ALGORITHM FOR SELECTION OF CONTENTS IN A

LOCATION-BASED SYSTEM

Adolfo Aumaitre, Vicente Carrillo and F. Javier Ortega

Escuela Técnica Superior de Ingeniería Informática, Avda. Reina Mercedes s/n, 41012, Seville, Spain

Keywords: Location-based services, radio, geo-referenced, algorithm.

Abstract: This article presents a geo-referenced and personalized radio system for mobile terminals that works over

wireless broadband networks. The architecture of the system allows the combination of thematic contents

that are chosen by the user and contents with local information that depends in the first instance on GPS

coordinates, velocity and direction of travel of the user and in the second instance on the preferences of the

user, who can configure its terminal according to its interests. The requesting, downloading and playing of

local information is managed by a predictive algorithm that takes into account not only the position of the

terminal or the preferences of the user but also other features as the area where the information must be

played and the loading time.

1 INTRODUCTION

Positioning is used to find out the spatial location of

a person in terms of longitude, latitude and altitude.

There are some technologies for positioning, but the

most well-known is GPS (Global Positioning

System) that consists of 24 satellites that are in

medium Earth orbit transmitting signals. GPS

receivers take this information and use triangulation

to calculate the user’s exact location. The European

system, GALILEO, will be operative in 2011-12 and

will provide a higher precision.

Positioning is the key concept to develop

location-based services, which provide information

taking into consideration the position of the users.

There are many possible applications for location-

based services such as health care, navigation,

advertising…

The first generation of located-based services

was not successful, but the apparition of low-cost

GPS receivers and new geographic content changed

dramatically the situation. Nowadays there are

highly successful devices such as GPS navigators.

In the other hand, the radio is the most

widespread communication mean. Another

important feature of radio is that users do not need a

high degree of attention, so they can develop other

activities while they are listening to radio. This

feature allowed the evolution of use cases from

static (a person listening to radio at home) to

dynamic (a driver listening to radio).

The combination of the location-based services

and the radio establishes the basis for a new service

of information and entertainment of general purpose.

A new generation of digital, geo-referenced and

personalized radio that overcomes the limitations of

the current standards of digital broadcasting (DAB),

taking advantage of the features of the bidirectional

transmission of data by packets (UMTS, HSDPA,

WiMAX).

At present there are many works about location-

based services, but few of them deal with the

application of these services to a radio system.

I. Takei et al. proposed a position-adaptive

broadcasting system for mobile terminals. The

approach is similar to the one proposed in this

article, but their solution consists of broadcasting

summaries of local information to all the terminals.

Then every terminal selects which summaries are

going to be extended and downloads the complete

contents from the corresponding server. This

solution is not feasible in this problem because the

contents of the thematic channel are played

asynchronously in every terminal and broadcasting

all the summaries of advertising may overload the

network.

This article presents a new radio system for

mobile terminals. It allows the combination of

thematic contents that are chosen by the user (in a

157

Aumaitre A., Carrillo V. and Javier Ortega F. (2007).

ALGORITHM FOR SELECTION OF CONTENTS IN A LOCATION-BASED SYSTEM.

In Proceedings of the Second International Conference on e-Business, pages 157-161

DOI: 10.5220/0002109301570161

 SciTePress

similar way to a radio channel) and contents with

local information that depends on the location,

velocity and direction of travel of the user. The

second section is an introduction to the system

architecture. The third section concentrates in the

algorithm used by the terminals to select the

contents that are played. The proposed algorithm is

an adaptation of Takei’s algorithm that fulfils the

requirements of the problem exposed in this paper.

The fourth section illustrates the functioning of the

algorithm by means of a simulation.

2 ARCHITECTURE OF THE

SYSTEM

In this section there is a brief resume of the business

model that involves the providers, the customers and

the technological means that allow their

interactuation. To illustrate this, there will be

expounded some habitual use cases that the system

supports.

The profile of the customer consists of a person

or group of persons that carry a terminal that is

playing auditive information. There may be portable

terminals that can be carried by a person, for

example a small radio receiver, but the most usual

profile consists of a person that is travelling by car

(or some vehicle of that kind) and is listening to the

terminal that is installed in its car. The choice of this

profile as predominant is based on the high

percentage of drivers that listen to radio, due to their

capability for developing activities (in this case

driving a car) while are consuming auditive

information. Other means that integrate sound and

images are not so suitable for drivers because can

cause distractions that jeopardize the safety.

The system allows to the customers consuming

two kinds of contents: one of them is a thematic

channel. The information is sent from the server to

all the terminals that select the channel (in a similar

way to a radio channel), but the innovative aspect is

that the content is asynchronous. Every terminal can

play different parts of the content at the same time.

This feature is necessary because the second kind of

content, the advertising, is asynchronous too because

is dependent on the position of the terminal. If the

thematic channel was synchronous, there may be

lost of information produced when the thematic

channel was replaced by the ads. So the terminals

request the data from the server when they need to

play some information of the thematic channel. If

some ad is inserted, the terminal continues playing

the thematic channel at the point it was interrupted.

