2022 International Conference on Advanced computing Technologies & Applications (ICACTA), Mar.

04 – 05, 2022, Coimbatore, INDIA

Design and Implementation of Virtual Tour Guide

App
2022 International Conference on Advanced Computing Technologies and Applications (ICACTA) | 978-1-6654-9515-8/22/$31.00 ©2022 IEEE | DOI: 10.1109/ICACTA54488.2022.9752804

Baivab Maulik Aditi P Nayak

Department of Computer Science Department of Computer Science
PES University PES University
Bangalore, India Bangalore, India
[email protected] [email protected]

Sanjana U Simmi Alok

Department of Computer Science Department of Computer Science
PES University PES University
Bangalore, India Bangalore, India
[email protected] [email protected]

Divyaprabha K N
Department of Computer Science
PES University
Bangalore, India
[email protected]

Abstract—This paper is about a flutter-based application The tourism sector is seen to be thriving more and more
which integrates different functionalities into a single app hence since the growth of smartphones and mobile applications. The
helping the tourists get important information and functionalities tourists find it easier and convenient to have access to
at the same place without the need to explore multiple information such as tourist destinations, lodging, dining
applications. This application offers 3 main services. The first
options, booking and many more simply through a mobile
service does image recognition with the help of google cloud
vision API and provides information about the recognized tourist application which is just a touch away. In recent years
attraction using web scraping techniques. The second extensive progress has been seen in travel-based apps. These
functionality of the app is planning a trip which provides apps are acting as a replacement to the traditional sources of
content- based recommendation and optimal route generation by information to the tourists such as guidebooks, brochures and
solving the travelling salesman problem (TSP) using genetic travel agents.
algorithm. Both these services are implemented as Rest API in
Django. The third major feature is finding a tour guide virtually This paper will discuss one such app which mainly
and obtaining information from them directly. Thus, also comprises 3 major functionalities being image recognition and
creating job opportunities for many tour guides. gaining knowledge about tourist attractions, planning a
personalized tour based on the interests of the user and
Keywords—flutter; image recognition; web scraping; Django; obtaining an optimal route which has been implemented using
TSP; genetic algorithm the genetic algorithm and finding a tour guide virtually and
obtaining information from them directly. Thus, also creating
I. INTRODUCTION job opportunities for many tour guides.
The rapid advancement in smartphones has had a major
influence on the lives of people and has brought in significant II. LITERATURE SURVEY
changes in the modern society. This led to a notable Tourism and travel are two areas where users may profit
enhancement in the development of mobile applications. This from smart phone applications. A growing trend has been
has expanded the capabilities of mobile phones in addition to observed in the last decade: the traveler is often connected to mobile
conventional audio communication. Tourism industry is one of devices [1]. Consequently, in the domain of tour guide and trip
the major sectors which has benefited from the advancements planning many researchers and developers have come up with
of mobile applications. research papers and applications that help tourists. The

978-1-6654-9515-8/$31.00©2022IEEE
Authorized licensed use limited to: T.C. Cumhurbaskanligi Kutuphanesi. Downloaded on November 08,2024 at 11:53:28 UTC from IEEE Xplore. Restrictions apply.
 2022 International Conference on Advanced computing Technologies & Applications (ICACTA), Mar. 04 – 05, 2022, Coimbatore, INDIA

