2011 International Conference on Indoor Positioning and Indoor Navigation (IPIN), 21-23 September 2011, Guimarães, Portugal

A Smart-Phone Indoor/Outdoor
Localization System
Carlos Pereira, Ludimar Guenda and Nuno Borges Carvalho
Instituto de Telecomunicações, Dep. Electrónica, Telecomunicações e Informática
Universidade de Aveiro

Abstract—This paper depicts the operation of an integrated From that moment on, all friends can access this position
system designed for both outdoor and indoor localization, and see it on the map.
with special emphasis on the latter. It is demonstrated how A website where the users can register and see their
three different methods of localization can be combined, locations is also available. It is a useful functionality to
based on WiFi access points and mobile cells. Not only there
track mobile devices (see Fig. 1).
is no need to know the exact location of the access points to
start using the system, but it also self-improves its own
accuracy over time. The application developed to implement
this method works on Android OS running devices and it is
publicly available for download with the name “Locate Me”.
Keywords—Localization, Android, GPS, Wireless Networks,
indoor positioning, WiFi, Locate Me

I. INTRODUCTION
In past recent years there has been a rising usage of
localization systems, both for personal and object
tracking. The Global Positioning System (GPS) is the
evidence of that. However, it has a determinant
drawback: it works neither for indoor environments nor
when surrounded by big structures. For this reason, Figure 1. View of the website
alternative methods of localization have been studied.
At the same time and due to their decreasing cost,
there are WiFi access points almost everywhere. It makes
Coordinates
them very attractive to use in indoor localization systems.
obtained by
There are many studies already made with different
map pointing
approaches using WiFi access points. One example of
those methods is the localization by the roadside WiFi
networks [1]. This method uses sectorized antennas to
estimate the location of the access points. It is an accurate
method but needs specific equipment to build the
database. Other techniques of localization assume that the
exact location of the access points is known and estimate
the location of the mobile device taking into account the
received signal strength [2]. They also use filters to
minimize the fading effects that are predominant in
indoor environments. A more complex method like
SMART [3] uses other parameters besides WiFi signal to
increase the accuracy of the results. They use the
microphone, the camera and the accelerometer modules
to measure additional information and create precise
fingerprints. Figure 2. Add place example
In this work, we try to combine some of those
techniques and develop a functional application for One of the main advantages of the now proposed
Android OS running systems. It was called “Locate Me”. approach is that the users are allowed to add WiFi places
It makes use of the GPS module to get the location of to the database, which means that the locations’ database
the device in outdoor environments, but when it cannot can be updated manually, taking only into account the
find signal (the GPS module does not return any users’ location input, the mobile cells and APs sweep
coordinates to the operating system), it uses alternative (Fig. 2). This WiFi places works by selecting a position
methods to be located. These will be explained later. in the map, manually and adding a corresponding text tag.
In summary, it is a localization system to find friends
wherever they are. The application sends the current
location of the device to the server where it is stored.
 2011 International Conference on Indoor Positioning and Indoor Navigation (IPIN), 21-23 September 2011, Guimarães, Portugal

II. LOCATE ME – INDOOR/OUTDOOR LOCALIZATION

SYSTEM
Locate Me is a practical application of outdoor and
indoor location methods. It was developed for mobile
devices running Android OS and takes advantage of the
GPS and WiFi modules to acquire the location of a
person. To represent the users’ location it uses the Google
Maps Android API and it has two views available: road
view and satellite view (Fig. 3).

Figure 4. Locate Me print screen with plan

Figure 3. Map viewer

Besides that, there is an option to draw a building plan

overlay in the map. For this reason, it is possible to get
accurate representations of the users’ locations,
particularly inside buildings with a great number of WiFi
access points and using the fingerprint localization
method previously explained.
Fig. 4 shows an example of this type of
representation:
In addition to those functionalities it is possible for
the users to add friends in order to see their locations.
They will also be represented on the map with (see
Fig. 4).
As a complement to the Locate Me mobile application
there is also a website available. On the website it is
possible to track your own devices by adding them to Figure 5. Localization request sequence
your account. There is a history mode where it is possible
to see the last 100 positions. 1) Fingerprint localization
The fingerprint method is an alternative localization
III. LOCALIZATION ALGHORITMS method intended to be used in the absence of GPS signal.
This location system is based on 4 different methods This localization process is initiated by the user, who
of localization, one for outdoor and three for indoor points his current location on the map. The coordinates of
environments. this position will then be registered by the system. After
As shown in Fig. 5, the first location method used that, the device performs a scan for WiFi access points
(after a location request) is GPS. It works in open areas and mobile stations. Finally, the result of this scan, the
but fails when inside of buildings or just surrounded by name, the description and the floor of this new location
them. To avoid these situations we developed 3 are sent to the database. This process is actually an initial
alternative methods of localization: localization by calibration of the systems initiated by the user.
places’ fingerprint, by geo-referenced WiFi access point After the WiFi points are added to the database, they
and by geo-referenced cells. These requests can be can be used to locate the devices, in order to do so the
configured with a minimum period of 10 seconds device perform a WiFi scan for 1 second and send the
between each location request. result to the server. This result includes the MAC address,
 2011 International Conference on Indoor Positioning and Indoor Navigation (IPIN), 21-23 September 2011, Guimarães, Portugal

the RSSI and the SSID of each detected access point. On automatically generated by the application and its
the server side the access points are sorted by received precision is also improved with its use.
power (RSSI) and after that, it uses the algorithm Every time there is a GPS signal available, the system
exemplified on Fig. 6 to try to find the best match performs a WiFi scan and sends the result with the GPS
between the WiFi scan result and the locations coordinates to the server. If a new access point was
fingerprints in the database. detected it will be added to the database. If not, and it
Fig. 6 shows how the algorithm works for a already exists in the database its coordinates will be
hypothetical scan result with 3 WiFi access points: AP_1, recalculated by the following method:
AP_2 and AP_3, with received powers of rssi_1, rssi_2
and rssi_3 respectively, and rssi_1 > rssi_2 > rssi_3 .
_ =
_ + 1
The algorithm searches for AP_1 in the location
_ ∙
_
_  + 
fingerprints. If it finds a match, it tries AP_1 and AP_2
_
_  =

