DOKK Library
                                                                                     Polish Cartographical Review
                                                                                         Vol. 54, 2022, pp. 108–122
                                                                                       DOI: 10.2478/pcr-2022-0008
                                                                                               Received: 28.08.2022
JAKUB ŁOBODECKI                                                                                Accepted: 13.10.2022
Warsaw University of Technology
Faculty of Geodesy and Cartography
Warsaw, Poland;
Warsaw University of Technology
Faculty of Geodesy and Cartography
Warsaw, Poland;

  Developing a simulator of a mobile indoor navigation application
                as a tool for cartographic research
   Abstract. Solutions designed for indoor navigation are extremely rare compared to outdoor navigation;
however, the potential for development is, therefore, very high. Several pilot projects exist in airports, universities,
hospitals, and shopping centres. The difficulties in development are currently mainly due to the continuing low
quality of indoor positioning and lack of widespread access to high-quality building models. A strong methodo-
logical basis for how the interior and exterior of buildings can be cartographically represented in navigation
applications has also not yet been developed. Therefore, an attempt was made to design a virtual environ-
ment dedicated to supporting the design of indoor navigation applications. Authors present the results of
a study aimed at creating a concept of a simulation environment accompanied by the assessment and prelimi-
nary validation of its technological feasibility in terms of the method and technology used – although it does
not yet constitute a target study. The result was a fully functional prototype of a virtual test environment, which
was successfully used in a pilot study on the effectiveness of different types of navigation guidance. The par-
ticipants’ behaviour within the desktop virtual environment was investigated and their opinions were collected
through a questionnaire. This research proved the technological feasibility of the proposed concept and
demonstrated the usefulness of the Unreal Engine game engine in building new tools to support the work of
cartographers. The created environment will be further developed and used in indoor mapping research.
   Keywords: indoor cartography, indoor navigation, map design, Unreal Engine, virtual reality

1. Introduction                                                   for example, positioning information using an
                                                                  already existing Wi-Fi network in the building,
   The development of indoor navigation appli-                    Bluetooth beacons, or mobile phones’ inertial
cations has been slowed and limited by particu-                   sensors (Sakpere et al., 2017).
lar practical and technological difficulties which                   An excellent example of the utilisation of the
have prevented the direct transfer of solutions                   above technology is a US hospital, which, in
known from other types of navigation. The main                    2015, became the first medical facility in the US
difficulty faced by the designers of indoor navi-                 to implement a navigation system in its building
gation systems is the lack of widely available                    (Health IT Outcomes, 2015). An example of
technology for obtaining precise location infor-                  a development in indoor mapping is the InMapz
mation within buildings. When a user is indoors,                  application, which is widely available and con-
the signal from the GNSS satellite network is                     stantly expanding the number of buildings it
limited or completely unavailable. In such situ-                  covers (InMapz, 2022); it allows static floor
ations, the only option is to use one of the other                plans to be converted into digital twins through
technologies available today (Marciniak, 2018);                   an automated process.
© 2022 Authors. This is an open access article licensed under the Creative Commons Attribution 4.0 License.
                 Developing a simulator of a mobile indoor navigation application as a tool...          109