function of the guide is being taken over by other services explains how to solve the route planning problem using a
and applications as technology advances.[2]. greedy algorithm. The results of experiment on both synthetic
and actual data sets demonstrates that the algorithm is capable
In paper [2] , the author proposes an android mobile
of find an optimal path. [10] presents an offline method for
application that enables the users to recognize the tourists
producing an ideal vacation route plan in order to maximise
attractions from their mobile phone with just an image of
the number of destinations visited throughout the vacation
their desired tourist attractions and a mini guide on the
period. The suggested method would compute the arrival time
recognized tourist attraction. It uses image recognition tools
and minimum stay time to see if it exceeds the stop's working
like OpenCV and API like google maps, google places and
hours.
GPS technology. The user would have to select an image,
which would then be compared to the database's images, and In the paper [11] is a combination of features available in
the monument would be identified. In [3] study has been other travel plan apps like checking hotel availability, booking
conducted on building an android app for tourism in hotel rooms, find the tourist attractions in the places that they
Indonesia. The app had features like scanning a landmark want to visit, calculating optimal routes between the different
and this was done with help of google cloud vision API. user specified venues/destinations etc., while having the
Another main feature of this app was to provide information constraints of time, locations, meeting days and the users’
on all ongoing/upcoming events in Indonesia. The app was personal preferences dining, hotel and also offering
evaluated and 94% of users found it really helpful. suggestions to the users based on their preferences. The
algorithm that has been used to find the best path is modified
Recommender System has become a hot topic and aims
by Dijkstra's algorithm. Route planning can also be a
to help users find items that match their interest [4].
Travelling salesman problem When starting and end points are
Reference [5] discusses one such mobile application named
the same.
TAIS. TAIS recommends tourist attractions according to
user’s preferences and provides details of the attractions. The research [12] uses genetic algorithms' optimization
The information about attractions come from multiple capabilities to create a viable TSP solution. The procedure
internet sources like websites / blogs that provide begins by calculating the Euclidean distance between the
multimedia information and comprehensive text-based places that the salesman will visit. Value encoding is used to
information about different places of interest. TAIS not only create the first chromosomal pool. The best-fitting
recommends places but also provides routing service chromosomes are then chosen using a roulette wheel selection
through OpenStreetMap web API. Tourists can use the method, followed by m-point crossover. Apply exchange
routing service to create a footpath, discover fellow mutation to the previously created offspring. This technique is
passengers going in the same direction, and find public now continued until the genetic algorithm converges.
transportation using Yandex. Schedule API to get to their
desired location. III. TECHNOLOGIES USED
A significant type of context aware apps is location
dependent services, which respond to location-related A. Flutter
inquiries. [6]. In [6] the author discusses implementation of Flutter is a free and open-source user interface software
Tour- guide which is a location based mobile application. The development kit from Google. It's used to develop cross-
application has a central server and mobile clients, which platform apps for Android, iOS, Linux, Mac, Windows,
follow the client server architecture. The mobile phones are Google Fuchsia, and the web with a single codebase. It was
aware of their location and can connect with the server employed in the development of our tour guide app.
through a wireless network.. Users can get details of the
attraction in terms of text, picture and video. With the help of B. Django
GPS the application can give location based information. In
addition, the article highlights several techniques taken to Django is a Python-based web framework that follows the
increase performance, conserve battery power, and decrease model–template–views design pattern. The recommendation
communication overhead. Data caching methods are used in API and the best route generating API were constructed using
order to reduce wireless network traffic cost by processing this framework, and both were written in Python and could be
user’s query locally. hosted anywhere.
Due to various customised criteria, tourism path design is a C. GPS
difficult challenge[7]. Research has been conducted on
algorithms that find the shortest path. Paper [8] proposes an The Global Positioning System (GPS) is a navigation
artificial intelligence based mobile trip planner. A star and ant system for air, sea, and land transport that uses satellites, a
receiver, and algorithms to synchronise location, velocity, and
colony optimization algorithms are used for route planning.
time data. This is used to determine the user's present location
Provided places of interest by the user, the app will draw an
in order to provide them with more relevant recommendations.
optimized path through the route along time constraints. It
maintains track of prior journeys and activities taken by users.
As a result, it is capable of making fair recommendations. [9]

978-1-6654-9515-8/$31.00©2022IEEE
Authorized licensed use limited to: T.C. Cumhurbaskanligi Kutuphanesi. Downloaded on November 08,2024 at 11:53:28 UTC from IEEE Xplore. Restrictions apply.
 2022 International Conference on Advanced computing Technologies & Applications (ICACTA), Mar. 04 – 05, 2022, Coimbatore, INDIA

D. Google Maps API: C. Fitness Function

The Google Maps Platform is a collection of APIs and A fitness function is a function that takes a potential
SDKs that allow developers to integrate Google Maps into solution to a problem as input and outputs how "fit" or
mobile apps and websites, as well as obtain data from Google
"excellent" the answer is in relation to the situation at hand.
Maps. This API is utilised in this application to provide the
user's current location as well as the best path for them to take. In this application the fitness function to be 1/(total distance
of the route) or 1/(total time taken to cover the route) i.e.
E. Google Cloud Vision: (inverse of total distance or total time) since the objective is
This app is making use of cloud vision API for landmark to identify a route with the shortest overall distance or time.
detection. This API accepts an image as input. By making The route having the least total distance or total time will
use of image recognition techniques, the API detects the have the highest fitness value and hence will have a high
landmark of which the image was taken and returns the probability of being the optimal route.
relevant information like latitude, longitude etc.
D. Travelling Salesman problem using genetic algorithm
IV. FORMULAS AND ALGORITHMS

