2015 IIAI 4th International Congress on Advanced Applied Informatics

Application for e-Tourism:

Intelligent Mobile Tourist Guide

Alexander Smirnov1,2, Alexey Kashevnik1,2, Andrew Ponomarev1,2, Maksim Shchekotov1,2, Kirill Kulakov3
1
Laboratory of Computer Aided Integrated Systems
SPIIRAS, St.Petrsburg, Russia
2
International Research Laboratory «Intelligent Technologies for Socio-Cyberphysical Systems»
ITMO University, St.Petrsburg, Russia
3
Department of Computer Science
Petrozavodsk State University (PetrSU), Petrozavodsk, Russia
{smir, alexey, ponomarev, shekotov}@iias.spb.su, kulakov@cs.petrsu.ru

Abstract—This paper presents an e-Tourism application for
supporting tourists in a region. The application recommends the
tourist attractions that are interested to him/her based on the
tourist preferences and the current situation in the region.
Attractions and their descriptions & images are extracted from
accessible Internet sources (like Wikipedia, Wikivoyage,
Panoramio). They are ranged by the special recommendation
service of the application. Recommendations are based on ratings
set by the tourists that use the application. The paper describes
the service-based application architecture, ontology for
intelligent mobile tourist guide services interaction, and
evaluation. Developed application is accessible for download in
Google Play market for Android device users.
Keywords—e-tourism, mobile application, tourist
ontology, service-based architecture
The authors presented the detailed analysis of related work
and classification of mobile tourist guides applications in [2]. It
covers lot of travel-related applications (e.g., [3-10]) as
described in the scientific papers as accessible in Google Play,
Nokia Store, and Apple Store. The analysis shows that mobile
tourist guides application development is a popular research &
development direction, which combines information services
and location-based services to provide the tourist actual and
suitable information before, after, and during the trip.
The rest of the paper is structured as follows. Tourist guide
architecture is presented in Section II. Section III describes
application service interactions in smart space. Section IV
present implementation and evaluation of the presented in the
guide,
paper application. The results are summarized in Conclusion.

I. INTRODUCTION II. APPLICATION ARCHITECTURE

Presented in the paper intelligent mobile tourist guide –
TAIS is a mobile application that proactively recommends the
tourist attractions around, which is currently better to attend.
The recommendation is based on tourist’s preferences and
current situation in the location region. Attractions are
extracted from different Internet sources on the fly (e.g.,
Wikipedia, Wikivoyage, Panoramio). The main idea of the
approach is that TAIS consists of several services, which are
united by the smart space technology for providing the tourist
with information according to personal preferences.
Application architecture is based on Smart-M3 information
sharing platform [1] that implement smart space concept. It
makes possible to significantly simplify further development of
the system, include new information sources and services, and
to make the system highly scalable. The key idea of this
platform is that the formed smart space is device, domain, and
vendor independent. Smart-M3 assumes that devices and
software entities can publish their embedded information for
other devices and software entities through simple, shared
information brokers. Platform is open source and accessible for
download at Sourceforge1.
1
Smart-M3 at Sourceforge, URL: http://sourceforge.net/projects/smart-m3
The intelligent mobile tourist guide consists of the
following main components (Fig. 1):
• client application installed to the user mobile device
that shares tourist context with the smart space and
provides the tourist results of application operation;
• attraction information service that implements
retrieving and caching the information about attractions;
• recommendation service that evaluates attraction/image/
description scores based on ratings that have been saved
to internal database earlier (detailed description of
developed recommendation method is given in [11]);
• region context service that acquires and provides
information about current situation in the considered
region (e.g., weather, traffic jams, closed attractions);
• ridesharing service that finds matching of the tourist
movement with accessible drivers; it provides the
tourist possibilities of comfort transportation to the
preferred attraction (detailed description of developed
algorithm is presented in [12]);
• public transport service that finds information about
public transport to reach the preferred attraction.

