Content uploaded by Odej Kao
Author content
All content in this area was uploaded by Odej Kao
Content may be subject to copyright.
Suitability of Positioning Techniques for
Location-based Services in wireless LANs
Ulf Rerrer and Odej Kao a
aDepartment of Computer Science, University of Paderborn, Germany
email: {urerrer,okao}@upb.de
Abstract - The steady rise of mobile computing devices and local-area wireless
networks has fostered a growing interest in location-aware systems and services.
Local services will be one of the most exiting features of the next generation wireless
systems. This paper gives an overview of the main concerns of location-based
services (LBS) which are already established in the area of mobile phone networks.
Transferring LBS to an indoor wireless network, especially WLAN, is the next
step. As a first step for any kind of LBS a positioning system is required. Different
techniques exist in all major wireless networks which have basic procedures in
common. This paper presents, analyses and categorises state-of-the-art techniques
for indoor positioning and examines the suitability to support LBS in WLANs.
1 Introduction
The emerging world of mobility is characterised by a multiplicity of new technologies, applica-
tions, and services. With the ability of mobility, location identification has naturally become a
critical attribute, as it opens the door to a world of applications and services that were unthink-
able only a few years ago. The steady rise of mobile computing devices and local-area wireless
networks has also fostered a growing interest in location-aware systems and services. Our every-
day life changes because Internet applications are nowadays involved everywhere. Activities like
gathering information from any networked site on world, exchanging documents, information
transfer via e-mail, telephone and video conferences, e-learning or web-shopping are done every-
where, anytime. The common characteristics of all these operations are the elimination of the
spatial distance between the involved partners and the time constraints. Furthermore, the infor-
mation can be accessed with different devices, from classic hardwired PCs via WLAN-capable
notebooks and PDAs to mobile phones.
The introduction of wireless communication in recent years opened new opportunities for
communication and interaction provided by so-called location-aware or location-based systems
and services. A key feature of such systems is that the application information and/or interface
presented to a user is a function of his or her physical location. The granularity of position
information needed could vary from one application to another. For example, locating a nearby
printer requires coarse-grained positioning information whereas locating a book in a library
requires a more fine-grained information.
The rest of the paper is organised as follows: In Section 2, we describe location-based services,
categorise them and point out the main concerns influencing location-aware services while in
Section 3 a technology overview for indoor-positioning is given. Section 4 outlines the potential
of WLAN for positioning and summarises techniques and algorithms. Conclusions and future
work are presented in Section 5.
2 Location-based services
Services depending on the current position of the user are well known in mobile phone networks.
Provider established different commercial location-based services (LBS) like a ”friend finder”
[1] or a ”nearest hotel finder” to bring more comfort to their customers. These services can be
generalised and transferred to other wireless networks. In this paper we verify the requirements
and the possibilities to apply LBS in wireless LAN (IEEE 802.11). The underlying positioning
technique is essential for any application build on this basis and is discussed in Section 3.
2.1 Categories for LBS
Services based on the location of devices have different accuracy needs, but can be categorised
by the provided service type [7] as follows:
Emergency The clearest application for location-based services is summarised in this category.
Individuals – unaware of their exact location – in a case of an emergency (injury, criminal
attack, etc.) use their mobile device. In a case of life-threatening injuries a call for
assistance is possible automatically revealing the exact users location and alarm emergency
forces immediately using the positioning capability of the mobile device. LBS of this
category can be applied indoor and outdoor.
Navigation Entering a foreign city or area the user’s needs for directions within the current
geographical location can by satisfied with applications in this category. These services
allow to find special places (shops, hotels, gas stations, etc.) depending on the users
location with detailed maps or route descriptions (positions, directions, traffic conditions,
points of interest) transferred to the mobile device. Well established car navigation systems
already provide these services. The potential of general LBS in this category is the indoor
navigation combining the flexibility and functionality of the mobile devices like handhelds.
Information Providing information to a user depending on his/her position are placed in this
category. Travel services provide information about local sightseeing objects, yellow page
services notify users of special local institutions like public swimming pools or tourist
centers. Infotainment services notify about local events or location-specific multimedia
content to interested users.