A. Distance calculation using the Haversine formula tsp_using_genetic_algorithm(population,population_size,

To calculate the distance between two places on Earth, mutation_rate, generations, elite_size)
the Haversine formula is utilised. The distance between the {
user's current location and tourist sites is calculated using population = initial population is generated randomly
this algorithm.
of size 'population_size'
Value of latitude in radians = latitude / (180/pi) for i in range(generations):
Value of longitude in radians = longitude/(180/pi) fitness_value = []
list = []
𝑑 selected_chromosomes = []
𝛼 −𝛼
= 2𝑟𝑠𝑖𝑛 𝑠𝑖𝑛 for route in population:
2
fitness_value.append(fitness(route))
+ cos(𝛼 ) cos(𝛼 ) 𝑠𝑖𝑛 (𝛽 − 𝛽 )/2 list = ranked list of routes based on fitness
selected_routes = select top 'elite_size' number of
Where,
routes from list
α1 = latititude of point one (in radians) for i in range(len(list) - elite_size):
α2 = latitude of point two (in radians) selected_routes.append(roulette_wheel_sele
ction(list))
β1 = longitude of point one (in radians)
routes = actual list of all the routes selected
β2 = longitude of point two (in radians)
set1 = first half of routes
r = radius of earth which is 6371km
set2 = second half of routes
B. Cosine similarity ans = crossover routes in set1 and set2
The similarity of two vectors in an inner product space for each route in ans:
is measured by cosine similarity. It detects if two vectors
are pointing in the same direction by measuring the cosine population = mutate(route, mutation_rate)
of the angle between them. final = ranked list of routes in population based on fitness
𝑥. 𝑦 optimal_route = topmost route in final
sin(𝑥, 𝑦) =
∥𝑥 ∥∥𝑦 ∥
return optimal_route
}

978-1-6654-9515-8/$31.00©2022IEEE
Authorized licensed use limited to: T.C. Cumhurbaskanligi Kutuphanesi. Downloaded on November 08,2024 at 11:53:28 UTC from IEEE Xplore. Restrictions apply.
 2022 International Conference on Advanced computing Technologies & Applications (ICACTA), Mar. 04 – 05, 2022, Coimbatore, INDIA

fitness(route)
{
return 1/total_distance(or total_time)
}

roulette_wheel_selection(list)
{
cum_sum = cumulative sum of fitness values
cum_perc = cumulative percentage of fitness values
a = generate a random number between 0 and 100
r = select the route having a cumulative percentage
just above
Fig. 1. Click a Picture
the random number generated
return r
}

V. IMPLEMENTATION
This section will discuss working of each feature of the
flutter based application, Smaarak- My tour guide.
Smaarak-My tour guide has been developed using flutter
and runs in both android and IOS operating systems. Entire
app contains around 8000 lines of code written in dart
language. The app requires access to the internet in order to
provide services. It also requires GPS access for better
recommendations. The server contains micro service APIs that
were implemented in python using the Django framework.
These APIs provide services like recommendation, optimal
route to the user app. Google’s cloud storage service firebase
has been used for databases.

A. Click a Picture
The user can click a picture of the tourist attractions or
upload a picture from the gallery. The landmark detection
API will recognize the place and then send the landmark to
Wikipedia API. The Wikipedia API returns the information Fig. 2. Recommendation
in pure text and which when fed to Google’s TTS engine for
audio translation. B. Plan a tour
The user has to select a city and enter the radius of search.
Radius of search is the distance in km within which
attractions will be recommended. The user has to select
different types of venues he would like to visit. Once submit
button is pressed, an http post request with body containing
city name, radius of search, geo coordinates, and list of
venue preferences will be sent to an API built using Django
and hosted. The API upon getting http post request will first
calculate the distance between every attraction in the
selected city and user’s current location using haversine