978-1-4799-9958-3/15 $31.00 © 2015 IEEE 40

DOI 10.1109/IIAI-AAI.2015.190
 Smart Space-Based Interaction transportation means to reach the preferred attraction, the
ridesharing service finds a driver, which goes from the tourist
Attraction Information
location to the same direction.
Client Application
Tourist Mobile Device

Other Devices in Smart Space

Service Detailed description of the supported TAIS application
scenarios is presented in the Section III and is based on
Recommendation presented in the Fig. 2 ontology. The ontology defines the main
Tourist Context Service concepts and relationships for the TAIS components
interaction. It contains classification of attractions (subclasses
Region Context of the class “Attraction”), different kinds of attraction
Service description (subclasses of class “AttractionDescription”),
classification of different transportation means (subclasses of
the class “Transportation”), user description (class “User”)
Public Transport Service Ridesharing Service that can be a tourist (class “Tourist”), location description
(class “Location”), and route description (class “Route”) that
consists of points (class “Point”), which is related to the class
Fig. 1. General arhitecture of intelligent mobile tourist guide “Location”. Other relationships between presented classes are
depicted in Fig. 2.
Thereby, the main tasks of the client application are share There are following properties have been determined for
information about tourist context, profile, and actions; the ontology.
communication with smart space; provide results to the tourist;
and share tourist ratings of attended attractions, browsed • Property “user_id” for the class “User” that specifies
descriptions and images with the smart space. identifier for the user in the intelligent mobile tourist
guide.
The attraction information service is responsible for four
main tasks. The first one is extraction of attractions titles • Property “role” for the class “User” that specify role
around the tourist from external sources and shares it in the that the user have at the moment.
smart space. The second task is caching acquired information
• Properties “lat” and “long” for the class “Location”
for quick access by tourists in the same location without
that specify latitude and longitude coordinates for the
additional requests to external sources at nearest future (guide
current user location.
application administrator sets live time of cached attraction).
The third task is providing internal identifiers for attractions, • Property “address” for the class “Location”.
which are stored in Geo2Tag platform2 (the platform provides
possibilities for quick search of identifiers by locations). The • Property “temperature” for the class “Location”.
fourth task is extraction of default images for attractions that • Property “weather_icon” for the class “Location”.
are stored in the internal database. Default images are defined
by the recommendation service as the best images for an • Property “weather” for the class “Location”.
attraction based on tourists ratings. The fifth task is extraction
of attraction details (lists of images and text descriptions) from • Property “traffic_jam_level” for the class “Location”.
external sources and sharing them with the smart space. In
addition, the last task is to set and refresh default images in the
internal database for the attractions based on the information is_a is_a
is_a is_a is_a
available in the smart space from the recommendation service. is_a is_a
The recommendation service is responsible for ranking is_a
attractions, images, and descriptions for providing the tourist favorite_transportation
has_route
the best attractions to see and the best images and description has_description
is_a is_a
of chosen attraction for acquaintance. The detailed description is_a
is_a
of this service can be found in [11]. is_a is_a
is_a
The region context service acquires information about is_a is_a
tourist location area from region specific services as soon as favorite_attraction
is_a is_a
the client application shares the tourist context with the smart visited_attraction is_a
space. The context information is shared with the smart space is_a is_a
is_a is_a
and used by other services. has_location

The public transport and ridesharing services provide has_location

alternatives for transportation means for the tourist to reach has_point
preferred attractions. While the public transport service
provides the tourist a route that describes the sequence of Fig. 2. Ontology for intelligent mobile tourist guide services interaction