Discovery and Tracking Services in this category help to find lost things or persons. Exam-
ples are finding stolen cars or lost children and elderly people in malls. Similar applications
allows companies to locate their field personnel (salesman or maintenance/repair crews)
or even product tracking for supply chain management.
Billing Depending on the location of users different location-sensitive billing systems can be
applied. With these systems consumers can be dynamically charged like a network operator
can price calls from mobile phones when the user is in his ”home zone”.
The main objective of this paper is to show the transfer of general location-based services
to an indoor wireless data network. Therefore the demands and requirements of every above
mentioned category have to analysed and verified. Therefore the main concerns for LBS are
introduced in the next section.
2.2 Main concerns for LBS
To decide which positioning technique is suitable for location-based services in wireless LANs
it is necessary to look at the major needs and targets which have a great influence on the
acceptance and usefulness of future applications and services (see Section 4). Those aspects can
be formulated as questions:
Scenario/Category In which category (see Section 2.1) can the planned scenario be classified?
Does the chosen technology and positioning technique inhibit future applications?
Technology Does the chosen technology need additional hardware on the infrastructure side
and the client side?
Accuracy Which fluctuations in terms of accuracy are acceptable to provide the service? What
is the minimum mean accuracy and worst case accuracy?
Interference Tolerance Outdoor wireless networks suffer from atmospheric interferences, in-
door wireless networks from shadowing and multi path propagation because of reflection
and refraction. Do the services tolerate these interferences?
Scalability When location-based services grow from one building to several buildings or even
many separated blocks of buildings, does the positioning technology scale as well as the
services? What can be done if a higher accuracy is required?
Security and Privacy Remote positioning and device tracking harm the users privacy. The
users admission is usually needed. Security aspects to guarantee only authorised access
have to be applied.
Place of Data Collection The location-sensible data can be collected and handled on server-
side or on client-side. Both places have pro and cons.
2.3 General location techniques
Many ways exist to locate a user or respectively a mobile device in a wireless network. A po-
sition can be symbolic (informal or abstract ”the station is in the center of the city”), absolute
or relative to some point [8]. Some general techniques are common in all different networks
(infrared, Bluetooth, RF-based, mobile phone or satellite-based networks). They can be cate-
gorised/described as follows:
Cell-of-Origin This technique is easy to realise and determines the current base stations (one
or more) to which the mobile device is currently connected. These base stations (BS) exist
in all wireless networks (mobile phone base stations, access point in WLAN, etc.). Due to
the known position and BS range a relatively exact position can be determined.
Signal Strength In this category the exact location of a mobile device is determined using
the current signal strength from a device to a BS. The signal strength decreases with
increasing distance, but multi path fading and shadowing have a dominant effect in indoor
and outdoor environments. Measurements to more than one BS can improve the accuracy.
Time-based A more accurate techniques than signal strength or cell-of-origin are time-based
systems. Approaches like time of arrival (TOA) and time difference of arrival (TDOA)
belong to this category. Similar to TOA is angle of arrival (AOA). Both determine the
position according to the time of a received signal [15]. This allows a high accuracy which
can be increase by repetitions or the number of BS and is also effected by multi path
fading and shadowing. A disadvantage is the need for a precise clock in the mobile device
for synchronisation.
These techniques are frequently used. They can be enhanced by additional effort considering
probability distributions [19, 13] or motion tracking [11] of device locations. Even neural networks
[18] are employed to achieve a higher accuracy.
3 Technology overview for indoor positioning
The major question arising from Section 2.2 is which underlying technology for an indoor posi-
tioning is the most suited for custom applications. Innumerable examples in recent years worked
at the problem to determine and track the users position in an indoor wireless network. The
satellite-based Global Positioning System (GPS) is established for outdoor services, but to make
GPS available within buildings can only be achieved with correction technology as in assisted-
GPS under great expenses. Infrared-based systems like ActiveBadges [17] are frequently used
for indoor systems, but suffer from short range transmitters and the huge amount of additional
hardware. Ultrasonic waves are another established and mature positioning technology used
in systems like Cricket [12]. They also need a lot of additional hardware and have a tolerable
accuracy.