978-1-6654-9515-8/$31.00©2022IEEE
Authorized licensed use limited to: T.C. Cumhurbaskanligi Kutuphanesi. Downloaded on November 08,2024 at 11:53:28 UTC from IEEE Xplore. Restrictions apply.
 2022 International Conference on Advanced computing Technologies & Applications (ICACTA), Mar. 04 – 05, 2022, Coimbatore, INDIA

distance formula and then take cosine similarities and send a

response back to the app. As indicated in the picture, the
answer will include recommendations depending on the
user's venue choices. Fig 5.4 Find a Guide

VI. CONCLUSION AND FUTURE WORK

Smaarak-My Tour guide is a reliable and effective
application that a traveller can use. The app makes travelling
easy and gives travellers the freedom to explore a place
however they need. This paper mainly proposes a practical
solution to problems in the tour guide domain.
This paper discussed the working of different services
offered by Smaarak. Smarak recognizes a tourist attraction
with just an image and gives an audio tour of the recognized
image. It recommends the best places according to preferences
and provides an optimal route. It also provides a service where
Fig. 3. Route on Map
a traveller can chat with local guides of the city and get their
guidance.
The user can then select any number of attractions from For future work, enhancement can be made to existing
the recommended attractions and opt for either optimal route features of this application. For example, a track of travel
calculation based on distance or based on time. On clicking history and use it for improving the recommendations. Live
calculate route by distance, a http request will be made to tracking of the route generated. Keeping in mind the poor
the API with name and coordinates of selected venues network connectivity in parts of the country, build a version of
the application that doesn’t depend as much on the internet for
as body. The optimal route will be calculated by genetic
use.
algorithm and the order of places will be sent as response to
the app. Using this optimal order, the route will be displayed Smaarak – My tour Guide app can make travelling a
on google maps. wonderful experience as tourists will not have to deal with
hiring a guide with no standard rates to compare with. Instead
they can focus more on the touring with resources and verified
C. Find a guide information on their fingertips.
This feature mainly focuses on interaction between the tourist and tour guide.
It is a chat feature which makes use of firebase Firestore database for saving
live user information and chats this feature is also integrated with payments
page where the tourist can pay the tour guide based on the service they have ACKNOWLEDGMENT
provided. This feature not only allows the users to choose and connect with
guides after comparing their rates to the standards set by others but also opens This project was supported by Dr. Sarasvathi V, Associate
up a bigger opportunity and market of clients for the tour guides. Professor, the Capstone Project Coordinator, who organised,
managed, and assisted throughout the process.
Special thanks to Dr. Sandesh B J, Chairperson, PES
University's Department of Computer Science and Engineering
and Dr. B.K. Keshavan, Dean of Faculty of PES University,
for his assistance and information and assistance.
Finally, Dr. M. R. Doreswamy, PES University's Chancellor,
Prof. Jawahar Doreswamy, PES University's Pro Chancellor,
and Dr. Suryaprasad J, PES University's Vice-Chancellor, for
providing with numerous chances and enlightenment along the
path.

978-1-6654-9515-8/$31.00©2022IEEE
Authorized licensed use limited to: T.C. Cumhurbaskanligi Kutuphanesi. Downloaded on November 08,2024 at 11:53:28 UTC from IEEE Xplore. Restrictions apply.
 2022 International Conference on Advanced computing Technologies & Applications (ICACTA), Mar. 04 – 05, 2022, Coimbatore, INDIA

REFERENCES Research on Engineering, Business, and Healthcare Applications of Data