2
Geo2Tag platform, http://www.geo2tag.org

41
 III. SERVICE-BASED APPLICATION SCENARIOS
The following scenarios have been identified and
developed for the intelligent mobile tourist guide application.
• Providing list of recommended attractions with images
and descriptions to be attended in the specified by the
tourist location. Attraction descriptions and images are
ranged by the recommendation service for presenting
the tourist.
• Providing the information how to reach the preferred
attraction by public transport or with drivers who are
going to the same direction.
• Rating attractions, descriptions, and images after
acquaintance.
A. Providing a List of Recommended Attractions
Fig. 3 shows service interaction sequence for getting a list
of recommended attractions for the tourist. Every service
publishes and subscribes for RDF triples. These triples are
based on the presented in the Section II ontology (Fig. 2).
Subscribe for a triple determines notifications when a required
information (related to the triple) appears in the smart space.
The client application shares tourist context with the smart
space via publication of the following triples when the tourist
searches for attractions around.
([user_id], “tais3:is_a”, “tourist”)
([user_id], “geo:long4”, “60.12”)
([user_id], “geo:lat”, “30.24”)
([user_id], “tais:date_time”, “2014-04-21 10:00”)
([user_id], “tais:role”, “traveler”)
The client application subscribes to the following predicate
([user_id], “recommended_attractions”, None)
and gets notification every time when a list of ordered
attractions around the tourist appears in the smart space. This
information is shown to the tourist.
The context service subscribes to changes in a tourist
context and generates the region-specific context, which is
shared with the smart space.
([user_id], “tais:weather”, “sunny”)
([user_id], “tais:weather_icon”, “http://..”)
([user_id], “tais:temperature”, “20”)
([user_id], “tais:traffic_jams”, “average”).
The attraction information service also subscribes to
changes in a tourist context for searching the list of attractions
in different sources when these changes appears in the smart
space:
(None, “tais:is_a”, None)
(None, “geo:long”, None)
(None, “geo:lat”, None).
3
Hereinafter tais means http://cais.iias.spb.su/XML/tais
4 graphical interface.
Geo:lat and geo:long predicates are defined by “Basic Geo (WGS84 lat/long)
Vocabulary” and must be prefixed with
“http://www.w3.org/2003/01/geo/wgs84_pos#”
42
Each attraction has unique internal identifier kept in the
Geo2Tag platform and can be quickly extracted by attraction
coordinates. Using this identifier the attraction information
service searches in the internal database default image. If it has
been assigned to this attraction previously the attraction
information service use the image as default for the attraction.
In other case, the attraction information service searches for
image in Internet sources and save it as default in the internal
database. The list of attractions with default images is shared
with the smart space,
([user_id], “tais:near_by”, NearbyXML),
where NearbyXML has the following structure.
<?xml version=”1.0” encoding=”UTF-8”?>
<Attractions xmlns=”http://cais.iias.spb.su/XML/tais”>
<Attraction>
<ID>[longitude]+[latitude]</ID>
<Name>The Hermitage</Name>
<Distance>0,2</Distance>
<Longitude>60,125</Longitude>
<Latitude>30,18</Latitude>
<DefaultImage>http://en.wikipe...</DefaultImage>
</Attraction>
<Attraction> … </Attraction>
</Attractions>
The recommendation service subscribes to the list of
attractions via the following triple:
(None, “tais:near_by”, None)
and gets notification every time when the attraction
information services publishes new list of attractions. After
that, the recommendation service queries an information about
the tourist and region contexts via the queries using the
following triples:
([user_id], “tais:date_time”, None)
([user_id], “tais:role”, “traveler”)
([user_id], “tais:weather”, “sunny”)
([user_id], “tais:temperature”, “20”).
The recommendation service evaluates the attraction scores
and publishes the ordered list of attractions based on acquired
context information of tourist and region [11].
([user_id], “recommended_attractions”,
RecommAttractionsXML),
where the RecommAttractionsXML has the following
additional tags for the “Attraction” section of the NearbyXML:
<Score>80</Score>
<Rating>5</Rating>,
where rating is the estimation of the current attraction set by
the tourist last time he/she visited this attraction (can be empty)
and the score is a value calculated value by recommendation
service for attractions ordering.
The client application receives notification for its
subscription and presents the results to the tourists using
 External Images Context
Client App. SS AIS RS Go2Tag
Sources Database Service

Sharing tourist context

information (location,
preferences, ...)
Notification about changes in the tourist context

Context of the tourist location region

Loop: for each external source

Notification about
changes in the tourist
context
Query attraction list nearby the tourist in radius R