Related to the kinds of contents there are two

kinds of providers. One of them is the contents

provider, that provides data for the thematic channel.

The other is the ad provider, that provides the geo-

referenced advertising.

There must be some technological resources that

support all the system. First of all there must be a

server that stores the thematic content and the geo-

referenced advertising, must receive the requests

from the terminals and must send the corresponding

data to every terminal. Another component of the

system is the terminal. The main use case consists of

a personal radio that has capability of playing the

data sent from the server and requesting the

necessary data to the server, This terminal can be

configured with the preferences of the customer. For

example, the user may not be interested in listening

information relative to hotels. The secondary case of

use consists of the web interaction. The terminal

may access to internet and apply information

extraction techniques to look for specific

information. The data found is converted to auditive

information with a voice synthesizer. Finally, the

terminal must have the capability to extend the

information about one specific ad. This is necessary

because one geo-reference ad must be a resume of

the information to avoid overloading the network

with information that may not be interesting for the

customer. If the customer was interested in some ad,

by pressing a button he can obtain a more detailed

information.

The communication between the terminal and the

server must be through a wireless wide band

network. Nowadays this technology is supported by

UMTS but there are incoming standards as HDSPA

and WiMAX that improve the speed of transmission

and support more users.

To illustrate the functioning of the system some

possible use cases will be explained.

The first case of use consists of a family that is

travelling by car through Guadarrama Mountains.

They choose the thematic channel “Discover

Madrid” and activate the traffic warnings. While the

presenter talks about the mountains, the local

festivals and monuments, there are inserted some

warnings about the traffic in the area and some ads

about services that can be found through the route.

The second case consists of the owner of a new

restaurant located near an office building. He wants

to attract more clients, so hires the service to insert

an ad during the hour previous to habitual lunch

time. The ad is played in a ratio of 500 metres

ICE-B 2007 - International Conference on e-Business

158

around the restaurant. If a customer is interested,

presses the button “action” of the terminal, and the

information is extended. The owner of the restaurant

pays a fee for the emission of the ad and another fee

for every person that pressed the button “action”.

The third case consists of a local radio station

that broadcasts programmes about the history and

curiosities of Madrid. This programmes are geo-

referenced and integrated in the system explained

above, inserting placements for advertising. The

benefits of the emission of advertising in those

programmes is shared with the local radio station.

This way, the local radio station increases its

benefits reusing contents that already owns.

Figure 1: Architecture of the system.

3 ALGORITHM OVERVIEW

Bellow is a description of the algorithm that

manages the contents that are played by the terminal.

This algorithm is the responsible for switching the

thematic and the local information channels and for

requesting, downloading and playing the data that is

related to the coordinates of the portable terminal.

The map is divided into rectangular pieces called

quadrants. At the beginning, the algorithm loads

into memory the quadrant related to the GPS

coordinates where the terminal is. But from then on,

it is necessary to predict which quadrants will be the

next ones to be crossed by the trajectory of the

terminal to avoid the delay caused by loading into

memory one quadrant when the terminal is already

into it.

Every quadrant is crossed by four lines that are

parallel to its sides. The distance between the sides

of the quadrant and these lines depend on the speed

of the vehicle. The higher the speed of the terminal,

the longer will be the distance. If the trajectory of

the terminal crosses one of this imaginary lines, the

terminal loads into memory all the quadrants that are

adjacent to the sub-quadrant where the terminal

enters. For example, in figure 2, the terminal will

probably enter into the sub-quadrant Q

. If it

happens, the quadrant Q

is loaded into memory. If

the vehicle enters into the sub-quadrant Q

, then the

quadrants Q

, Q

and Q

will be loaded into

memory.

Figure 2: Division of the map into quadrants.

Every quadrant contains the data relative to the

position of the points of interest and their influence

ratios. If the vehicle enters into the area delimited by

one influence ratio, the information related to that

point of interest must be played by the terminal.

Depending on the direction of the vehicle the

algorithm calculates the intersection of the trajectory

of the terminal and the influence areas. If one

influence area is intersected, the loading ratio is

added to the influence ratio. The loading ratio

represents the time necessary for requesting and

loading into memory the information relative to one

point of interest. It is directly proportional to the

speed of the vehicle and depends on the volume of

the information and the loading of the network too.

When the vehicle reaches one loading area, the

terminal begins to download from the server the data

relative to the point of interest. When the vehicle

reaches the influence area, the terminal plays the

information that was downloaded.

Figure 3: Influence and loading areas.

ALGORITHM FOR SELECTION OF CONTENTS IN A LOCATION-BASED SYSTEM

159

If the vehicle changes its trajectory when is

outside of any loading area, the new intersections

must be recalculated and the new loading ratios

reconsidered. If the vehicle is inside a loading area

but outside the influence area and its new trajectory

does not intersect the influence area, the information

is downloaded from the server and stores it in a

buffer, but it does not play it.