   In addition to location, another essential piece      on streets and roads for cars, or even on pave-
of information used by navigation applications           ments for pedestrians. Movement is much less
is the direction in which the person using the           orderly and also less predictable than for cars
navigation software is “looking” at any given time.      on roads; completely different factors determine
Most smartphones are equipped with a com-                whether people stop or turn around; movement
pass, which usually provides this information.           can also be vertical (e.g., using a lift). Moving
However, this is sometimes problematic due to            around a building is subject to different restric-
uncalibrated sensors or magnetic field inter­            tions to those for cars on the road (e.g., opening
ference caused by, for example, a nearby lift,           hours, staff-only sections). There are also sig­
electrical appliances, or other metal objects.           nificant differences in the environment: starting
   Another vital element that creates a signifi-         with the lighting, the way the building is sign-
cant difficulty in the implementation of indoor          posted, and changes in furnishings and décor.
navigation is the need for digital building plans.          Another aspect is the different ways in which
For most buildings which have architectural              users interact with the navigation application.
and construction plans, the plans are usually            In car navigation, the device used for naviga-
only available as hard copies. CAD drawings              tion is mounted on a special holder within the
are available for newer buildings, but, usually,         driver’s field of vision. In pedestrian naviga-
there is no standard to define how to record             tion, the device is mostly held in the hand, and
information about the structure of rooms, corri-         looking at the app’s screen requires lifting the
dors, or floors; although some standards do              device and often involves stopping; otherwise,
exist, for example, the IFC (Industry Founda-            there is a risk of tripping or bumping into another
tion Classes). It is therefore necessary, to digitise    person. In the case of in-car navigation, audio
and adapt existing plans or create them anew.            messages are only heard by the driver and
Documentation of this type is not widely avail-          passengers concerned. In the case of indoor
able, and the free mapping of building interiors         navigation, the inconsiderate use of voice an-
is limited for several reasons. The best situation       nouncements would be annoying to those
is in the case of new buildings, where the latest        around. On the other hand, the amount of noise
BIM (Building Information Modelling) metho­              in a shopping mall or airport may prevent the
dologies and technologies have been applied              audible message from being heard. The use of
during construction. However, using this data            direct experience which is designed for car na-
in a navigation application requires permits             vigation applications is, therefore, of very limited
and carrying out several processes. Solutions            use in pedestrian navigation.
combining BIM and GIS (geographic informa-                  The issues mentioned above are only a small
tion system) technologies are promising, none-           number out of many examples which indicate
theless (Isikdag et al., 2013).                          the specificity of indoor navigation and the
   The specificity of indoor navigation also ge-         need for research. The development of state-
nerates a whole other set of challenges for              -of-the-art technological solutions to support
cartographers and computer scientists. The               the design of indoor navigation applications
representation of the building has to take into          for high-quality cartographic information can
account three dimensions, at different levels of         be a crucial element in the development of car-
generalisation (Gotlib et al., 2020). Over years         tography (Chen & Clarke, 2020).
of cartographic development, the principles for
the generalisation and cartographic presenta-
tion of open space data have been developed.
                                                         2. The research environment concept
The cartographic modelling of buildings has              and methodology
only been needed relatively recently, so the
process of developing and testing the rules and             Due to the specificity of indoor navigation
principles for this area is still ongoing (Huang         applications and the need for these applica-
et al., 2018).                                           tions to be properly cartographically designed,
   In the process of designing an indoor navi-           a concept was proposed consisting of the pre-
gation application, it is crucial to understand          paration of a research environment in the form
how people and vehicles move and behave,                 of a simulator using game technology. The
which is different in buildings to what is familiar      process of conducting multifaceted research
110                                   Jakub Łobodecki, Dariusz Gotlib