Unfortunately, all these technologies suffer from significant drawbacks like poor scalabil-
ity, need for extensive deployment of sensors and/or prohibitive expenditure for deployment
and maintenance. The most promising technology uses radio frequency (RF) electro-magnetic
waves. RFID devices as a passive component and WLAN devices as an active components allows
a precise and low-cost positioning. Most systems [3, 6, 14] focus on the latter devices and its
technology because such systems have the potential to leverage existing data networking func-
tionality with location-aware services since the protocols in WLAN are robust and technically
mature. RADAR [3] developed at Microsoft Research was one of the first systems in this field
implementing simple algorithms to calculate the users position on a central server. RADAR
inspired many other researcher to develop a big variety of algorithms for more accurate posi-
tioning (see next Section) because generalising the system to multi-floored buildings or three
dimensions is still a nontrivial problem.
4 Suitable positioning for LBS in WLAN
To build a successful architecture for location-based services in WLANs many aspects have
to be considered (see Section 2.2). Using a pre-existing infrastructure with access points has
advantages in low costs and no necessary constructional changes. The access point have the
technology, coverage and density for an indoor positioning system.
Depending on the application an accuracy of a few meters is adequate – this accuracy can
be sufficient for every service categorised in Section 2.1. Simple positioning with a cell-of-origin
technique would not be appropriate because the cells are to wide [10] and due to the WLAN
standard [9] devices keep connections to access point even if a stronger signal from a new one
is available. More accurate techniques use signal strength to determine the devices’s location.
Here the positioning exactness is disturbed by reflections, refractions and other interferences,
but can be increased using intelligent algorithms. Algorithms like weighted k nearest neighbour
[4], bayesian algorithm [5] compare current measurement points with a database of reference
values do estimate the current position. For services in the information or billing category this
technique is enough.
Other approaches use more complex algorithms computing probabilistic signal strength dis-
tributions [19, 13] to achieve a higher accuracy for services in the navigation or emergency
category. Instead of calculating Euclidian distances to known signal vectors these algorithms
model how the signal strengths are distributed in different geographical areas based on a sample
of measurements collected at several known stations. A probabilistic framework is a complex
system which needs a long training phase, but achieves a high accuracy.
Motion tracking can feed history monitoring algorithms with additional information [2].
Using the premise that the mobile user cannot switch from one set of coordinates to another
totally arbitrary location, from one instant of time to the next helps to report a better position.
If privacy and/or security issues prohibit any history monitoring only approaches which try to
correct the current measurements with a differential correction [16] are suited. To conclude all
these principles the best solution for a positioning technique for WLAN is based on the pre-
existing infrastructure using the access points enhanced by probability distribution approaches.
In order to proof the suitability of WLAN positioning techniques for location-based services
we present a sample scenario for a prototype. A user with a wireless LAN-capable device enters
a region in the range of an access point (AP). The device registers and receives – open network
assumed – local address, gateway and DNS information. Subsequently, the AP transmits all this
information to a device management service, which stores the current device position along with
additional properties such as installed operating system, activated browser, available bandwidth
etc. Finally, the accuracy of the determined position is stored. The value of this attribute
expresses the uncertainty due to the location computation based on signal characteristics. If
two and more APs are reachable by the device, the accuracy is increased significantly.
The created device table is used in twofold manner. The contained information is forwarded
to the service management in order to filter the relevant data accordingly. On the other hand
this information can be used for broadcasting messages, for example all users in the range of
the AP can be identified immediately in case of emergency. In following, solely the usage of the
position for tailoring location-aware information will be described. The gained position data is
forwarded to the service management component, which firstly selects all services available for
the current device position. Subsequently, the activated browser presents a list, where the user
can choose the desired services. The response is used to activate the corresponding services,
which can be located anywhere on the Internet. In case of internal services like maps, solely
the relevant region has to be determined. A number of database queries delivers the contained
hardware devices in that area and their status, the personnel information about the people
working in the offices around etc. In case of external services such as bus departure schedule, an
adapted agent is used to query the provider’s web-site entering the relevant bus station and the
current time. The local service component must solely format the delivered data for the given
properties of the users display. The position of the user is monitored permanently. As soon as
the position changes the new data is given to the service management component in order to
evaluate, which services has to be updated. For example, the map and device information will
be adapted to the new current position. On the other hand no action for the bus departure
schedule is required in this case. The schedule can be updated e.g. every hour or on demand.