List of attractions nearby the tourist

Loop: for each attraction

Query internal identifier for an attraction and insert new one

Attraction identifier
Query default image

Default image

List of attractions with

identifiers and default
images from internal DB

Notification about accessible list of attractions and region context

List of attractions ordered by recommendation service

Notification about
ordered attraction
accessible
Query I default images nearby the tourist

List of attractions ordered Default Images

by recommendation service
Notification about with default images
ordered attractions with
default images accessible

Fig. 3. UML sequence diagram of services interaction for getting list of recommended attractions

<Departure geo:lat="60.97" geo:lon="32.93"/>

B. Providing Information How to Reach a Preferred <Arrival lat="60.97" lon="32.96"/>
Attraction
</Drive>
When a tourist would like to attend an attraction, the <Drive arr_time="10:20" dep_time="10:10"
client application shares the following triples: distance="4.3" type="Bus 56">
([user_id], “transport:detour”, 300) <Departure lat="60.97" lon="32.96"/>
([user_id], “transport:delay”, 5) <Arrival lat="60.69" lon="33.61"/>
([user_id], “transport:point”, [used_id]“dest”) </Drive>
([user_id]“dest”, “geo:long”, “60.52”). </Transportation>
([user_id]“dest”, “geo:lat”, “30.32”). The ridesharing service publishes the following triple,
Public transport and ridesharing services subscribe to the when it completes the calculations:
following triple: ([user_id], “tais:recomm_ridesharing”, RidesharingXML)
(None, “transport:point”, None) where PublicTransportXML contains list of points of driver
and get notifications every time when the tourist adds a new route with information about seats accessible:
destination point. After the route calculation, the public <?xml version=”1.0” encoding=”UTF-8”?>
transport service shares the following triple: <Transportstion xmlns=”http://cais.iias.spb.su/XML/tais”>
([user_id], “tais:recomm_transport”, <Point geo:lat="59.97" geo:lon="30.32" seats: 4/>
PublicTransportXML), <Point geo:lat="59.94" geo:lon="30.45" seats: 1/>
…
where PublicTransportXML contains information how to
</Transportation>
reach the attraction by public transport:
where object is the ridesharing path that defines a list of
<?xml version=”1.0” encoding=”UTF-8”?> points for the tourist and driver movements. Long and lat are
<Transportstion xmlns=”http://cais.iias.spb.su/XML/tais”> longitude and latitude of every point, alone determines if the
<Drive arr_time="10:10" dep_time="10:00" tourist goes from this point alone or with the driver, mp
distance="3.04" type=”Bus 562”> means that this point is a meeting point.

43
C. Rating Attractions, Descriptions, and Images screenshot), and/or estimate it. Detailed information about
After attending an attraction, the tourist has a chance to attraction contains a list of images that is associated with this
estimate it. In this case, the client application publishes the attraction and it description. This information is extracted by
following triple: the attraction information service from different internet
sources (e.g., Wikipedia, Wikivoyage, and Panoramio are
([attraction_id], “tais:rate_attraction”, used at the moment).
RateAttractionXML)
The tourist has possibility to estimate images using the
where RateAttractionXML contains rating for attraction, user, following options: “like image”, “dislike image”, “this image
and region contexts. is not applicable” to the attraction (see Fig. 5, left
<?xml version=”1.0” encoding=”UTF-8”?> screenshot). Based on these estimations the recommendation
<Evaluation xmlns=”http://cais.iias.spb.su/XML/tais”> service will re-order images for this or another tourist next
<UserID>[user_id]</UserID> time. The tourist can browse the attraction reaching path by
choosing “Show on the map” item in context menu (see
<Context>
Fig. 5, right screenshot). The ridesharing and public transport
<DateTime>2013-10-09 10:59</DateTime> services are responsible for calculating attractions reaching
<Weather>Sunny</Weather> path based on developed OpenStreetMap-based web
<Company>Alone</Company> mapping service [13]. It provides the tourist possibility to
</Context> shows paths in the interactive map, find fellow travelers who
<Rating>4</Rating> go to the same direction, and find public transport to reach
</Evaluation> chosen attraction.
The tourist also has possibility to estimate images related Application Evaluation shows that it response time not
to an attraction. In this case the client application publishes more that few seconds for every operation. For example, for
the following triple: the center of St.Petersbug (Vasilyevsky Island area)
([image_url],“tais:rate_image”,RateImageXML), recommendation of attractions use case takes about 3
where RateImageXML contains rating of the image: seconds for the acquiring (2,9 sec), ranking (0,1 sec) and
<?xml version=”1.0” encoding=”UTF-8”?> providing to the tourist up to 50 nearest attractions.
<Evaluation xmlns=”http://cais.iias.spb.su/XML/tais”>
Location Location