(including on the perception of navigation             in the design of buildings and physical signage
messages and ways in which buildings are               in interiors.
cartographically presented) will consist of simu-         The planned research environment could
lating the navigation process in a virtual world,      also serve as an educational environment in
which allows for more efficient research and           the future. There are well-known examples of
reduced costs compared to real-world tests (in an      the use of virtual reality in teaching geography
actual building and not its digital twin). A SWOT      (Šašinka et al., 2018). Building models in con-
analysis comparing the two approaches has              junction with a building viewing system, would
already been carried out (Kinateder et al., 2014b).    be able to function as educational material for
When navigating in a virtual building, it is pos-      architects and interior designers, or as a training
sible to generate multivariate graphic, auditory,      field for people with limited mobility.
and vibrational navigation cues along with                The proposed concept would also be a si-
multi-scale representations of the building, and       mulator-integrated way to collect information
identify areas causing wayfinding difficulties.        about the effectiveness of different types of
The use of simulators does not eliminate testing       navigational cues, and to collect feedback on
in a real-world environment but complements            the feelings, perceptions, and preferences of
it, and, in some aspects, allows for testing which     users participating in tests. The research could
is not possible in the real world, for example,        be conducted through a diagnostic survey or
for safety reasons. Good examples of this are          focused interviews.
the use of a driving simulator to test the impact         In the first research stage described in the
of a car navigation display’s placement on             master thesis by Łobodecki (2020), and in this
driver safety (Ishiko et al., 2014), and the social    article, only a simplified test study was carried
impact tests on the choice of escape route from        out on a selected group of users to test the
a burning tunnel carried out in virtual reality        proposed solution from a methodological and
(Kinateder et al., 2014a). It is also worth noting     technological point of view. The interaction with
research that examines the effect of how a driver      the virtual reality environment in the pilot study
navigates at the level of spatial information          was based on a desktop environment with
recall (Khan & Rahman, 2018). Thanks to the            keyboard and mouse interaction, rather than
virtual testing environment, it is possible to         head mounted VR googles. This is not yet
conduct multiple tests under identical conditions      a target study, but a preliminary validation of the
(lighting, level of congestion in the building, lift   method and technology used. For this reason,
waiting time, etc.).                                   the following section refrains from presenting
    An important factor that can significantly         the detailed results collected in the research
impact this type of research is the level of im-       questionnaire. Instead, only basic information
mersiveness of the simulation environment.             is presented to show the specifics of this type
This is dependent on the graphics quality (Jelfs       of survey. Conclusions from this stage were used
& Whitelock, 2000) and how realistic the ac-           to refine the technology and the proposed
companying sound design of the simulation is           method for conducting research in a virtual
(Çamcı & Hamilton, 2020). In addition, how the         environment. The results of specific research
simulation is controlled – with a mouse and            on, among other things, the perception of navi-
keyboard or with a professional simulator              gation directions using the proposed solution
equipped with appropriate sensors (e.g., pedals,       will be presented in another publication.
steering wheel and gearstick in the case of a car)        For testing purposes, a survey was conducted
– also has a significant influence. Other issues       to compare different types of navigational
related to the design of the navigation applica-       guidance. Comparisons of this type have already
tion can also be tested. For example, the quality      been carried out in the past, but in a different
of the positioning signal can be simulated,            environment and using different methods. An
including various simulated configurations of          outdoor environment comparison between
its components. Data analysis (data mining)            visual and audio guidance systems is avail­
techniques can also be used in the design              able by Chittaro and Burigat (2005). In the case
process (Sattarian et al., 2019). Analysing and        of indoor navigation, a survey of user preferences
understanding user movement will allow for             on how navigation directions are communicated
better application design and can also be used         was executed by De Cock et al. (2019). Also,
                   Developing a simulator of a mobile indoor navigation application as a tool...            111

