be estimated

with the help of interpolation calculus, by using only

the coordinates of the m strongest signals from n

signals received by the device, or by a combination

of both approaches. Meaning, ﬁrst selecting the m

strongest signals and then interpolating the coordi-

nates related to this signals. However, the result will

be an area or even a space.

Determining the position of the mobile device

only with the help of the strength of the received sig-

nals is highly inﬂuenced by the changing environ-

ment and the changing sending power of the consid-

ered access points. Thus, we can not assume, that

the strongest signal is received from the closest ac-

cess point. In Figure 1 the signal received from AP

could be stronger than the signal received from AP

Figure 1: WLAN SSID-Positioning.

2.5 SSID WLAN Whispering

In Fig. 1 the mobile device MD receives the signals

and SSID’s from the access points AP

, AP

, . . . , AP

To select the closest geographical vicinity of the mo-

bile device, we introduce the whispering approach.

Since a mobile device is able to control its WLAN ra-

dio interface it can control also its sending power. The

characteristics of its receiving antenna are not inﬂu-

enced thereby, so that the list of access points received

by the mobile device would not change. WLAN radio

whispering (Krempels and Krebs, 2008b) consists in

reducing the sending power of a mobile device to a

minimal value (less than 1mW) and querying a subset

of the visible access points for management informa-

tion (Fig. 2).

Due to the reduced sending power of the mobile

device only the access points, that are geographically

very close to the mobile device will receive its query

and will answer to it. Thus, the effect of whispering

is a ﬁlter that is robust against signal multi-path prop-

agation and power oscillations or automated adaption

of access points. An idealistic abstraction of the whis-

pering effect is shown in Fig. 3.

WINSYS 2009 - International Conference on Wireless Information Networks and Systems

140

Figure 2: Radio Whispering to Detect the Close Vicinity.

Figure 3: Answer of the Close Vicinity.

In the WLAN communication range of the mo-

bile device MD the access points AP

, AP

, . . . , AP

are visible (Fig. 1). AP

and AP

will receive the

information query send with very low power by the

mobile device (Fig. 2) due to their close vicinity to

it. Access point AP

answers to the query (Fig. 3)

and the mobile device can extract its position from

the SSID of AP

3 EXPERIMENTAL SETUP AND

ENVIRONMENT DESCRIPTION

Device whispering requires control of the signal send-

ing power at hardware level and a corresponding

driver that provides this functionality to application

software. Since only a few network drivers met this

requirement Linux was chosen as development plat-

form providing a suitable open source driver. Fig. 4

shows the architectural diagram of the implemented

software.

Network Interface

Firmware

Software

Driver

Application

Hardware

Figure 4: Software Architecture.

The developed application collects the positioning

information from the access points requesting the op-

erating system driver to scan the wiﬁ network. Thus,

the wireless network interface is instructed with the

help of ﬁrmware functions by the wireless interface

driver to perform the required network scan, then to

reduce the sending power to 1mW, and to send out

probe request packets to access points with tagged

SSIDs.

All measurements for the evaluation of the device

whispering approach have been taken at Cologne In-

ternational Airport

. The airport building consists of

two terminals Terminal 1 and Terminal 2 with three

ﬂoors. In both terminals the third ﬂoor is the depar-

ture level. The arrival level is situated at ﬂoor 2 in

Terminal 1 and at ﬂoor 1 in Terminal 2. In Termi-

nal 1 the three ﬂoors have a common open side close

to the elevators and the stairs. Thus, the WLAN sig-

nal strength varies very much in this area, due to the

high dependency of the signal quality from the posi-

tion and direction of the mobile’s WLAN radio an-

tenna. For the validation of the approach we tagged a

set of 13 WLAN access points operated by the com-

puter center of the airport with geo tags. Five access

points are installed in Terminal 1 at departure level,

six access points in Terminal 2 at departure level, and

2 access points in Terminal 2 at arrival level. Figure 5

shows a sketch of the airport building and the position

of tagged access points.

Cologne International Airport

EVALUATION OF DIRECTORY-LESS WLAN POSITIONING BY DEVICE WHISPERING

141

Arrival

Level

Departure

Level

Terminal 2

Terminal 1

Departure

Level

Arrival

Level

Shopping

Mall

Airport

Administration

Figure 5: Position of the tagged access points in the airport

building.

4 EVALUATION

Figure 6 shows a sketch of the airport building and

the position of measurement points 1-19. At each of

this measurement points the real position was deter-

mined manually to get a value to compare to. Then

the position was measured by triangulation with the

access points. Once using the old approach and once

using the new whispering approach. This was done

two times to get an idea of the stability of the mea-

surements. The deviation of the measured positions

from the actual positions is shown in Figure 7 not us-

ing whispering and in Figure 8 using whispering.

If we look at the ﬁgures and at the median of the

measurement errors, which are 45.5m without whis-

pering and 32.5m with whispering, we see an im-

proved positioning quality by about 29%. As we

only did the measurements 2 times, our ability to

make assumptions about the stability of the position-

ing method is limited. But looking at the mean val-

ues of the differences between the two measurements,

which are 23.53m without whispering and 14.37m,

the stability seems to have improved, too.

Terminal 1

Departure

Level

Arrival

Level

Airport

Administration

Shopping

Mall

Arrival

Level

Departure

Level

Terminal 2

Figure 6: Position of the tagged access points in the airport

building.