<UserId>[user_id]</UserId>

Recommende Clickabl List of images

<Rating>4</Rating>
</Evaluation>
on the map

For estimating an attraction description the following

triple is used:

e Map
([description_url],“tais:rate_description”,
Context

RateDescriptionXML) d attractions
where RateDescriptionXML has the same structure as
RateImageXML.

IV. EVALUATION
Implementation of the application has been developed Fig. 4. Intelligent mobile tourist guide screenshots: main screen, context
based on Smart-M3 information sharing platform. For menu with actions
accessing the platform Java KPI library 5 is used. Mobile
clients have been implemented using Android Java
Destination attraction

Development Kit6.
List of images

The main application screen is shown in Fig. 4, left

screenshot. The tourist can see images extracted from
Estimation

accessible Internet sources around, clickable map with

his/her location, context situation (weather), and the best
three attractions provided by the recommendation service.
When the tourist click to an attraction the following context
Path
Attraction

menu is opened (see Fig. 4, right screenshot). The tourist can

see detailed information about the chosen attraction (Fig. 5,
Location

left screenshot), browse attraction reaching path that is

proposed by the system route to an attraction (Fig. 5, right

5
Java KPI library, http://sourceforge.net/projects/smartm3-javakpi/ Fig. 5. Intelligent mobile tourist guide screenshots: attraction details and
6
Android Java Development Kit, http://developer.android.com/sdk/ route to the attraction