the first studies on the use of virtual reality in           make it easier for future researchers who are
indoor navigation analysis have been carried                 not programmers to continue the work. An
out (De Cock et al., 2022).                                  additional advantage is that this engine allows
  In the future, the prepared environment is                 the best visual effects to be achieved with the
expected to allow for a high quality, compre-                least amount of effort, as found in comparative
hensive study of these issues as well.                       tests (Stylized Station, 2021).
  The initial aim of this study was to develop                  The Unreal Engine is a complete set of tools
an environment for such testing rather than to               for, among other things, creating games and
conduct the testing itself, which will be carried            architectural visualisations. The engine is written
out in subsequent stages.                                    in C++, and its developers have implemented
                                                             mechanisms that support cross-platform capa-
3. Selection of game engine technology                       bilities. Supported platforms include Microsoft
                                                             Windows, macOS, Linux, iOS, Android, Nintendo
   The intensive development of three-dimen-                 Switch, PlayStation 4, Xbox One, HTC Vive,
sional visualisation technology is being driven              Oculus Rift, and PlayStation VR, among others.
mainly by the gaming industry. According to the              The first version of the engine was released in
State of Polish Video Games Industry (Kraków                 1998, so it has been in development for more
Technology Park & The Polish Gamers Obser­                   than 20 years. In March 2015, the engine was
vatory, 2020) report, the global electronic en-              made available for free to all developers. For
tertainment market is growing rapidly. In 2020,              commercial use, it is necessary to pay a royalty
the number of gamers reached 2.7 million,                    of five per cent of revenue after it exceeds one
which translates into industry revenues of                   million dollars.
US$ 175 billion.                                                The prepared environment was tested in both
   Because games, in addition to being an ad-                an older version of the engine, 4.25, and ver-
vanced technology, are a form of creative art,               sion 5.0, released in April 2022.
the diversity of individual titles shows how much
freedom electronic entertainment creators have.              4. Description of the developed research
At the same time, the game development pro-                  environment prototype
cess primarily uses existing elements. A col-
lection of such configurable elements forms                     The Main Building of the Warsaw University
software called a game engine.                               of Technology was selected as the area for
   Currently, Unity and Unreal Engine are the                virtual testing. The choice was determined by
two most popular game engines in the industry                its complexity and the availability of an advanc­
(Toftedahl, 2019). Even though Unreal Engine                 ed 3D model for this building, which allowed
has almost twice as many games released on                   for advanced testing. The three-dimensional
the Steam platform,1 Unity is almost ten times               model of this building was created as a result
more popular then Unreal in Google searches                  of transformations carried out on source spa-
(Google Trends, 2022), which indicates that it               tial data collected by the Department of Carto-
is the choice of individuals and hobbyists. Less             graphy of the Warsaw University of Technology
popular engines, but still worth mentioning, are             during various scientific and implementation-
CryEngine, Godot, and GameMaker.                             -type projects, including work related to the
   After analysing the literature (Christopoulou             creation of the Property Information System of
& Xinogalos, 2017; Ciekanowska et al., 2021),                the Warsaw University of Technology (Gotlib
the decision was made to choose the Unreal                   & Gnat, 2018). The transformations were per-
Engine software, produced by Epic Games.                     formed as part of a paper (Janicki, 2020) in
This choice was dictated by several specific                 which the author used the Unity game engine.
features of this engine. One particular feature              The source data were provided in FBX2 format,
of the Unreal Engine is the visual programming               which resulted from the preparation of the data-
system (called Blueprint). Using this system to              base according to the Indoor Data Model (IDM)
build a research environment prototype will
                                                                 FBX – one of the main 3D data storage formats for
      A platform which sells games:   exchanging data between applications.
112                                 Jakub Łobodecki, Dariusz Gotlib

  Fig. 1. The 3D model of the Warsaw University of Technology’s Main Building, displayed in CityEngine
                                      software (Janicki, 2020)

developed by the Department of Cartography              the DataSmith functionality, first presented in
of the Warsaw University of Technology.                 conjunction with CityEngine, in 2017.
  The model, prepared as described above,                 The export parameters were chosen to make
was used in this research with the authors’             the entire building model visible as a single
permission. For this purpose, it was initially          object in the game engine. The final size of the
exported from CityEngine (fig. 1) to a format           model, including all the textures, was 97 MB.
designed for the Unreal Engine (.udatasmith               The mechanics of moving around the virtual
and .udsmesh). This was possible thanks to              building model using a first-person view were