There are some problems related to the overlapping

of influence and loading areas. The simplest

solution consists of storing the information in a

playing buffer. This way all the information relative

to the points of interest is played, but if there are

some overlapped areas the information relative to

one point may be played when the terminal is

outside its influence area. Another solution consists

of assigning priorities to the points of interest. The

information pills are stored in the buffer sorted by

priority but when they have to be played, if the

terminal is outside the influence area, the

information is not played and deleted from the

buffer.

4 SIMULATIONS

The algorithm for selection of contents has been

programmed in Java. A graphic interface has been

developed to show the results of the simulations.

The interface draws one quadrant with the position

of the vehicle, its trajectory, the position of the

points of interest and their influence and loading

ratios.

Figure 4: Screensave from simulation.

To illustrate the functioning of the algorithm

some screensaves from one simulation are shown

bellow. The quadrant of the simulation contains

seven points of interest with their corresponding

influence ratios.

The vehicle begins at Avda. Reina Mercedes

street. Its trajectory intersects three influences areas,

so the loading ratio is added to the influence ratio of

the areas intersected, as shown in figure 4.

The vehicle follows the predicted trajectory, so

enters inside the loading area of the first point of

interest. At that moment the data relative to the point

is downloaded from the server and stored in the

playing buffer. In the figure 5, the vehicle is just

entering inside the influence area of that point. The

information stored in the buffer is played then.

Figure 5: Screensave from simulation.

The simulation continues and the vehicle enters

inside the influence area of a second point and the

data relative to that point is played too. But before

entering the influence area of the third point, the

vehicle turns to the left, as shown in figure 6. The

new intersections of the trajectory with the influence

area are calculated, and the corresponding loading

ratios added. Along this trajectory the vehicle

brakes, so the loading ratios decrease.

Figure 6: Screensave from simulation.

ICE-B 2007 - International Conference on e-Business

160

Finally the vehicle continues the predicted

trajectory through the street and crosses the

influence area of two points of interest more, so their

contents are played too.

Figure 7: Screensave from simulation.

5 CONCLUSIONS

The problem of contents selection in location-based

services is well-known and studied in the case of

mobile phones, but there are few works about the

same problem in radio systems. The dissemination

services for advertising or news services in mobile

phones scenes do not have to manage a big amount

of data. In the case of radio systems the quantity of

information that has to be sent is much higher, so the

architecture of the system has to be considerably

modified.

Even the existing works about radio systems

can not be applied to solve the problem exposed in

this article due to the asynchronous nature of the

data that should be played by the terminals. The

solution proposed in this paper for contents selection

is a modification of Takei’s algorithm that takes

into account the requirements exposed before.

This project is currently in progress and there

are many aspects that require further research,

specially the ones related with the architecture of the

system.

ACKNOWLEDGEMENTS

This work has been partially funded by the projects

TIN–2004–07246–C03–03 and OG–136/06 P–

2006/1444.

REFERENCES

Adams, P. M., Ashwell, G. W. B., Baxter, R. (2003)

Location-based services – an overview of the

standards, BT Technology Journal, vol. 21, January,

pp. 34-43.

Dru, M. A., Saada, S. (2001) Location-Based mobile

services: the essentials. Alcatel Telecommunication

Review 1, pp. 71-76.

Ito, M., Matsui, Y., Kondo, T., Shigeno, H., Matushita, Y.

(2001). A broadcasting information providing method

using metadata based on geographical relation. IPSJ

(Information Society of Japan) Journal, vol. 42, n. 7,

July, pp. 1866-1875.

Kannan, P., Chang, A., Whinston, A. (2001) Wireless

commerce: marketing issues and possibilities.

Proceedings of the 34

Annual Hawaii International

Conference on System Sciences (HICSS-34), vol. 9,

January, p. 9015.

Rao, B., Minakakis, L. (2003) Evolution of mobile

location-based services. Communications of the ACM,

vol. 46, December, pp. 61-65.

Schmidt-Belz, B., Zipf, A., Poslad, S., Laamanen, H.

(2003). Location-based mobile tourist services – first

user experiences. Frew, A. J., Hitz, M. & O’Connor,

P. (eds.): Information and Communication

Technologies in Tourism 2003, pp. 115–123. Springer.

Takei, I., Sato, J., Tada, H., Taniguchi, K. (2004) A

position-adaptive broadcasting system for moving

terminals that accommodates different users

requirements. Proceedings of the European Wireless

2004.

Varshney, U. (2003) Location management for mobile

commerce applications in wireless Internet

environment. ACM Transactions on Internet

Technology (TOIT), vol. 3, n. 3, August, pp. 236-255.

Yunos, H. M., Gao, J. Z., Shim, S. (2003) Wireless

advertising’s challenges and opportunities. Computer,

vol. 36, n. 5, May, pp. 30-37.

Zipf, A. (2002) User-adaptive maps for location-based

services (LBS) for tourism. Proceedings of the 9

International Congress on Tourism and

Communication Technologies in Tourism (ENTER

2002), Innsbruck Austria, Springer Computer Science,

Heidelberg Berlin.

ALGORITHM FOR SELECTION OF CONTENTS IN A LOCATION-BASED SYSTEM

161