44
 The dependency of query transaction execution time
based on increasing of number of tourists is presented in
Fig. 6. For the experiments, a test application has been
developed that generates triples that describe tourists in
smart space (one tourist is described by approximately 30
triples) and calculates query & insert transaction execution
time. Insert transaction execution time does not depend on
count of triples in smart space while a query transaction has a
linear dependency (see Fig. 6). One thousand tourists is
described by approximately twenty-seven thousands of
triples and one thousand of subscribe transactions. For this
count of triples and subscribe transactions response time of
Smart-M3 platform is approximately 0.3 seconds. The
average size of XML files that is used to describe list of
attractions nearby is 5 Kb for 25 attractions. For the
experiments Smart-M3 platform core was located in the
following computer (see Table I).
TABLE I. HARDWARE CHARACTERISTICS FOR EXPERIMENTS
Characteristic name Characteristic value
Host operation system Windows Server 2008
Virtual operation system Debian 7.6 64 bit
Hardware Intel Xeon E5620 2.4 GHz, RAM 1.4 Gb
Network Ethernet, 1000 Mbit/s
V. CONCLUSION
The paper presents intelligent mobile tourist guide
application for providing the tourist with attraction around
based on his/her preferences and current situation in the
region. Proposed service-based architecture consists of
several services that interact with each other using the
ontology for semantic interoperability support. The smart
space technology is used to provide information sharing
between different application modules. Developed
application is based on open source Smart-M3 platform that
implement smart space concept. Evaluation shows that
application finds and ranges the attractions around by the
feasible time and can be used for the on-the-fly tourist
support during the trip. Application is accessible for
download in Google Play market for Android device users7.
ACKNOWLEDGMENT
The presented results are part of the research carried out
within the project funded by grants # 13-07-00336, 13-07-
00271, 13-07-12095 of the Russian Foundation for Basic
Research. This work was partially financially supported by
Government of Russian Federation: grant 074-U01, by the
Ministry of Education and Science of the Russian
Federation: project # 14.574.21.0060 (RFMEFI57414X0060)
of Federal Target Program "Research and development on
priority directions of scientific-technological complex of
Russia for 2014–2020", and by the project # 2.2336.2014/K
from the project part of state research assignment.
7 FRUCT, Saint-Petersburg, Russia, IEEE, 2014, pp. 135-143.
Intelligent mobile tourist guide – TAIS,
https://play.google.com/store/apps/details?id=ru.nw.spiiras.tais
45
Fig. 6. Query transaction execution time dependency on number of tourists
in intelligent mobile tourist guide
REFERENCES
[1] J. Honkola, H. Laine, R. Brown, O. Tyrkko, “Smart-M3 Information
Sharing Platform,” In proc. IEEE Symp. Computers and
Communications, IEEE Comp. Soc., 2010, pp. 1041-1046.
[2] A. Smirnov, A. Kashevnik, N. Shilov, N. Teslya, A. Shabaev,
“Mobile Application for Guiding Tourist Activities: Tourist Assistant
– TAIS,” In proc. of the 16th Conference of Open Innovations
Association FRUCT, Oulu, Finland, IEEE, 2014, pp. 94-100.
[3] R. Anacleto, L. Figueiredo, F. Almeida, P. Novais, “Mobile
application to provide personalized sightseeing tours,” Network and
Computer Applications, vol. 41, 2014, pp. 56-64.
[4] J. Borras, A. Moreno, A. Valls, “Intelligent Tourism Recommender
Systems: A Survey,” Expert Systems with Applications, vol. 41, no.
16, 2014, pp. 7370-7389.
[5] C. Emmanouilidis, R. Koutsiamanis, A. Tasidou, “Mobile Guides:
Taxonomy of Architectures, Context Awareness, Technologies and
Applications,” Network and Comp. App., vol. 36, 2013, pp. 103-125.
[6] D. Gavalas, C. Konstantopoulos, K. Mastakas, G. Pantziou, “Mobile
Recommender Systems in Tourism,” Network and Computer
Applications, vol. 39, 2014, pp. 319-333.
[7] K. Al-Rayes, A. Sevkli, H. Al-Moaiqel, H. Al-Ajlan, K. Al-Salem, N.
Al-Fantoukh, “A Mobile Tourist Guide for Trip Planning,” IEEE
Multidisciplinary Eng. Education Magazine, vol. 6(4), 2011, pp. 1-6.
[8] O. Garcia, R. Alonso, F. Guevara, D. Sancho, M. Sánchez, and J.
Bajo, “ARTIZT: Applying Ambient Intelligence to a Museum Guide
Scenario,” Ambient Intelligence - Software and Applications,
Springer-Verlag Berlin Heidelberg, 2011, pp. 173–180.
[9] K. Kulakov and A. Shabaev, “An approach for creation smart space-
based trip planning service,” in Proc. 16th Conf. of Open Innovations
Association FRUCT, Oct. 2014, pp. 38–44.
[10] A. Varfolomeyev, D. Korzun, A. Ivanovs, and O. Petrina, “Smart
personal assistant for historical tourism,” in Proc. 2nd Int'l Conf. on
Environment, Energy, Ecosystems and Development, Athens, Greece,
2014, pp.9-15.
[11] A. Smirnov, A. Kashevnik, A. Ponomarev, N. Teslya, M.
Shchekotov, S. Balandin, “Smart Space-Based Tourist
Recommendation System: Application For Mobile Devices,” Internet
of Things, Smart Spaces, and Next Generation Networks and
Systems, LNCS 8638, Springer International Publishing Switzerland,
2014, pp. 40-51.
[12] A. Smirnov, N. Shilov, A. Kashevnik, N. Teslya, S. Laizane, “Smart
Space-based Ridesharing Service in e-Tourism Application for
Karelia Region Accessibility. Ontology-based Approach and
Implementation,” In proc. 8th Int. Joint Conference on Software
Technologies, July 29-31, Reykjavik, Iceland, 2013, pp. 591-598.
[13] N. Teslya, “Web Mapping Service for Mobile Tourist Guide,” In
proc. of the 15th Conference of Open Innovations Association