                           Fig. 2. First-person view inside the virtual building
                 Developing a simulator of a mobile indoor navigation application as a tool...         113

implemented and assisted by the standard                 The following operating scenario was assumed:
computer game control of a virtual character,               1. On a desktop computer, the user launches
using a keyboard and mouse.                              the research application, after which, intro-
   The user moved around the virtual building            ductory information about the research and
using a first-person view (fig. 2). Moving was           the rules for using the application is obtained.
done by using the keys for the letters WASD              Additional statistical information, such as age
(as in video games) or by using the arrow keys           and gender, is also collected at this point for
(which may be more convenient for those with­            research purposes.
out gaming experience). Looking around was                  2. A 3D model of the facility’s interior is dis­
done by moving the mouse.                                played, which the user can navigate freely using
   In addition, elements were created to mimic           a keyboard and mouse.
the various components of mobile navigation                 3. The application delivers successive test
systems. Navigation clues were displayed on              routes to be followed and then navigates the
a partially visible virtual mobile phone. It was         user through the use of appropriate navigation
also possible to display a map of the building           messages (in different versions and at differ­
with the user’s current position marked on it.           ent points of their presentation), as well as
Additionally, this map was oriented in the di-           enabling the display of an on-demand naviga-
rection in which the virtual character was cur-          tional support map in the form of a 2D visuali-
rently looking.                                          sation.
   Upon completion of each stage of the ex-                 4. The application monitors the user’s beha-
periment, a set of statistical data was sent,            viour and regularly uploads information to a data-
allowing a research analysis to be carried out.          base.
The data collected included the time taken to               5. After the test runs, the application collects
complete a stage or the number of additional             additional information from the user for research
prompts the participant needed to complete               purposes, such as which type of navigational
a task. The final element of the system was              directions they preferred.
that the participants were able to freely provide           6. The researcher analyses the collected data.
their own comments and observations about the
research process and the way it is navigated.             6. The pilot study
   The environment described above created
a prototype of the research application which               In the initial part of the survey, participants
was tested in this study. The term “research             were introduced to the context of the survey
application” is understood here as an applica-           and how it would be conducted.
tion that is used to conduct scientific research.           The user was then given tasks to complete
The details and results of the tests carried out         which involved reaching various destinations.
are described later in this paper.                       The user (test participant) took the different
                                                         routes several times using different navigation
5. Study design                                          messages.
                                                            In the experiment, four different types of
   As part of the testing environment, it was            navigation messages were prepared, which
decided to test several types of navigation              were limited to three functions:
messages. The following types of guidance                   – forward movement,
were designed and placed in the simulator:                  – movement to the left,
   – a graphic direction arrow similar to traffic           – movement to the right.
signs, in two versions: large size and small size;          The first type of message was a graphic
   – a graphic direction sign, in different colours;     message resembling a road sign indicating
   – an audio message;                                   a “mandatory direction of movement” (fig. 3) in
   – a text message.                                     two size variants.
   It was assumed that the messages would be                The second graphical solution proposed
displayed by the test application (simulator) at         used colour to distinguish left and right turns
different moments during the user’s approach             (fig. 4). Furthermore, a texture indicating the
to the point at which the manoeuvre was to be            corresponding direction was additionally super-
performed (e.g. a fork in the corridor, a staircase).    imposed in order to make it easier to associate
114                                   Jakub Łobodecki, Dariusz Gotlib