_
simultaneously; if not, it skips the AP_1 and searches for
AP_2. This algorithm works using a recursive function
_  ∙
_
_  ∙
_ +   ∙ 
and as a result of that all access points found will be
_  =

_ ∙
_
_  + 
taken into account.

_  − last calculated position of the access point;
  − last position where access point has been detected;

_  − new position of the access point;

_
_  − mean value of the received signal strength;
 − last received signal strength of the access point

_
_  − new mean value to the received signal strength;

_ − number of updates already made;

_ − number of updates already made plus the
current update;

(*) All signal strengths should be in milliwatt;

(**) The coordinates of the access points (latitude and
longitude) are both calculated by applying this method.

In brief, when an access point is detected, the new

coordinates will be used in a calculation with a result
proportional to the signal strength and the number of
updates, approaching the previous position to the current
one.

3) Localization using georeferenced mobile base

stations
This type of localization is quite similar to the
previous one. However, it provides a worse precision,
since the covered area of a base station is bigger than the
covered area of an access point. The number of detected
cells by a smart-phone is usually much smaller than WiFi
access points and subsequently the amount of
distinguishable places will be reduced. This algorithm
uses the same formulas of the “localization using access
points” to determine the base station localization.
Even taking that into account this method should not
be underestimated because in some situations it could be
the only way to be located.
With a combination of these three localizations
methods plus the GPS, it is possible to get a localization
system for indoor and outdoor environments.
Figure 6. Fingerprint localization algorithm

2) Localization using georeferenced access points IV. DATABASE STRUCTURE

Nowadays, nearly all mobile devices have
This application works close together with a database,
simultaneously a WiFi module and a GPS module, so
the database is structured into four groups of tables
taking advantage of that we can create a database with
specially designed for Locate Me application: users info,
geo-referenced WiFi access points. This database is
building plans, places fingerprints and geo-referenced
access points.
 2011 International Conference on Indoor Positioning and Indoor Navigation (IPIN), 21-23 September 2011, Guimarães, Portugal

1) Users info
In this section the information regarding users is
stored. One table stores the name, status and the device
ID of each user, another one stores the users locations
history and finally ther is also a column to store the
friendly relations.
This is the information shared with the website and
used to display the users’ positions on the map.

2) Building plans
In order to overlay the map with the building plans,
there is a table where “PNG” images are stored together Figure 7. Locate Me users in the world
with the geographical information: left longitude, right
longitude, bottom latitude, top latitude and floor. With 6000

this information it is possible to overlay the plan in the

Number of georreferenced APs

5000
right place and with the right scale on a 2D map.
4000

3) Places fingerprints 3000

As a result of the fingerprint localization it is 2000

necessary to store the detected WiFi access points, their
1000
received power and where they were found. The name,
the description and the floor of the place are also stored. 0

4) Geo-referenced access points

Date
Two tables are used to geo-reference the access
points, one for the WiFi access points and another one for Figure 8. Georeferenced APs in the database
the mobile cells. There is no private data stored on those 450
tables. In case of WiFi access points it just stores the 400
Number of georreferenced cells

MAC (Media Access Control) address, the SSID (Service 350

Set Identifier) and the RSSI (Receive Signal Strength 300
Indicator). 250

200
V. CONCLUSION 150

In this paper we show that it is possible to build a 100

50
robust localization system that works both in indoor and
0
outdoor environments.
The indoor localization can be obtained using
different complementary methods and for this reason the Date
database can be self-constructed and improved by the
Figure 9. Georeferenced cells in the database
users at the same time. As a result of the used methods it
is not necessary to know the exact location of the access
points to start using the localization system. In addition to REFERENCES
this, the more it is used, the more accurate it gets.
[1] A. P. Subramanian, P. Deshpande, G. Jie, and S. R.
After three months of the first version release it was
Das, "Drive-By Localization of Roadside WiFi
possible to achieve about one thousand users all around
Networks," in INFOCOM 2008. The 27th
the world. The users’ dispersion can be seen on Fig. 7.
Conference on Computer Communications. IEEE,
With the fingerprint localization method it is possible
2008, pp. 718-725.
for users to add personal places to be located(for example
[2] J. A. Besada, A. M. Bernardos, P. Tarrio, and J. R.
different room in the house). The number of
Casar, "Analysis of tracking methods for wireless
distighuishable places changes with the detected WiFi
indoor localization," in Wireless Pervasive
access points.
Computing, 2007. ISWPC '07. 2nd International
The geo-referenced WiFi access points and mobile
Symposium on, 2007.
cells registered in the database are also increasing and
[3] Z. Peng, W. Dan, and S. Yi, "SMART:
now there are over five thousand for WiFi and over four
Simultaneous indoor localization and map
hundred for mobile cells. Fig. 8 and Fig. 9 shows these
construction using smartphones," in Neural
evolutions:
Networks (IJCNN), The 2010 International Joint
Conference on, 2010, pp. 1-8.