5 APPLICATION SCENARIO

Many indoor navigation and guidance applications

suffer on high positioning costs and on low position-

ing accuracy. The business cases of a subset of this

systems are based on low cost or free positioning

and do not require high accuracy positioning. Thus,

it seems that even with a low positioning accuracy

(less than twenty-ﬁve meters) navigation and posi-

tioning applications could be deployed and used. In

Fig. 9 a guidance scenario is shown that could be im-

plemented with the help of the positioning approach

discussed in this paper. The scenario is based on a

planned trip consisting of a travel chain. Each el-

ement of the chain has an expected duration and a

travel mode (e.g. walking, ﬂying, traveling by bus,

etc). For the most travel modes the operation vehicle

(e.g. bus, train) and its route is known in advance.

Thus, the positioning accuracy could be improved if a

determined (rough) position is mapped to well known

trajectories of a planned route at the respective time,

e.g. corridors, stairs, etc. The scenario in Fig. 9 shows

a travel chain element with the travel mode walking.

A traveler is guided with the help of discrete position

points mapped to his planned route to the right gate,

e.g to take his plane.

WINSYS 2009 - International Conference on Wireless Information Networks and Systems

142

100

110

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Deviation [m]

Measurement Point Id

Measurement 1

Measurement 2

Figure 7: Deviation of the measured positions without usage of the whispering approach.

100

110

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Deviation [m]

Measurement Point Id

Measurement 1

Measurement 2

Figure 8: Deviation of the measured positions using the whispering approach.

EVALUATION OF DIRECTORY-LESS WLAN POSITIONING BY DEVICE WHISPERING

143

Bus

Walk

Train

Walk

Plane

Gate 1

Gate 2

Gate 29

Gate 30

Figure 9: Application Scenario.

6 CONCLUSIONS

In this paper we presented the whispering technique

to improve the positioning accuracy in directory-less

indoor WLAN positioning. The advantage of this ap-

proach is that there is no need to establish an Inter-

net connection, and it is applicable indoor and out-

door. The positioning accuracy is determined by the

number of access points which can be seen by a mo-

bile device, their radio range and how ﬁne the send-

ing power of the WLAN radio of the device itself can

be adjusted. We could show on the K¨oln-Bonn air-

port that the approach gives better results than WLAN

positioning without whispering. A limiting factor is

the hardware’s and driver’s capability to reduce the

sending power. Current systems are able to reduce

the sending power to minimum of 1mW. But to im-

provethe positioning results further a adjustable send-

ing power between 10µW and 1000µW is required.

So, one future step is to modify the WLAN hardware

to support this low sending power levels. Another ap-

proach can be to combine this approach with other

positioning systems, as for example GPS, to a hybrid

positioning system and expand the 802.11 standard to

support context information for access points.

REFERENCES

Jan, R.-H. and Lee, Y. R. (6-9 Oct. 2003). An Indoor Ge-

olocation System for Wireless LANs. Parallel Pro-

cessing Workshops, 2003. Proceedings. 2003 Interna-

tional Conference on, pages 29–34.

Kaemarungsi, K. (2006). Distribution of wlan received sig-

nal strength indication for indoor location determina-

tion. pages 6 pp.–.

Krempels, K.-H. and Krebs, M. (2008a). Directory-less

WLAN Indoor Positioning. In Proceedings of the

IEEE International Symposium on Consumer Elec-

tronics 2008, Vilamoura, Portugal.

Krempels, K.-H. and Krebs, M. (2008b). Improving

Directory-Less WLAN Positioning by Device Whis-

pering. Proceedings of the International Conference

on Wireless Information Networks and Systems, Porto,

Portugal.

LAN MAN Standards Committee (1999). ANSI/IEEE Std

802.11, 1999 Edition, Part 11: Wireless LAN Medium

Access Control (MAC) and Physical Layer (PHY)

Speciﬁcations. IEEE Standard.

Lin, T.-N. and Lin, P.-C. (2005). Performance compari-

son of indoor positioning techniques based on location

ﬁngerprinting in wireless networks. volume 2, pages

1569–1574 vol.2.

Liu, H., Darabi, H., Banerjee, P., and Liu, J. (2007). Sur-

vey of wireless indoor positioning techniques and

systems. Systems, Man, and Cybernetics, Part C:

Applications and Reviews, IEEE Transactions on,

37(6):1067–1080.

Wallbaum, M. (5-8 Sept. 2004). Tracking of Moving Wire-

less LAN Terminals. Personal, Indoor and Mobile Ra-

dio Communications, 2004. PIMRC 2004. 15th IEEE

International Symposium on, 2:1455–1459 Vol.2.

Wallbaum, M. and Diepolder, S. (19-19 July 2005). Bench-

marking Wireless LAN Location Systems Wireless

LAN Location Systems. Mobile Commerce and Ser-

vices, 2005. WMCS ’05. The Second IEEE Interna-

tional Workshop on, pages 42–51.

Wallbaum, M. and Spaniol, O. (Oct. 2006). Indoor Posi-

tioning Using Wireless Local Area Networks. Modern

Computing, 2006. JVA ’06. IEEE John Vincent Atana-

soff 2006 International Symposium on, pages 17–26.

Yeung, W. M. and Ng, J. K. (21-24 Aug. 2007). Wireless

LAN Positioning based on Received Signal Strength

from Mobile Device and Access Points. Embedded

and Real-Time Computing Systems and Applications,

2007. RTCSA 2007. 13th IEEE International Confer-

ence on, pages 131–137.

Zhao, Y., Zhou, H., Li, M., and Kong, R. (2008). Implemen-

tation of indoor positioning system based on location

ﬁngerprinting in wireless networks. pages 1–4.

WINSYS 2009 - International Conference on Wireless Information Networks and Systems

144