5 Conclusion
The main objective of this paper is to show the transfer of location-based services – which are well
known in the field of mobile phone networks – to an indoor wireless network, especially wireless
LAN. As a first step a positioning technique is required to this kind of services. Therefore a
categorisation of established services and general techniques for positioning is presented. After
analysing relevant existing systems the best suitable technique is promoted. It reflects the trend
for WLANs in mobile networks and builds a foundation for LBS in WLANs.
As future work, we plan to develop an architecture for location-based services using WLAN
positioning techniques. Furthermore, we aim to integrate semantic web approaches to support
dynamic discovery and composition of services.
References
[1] AT&T-Wireless, http://www.attws.com/mmode/features/findit/FindFriends/, 2001
[2] V. Abhijit, C. Ellis, and X. Fan ”Experiences with an Inbuilding Location Tracking System:
Uhuru” IEEE Personal, Indoor and Mobile Radio Communications (PIMRC’03), 2003
[3] P. Bahl and V. N. Padmanabhan. ”RADAR: An In-Building RF-based User Location and
Tracking System” Proceedings of IEEE InfoCom’00, Vol. 2, pp. 775–784, March 2000
[4] R. Battiti and M. Brunato ”Statistical Learning Theory for Location Fingerprinting in
Wireless LANs”, October 2002
[5] P. Castro, P. Chiu, T. Kremenek, and R. Muntz ”A Probabilistic Room Location Service
for Wireless Networked Environments” Ubiquitous Computing (UBICOMP’01), 2001
[6] Ekahau, ”Ekahau Positioning Engine”, http://www.ekahau.com, 2004
[7] G. M. Giaglis. ”Towards a Classification Framework for Mobile Location Services” Mobile
Commerce: Technology, Theory, and Applications, pp. 67-85, 2002.
[8] J. Hightower and G. Borriello. ”Location Systems for Ubiquitous Computing” IEEE Com-
puter Journal, 34(8), pp.57–66, August 2001
[9] The IEEE Standard 802.11 - Wireless LAN Medium Access Control (MAC) and Physical
Layer (PHY) specifications, 1999
[10] O. Kao and U. Rerrer ”A Platform for Location-Aware, Ad-hoc Collaboration in Wireless
Networks” Workshop on Positioning, Navigation and Communication (WPNC’04), 2004
[11] H. Levy and Z. Naor. ”Active Tracking: Locating Mobile Users in Personal Communication
Service Networks” ACM Wireless Networks Journal, 5(6), pp. 467-477, 1999
[12] N. B. Priyantha, A. Chakraborty, and H. Balakrishnan. ”The Cricket Location-Support
System” Proceedings of ACM MobiCom’00, August 2000
[13] T. Roos, P. Myllymaki, H. Tirri, P. Misikangas, and J. Sievanen. ”A Probabilistic Approach
to WLAN User Location Estimation” Journal of Wireless Information Networks, 9(3), 2000
[14] H. Tarumi, K. Morishita, M. Nakao, and Y. Kambayashi. ”SpaceTag: An Overlaid Virtual
System and its Applications” Proceedings of ICMCS’99, Vol. 1, pp. 207–212, 1999
[15] M. Vossiek, L. Wiebking, P. Gulden et al. ”Wireless Local Positioning - Concepts, Solutions,
Applications” Proceedings of IEEE RAWCON’03, pp. 219–224, 2003
[16] Y. Wang, X. Jia, H.K. Lee and G.Y. Li ”An indoor wireless positioning system based on
WLAN infrastructure” 6th International Symposium on Satellite Navigation Technology
Including Mobile Positioning & Location Services, July 2003
[17] R. Want, A. Hopper, V. Falco, and J. Gibbons. ”The Active Badges Location System”
ACM Transactions on Information Systems, 10(1):91-102, January 1992
[18] M. Winter and G. Taylor. ”Modular neural networks for map-matched GPS positioning”
IEEE Web Information Systems Engineering Workshops, pp. 106-111, December 2003
[19] M. A. Youssef, A. Agrawala, and A. U. Shankar ”WLAN Location Determination via
Clustering and Probability Distributions” Proceedings of IEEE PerCom’03, 2003