Science and Analytics, pages 98-114.
V. Arulkumar An Intelligent Face Detection by Corner Detection using
[1] de Farias, I., Júnior, N. L., & Teixeira, M. M. (2017, June). Urbis: A Special Morphological Masking System and Fast Algorithm, 2021 2nd
touristic virtual guide. In 2017 12th Iberian Conference on Information International Conference on Smart Electronics and Communication
Systems and Technologies (CISTI) (pp. 1-6). IEEE. (ICOSEC) pp 1556-1561, Nov 2021.
[2] Fatima, R., Zarrin, I., Qadeer, M. A., & Umar, M. S. (2016, July). Mobile
travel guide using image recognition and GPS/Geo tagging: A smart
way to travel. In 2016 Thirteenth International Conference on Wireless
and Optical Communications Networks (WOCN) (pp. 1-5). IEEE.
[3] Irmanti, D., Hidayat, M. R., Amalina, N. V., & Suryani, D. (2017).
Mobile smart travelling application for indonesia tourism. Procedia
computer science, 116, 556-563.
[4] Singh, P. K., Pramanik, P. K. D., Dey, A. K., & Choudhury, P. (2021).
Recommender systems: an overview, research trends, and future
directions. International Journal of Business and Systems
Research, 15(1), 14-52.
[5] Smirnov, A., Kashevnik, A., Shilov, N., Teslya, N., & Shabaev, A. (2014,
October). Mobile application for guiding tourist activities: tourist
assistant-tais. In Proceedings of 16th Conference of Open Innovations
Association FRUCT (pp. 95- 100). IEEE.
[6] Shi, X., Sun, T., Shen, Y., Li, K., & Qu,W. (2010, June). Tour-guide:
Providing location-based tourist information on mobile phones. In 2010
10th IEEE International Conference on Computer and Information
Technology (pp. 2397-2401). IEEE.
[7] Xu, Y., Hu, T., & Li, Y. (2016, August). A travel route recommendation
algorithm with personal preference. In 2016 12th international
conference on natural computation, fuzzy systems and knowledge
discovery (icnc-fskd) (pp. 390-396). IEEE.
[8] Aydin, A., & Telceken, S. (2015, June). Artificial intelligence aided
recommendation based mobile trip planner for Eskisehir city. In 2015
IEEE 10th Conference on Industrial Electronics and Applications
(ICIEA) (pp. 1650-1654). IEEE.
[9] Bao, J., Yang, X., Wang, B., & Wang, J. (2013, November). An efficient
trip planning algorithm under constraints. In 2013 10th Web
Information System and Application Conference (pp. 429-434). IEEE.
[10] Chia, W. C., Yeong, L. S., Lee, F. J. X., & Ch'ng, S. I. (2016, August).
Trip planning route optimization with operating hour and duration of
stay constraints. In 2016 11th International Conference on Computer
Science & Education (ICCSE) (pp. 395-400). IEEE.
[11] Jafri, R., Alkhunji, A. S., Alhader, G. K., Alrabeiah, H. R., Alhammad,
N. A., & Alzahrani, S. K. (2013, December). Smart Travel Planner: A
mashup of travel-related web services. In 2013 International Conference
on Current Trends in Information Technology (CTIT) (pp. 181- 185).
IEEE.
[12] Juneja, S. S., Saraswat, P., Singh, K., Sharma, J., Majumdar, R.,
&Chowdhary, S. (2019, February). Travelling salesman problem
optimization using genetic algorithm. In 2019 Amity International
Conference on Artificial Intelligence (AICAI) (pp.264-268)IEEE.
[13] R. Lokeshkumar, E. Maruthavani, and A. Bharathi , A new perspective
for decision makers to improve efficiency in social business intelligence
systems for sustainable development, International Journal of
Environment and Sustainable Development 2018 17:4, 404-416.
[14] S. Balakrishnan, V. K. and M. S. S. Hameed, "An Embarking User
Friendly Palmprint Biometric Recognition System with Topnotch
Security," 2021 5th International Conference on Intelligent Computing
and Control Systems (ICICCS), 2021, pp. 1028-1032, doi:
10.1109/ICICCS51141.2021.9432230.
[15] V. Arulkumar A Survey on Multimedia Analytics in Security Systems of
Cyber Physical Systems and IoT, 2021 2nd International Conference on
Smart Electronics and Communication (ICOSEC) pp 1556-1561, Nov
2021. 10.1109/ICOSEC51865.2021.9591754
Ravi Kumar Poluru & Lokesh Kumar R (2021) An Improved Fruit Fly
Optimization (IFFOA) based Cluster Head Selection Algorithm for
Internet of Things, International Journal of Computers and Applications,
43:7, 623-631, DOI: 10.1080/1206212X.2019.1600831
Selvan.C, S.R.Balasundaram, (2020), Data Analysis in Context Based
Statistical Modeling in Predictive Analytics, IGI Global, Handbook of

978-1-6654-9515-8/$31.00©2022IEEE
Authorized licensed use limited to: T.C. Cumhurbaskanligi Kutuphanesi. Downloaded on November 08,2024 at 11:53:28 UTC from IEEE Xplore. Restrictions apply.