                                                          Figures 7, 8, 9, and 10 show the appearance of
                                                       the prepared research application when in use.
                                                          The upper part of the virtual mobile phone
                                                       screen is displayed in the bottom left corner of
 Fig. 3. Navigation message in the form of arrows      the screen (fig. 8). Here, the various types of
                similar to road signs                  navigation messages appear. In addition, the
                                                       time since the start of the current stage is dis­
                                                       played to motivate the user to complete the
                                                       task more quickly.
                                                          When the SPACE bar is pressed, the rest of
                                                       the phone is made visible, which simulates the
      Fig. 4. Colour-coded navigation message          user looking at their smartphone during navi-
               with direction indication               gation (fig. 9). The navigation map of the building
                                                       is visible and oriented relative to the direction
                                                       in which the user is looking. The map shows
a particular colour with a direction, and also to      the user’s current position and the route the
enable people with daltonism to recognise the          participant should take. While the map is dis­
direction. The choice of colours was inspired by       played, it is impossible to move, and each use
how directions are marked in nautical sailing          of the map “consumes” (simulates) the battery
(Czajewski, 1991).                                     power of the virtual phone. This functionality
   The third type of message was a text message        was aimed at getting the participants to follow
displaying text with a simple command “Turn            the route as quickly as possible and rely mainly
left”, “Turn right”, or “Continue straight ahead”      on the navigation messages rather than looking
(in Polish: “Dalej prosto”) (fig. 5).                  around the virtual building.
                                                          When the user reaches the door marked
                                                       with a star (the destination – fig. 10), a summary
                                                       of the survey’s progress is displayed in the
                                                       form of a series of stars indicating the number
                                                       of successful tasks completed (fig. 11). It is
                                                       also possible to abort the survey before it has
                                                       been fully completed to give the user the
                                                       option of reducing the overall survey time. In
                                                       such a case, only fully completed stages are
                                                       considered for further analysis.
         Fig. 5. Example of a text message                In the prototype application, seven different
      (Message text: “Continue straight ahead”)        stages were implemented, requiring the navi-
                                                       gation of seven diverse routes using several
  The fourth and final type of message was             variants of the navigation directions (tab.1). In
a voice message, symbolised by the display of          target studies, the choice of routes should be
a speaker icon in the message box (fig. 6). At the     carefully considered. Their difficulty, repeatability,
same time, a voice instruction with the same           and length are important. This is a complex issue
content as the text message was played.                and will be the subject of further research.
                                                          At the end of the study, a final screen was
                                                       displayed thanking the user for participating
                                                       in the study and presenting further survey
                                                       questions concerning the following:
                                                          – the best type of navigation message as
                                                       perceived by the user,
                                                          – the need for indoor navigation solutions,
                                                          – general comments and observations after
                                                       participating in the survey.
                                                          In addition, participants could enter their
           Fig. 6. Voice message indicator             email address to receive a summary of the
                (source:                 survey results.
                Developing a simulator of a mobile indoor navigation application as a tool...          115

Fig. 7. Screenshot of the developed application from the experiment’s starting point. A view of the 3D space
   from a first-person perspective is shown, as well as the interface elements: a virtual phone displaying
         navigational directions and text with supporting instructions (the text shown on this screen:
             “Move around using the WASD or arrow keys. Look around by moving the mouse.”)

                 Fig. 8. Screenshot showing the appearance of the developed application

  The developed prototype application displayed         pant had to be in the area which triggered the
navigation messages when the player (partici-           message (trigger box), for example, approach­ing
pant) fulfilled two conditions. First, the partici-     a fork in the route. Second, the player had to
116                                    Jakub Łobodecki, Dariusz Gotlib

      Fig. 9. The map tooltip displayed on user request (the text guidance on this screen is “Turn right”)

“look” in the right direction (using proper mouse            In order to collect and save research data,
operation). The frequency and size of the trigger         three communicating components were used:
areas varied between the different stages.                   1. An add-on to Unreal Engine 4 called VaRest
   In addition to information about the participant,      allowed HTTP requests of the POST and GET
the application collected information about               types to be sent quickly. The request was ac-
their behaviour, as shown in table 2.                     companied by a JSON language element con-

          Fig. 10. Screenshot of a star indicating the destination of a stage. The Polish text reads:
                  “Your task is to follow the directions to reach the door marked with a star”
                 Developing a simulator of a mobile indoor navigation application as a tool...              117

   Fig. 11. Screenshot displayed between each stage. On the application screen, the text in Polish reads,
         top to bottom: “Good job!”, “Get all the stars”, “Next”, “Finish the experiment before the end”

Table 1. Description of the implemented stages and the navigation message variants
 Stage                                      Description
                                                                                               message variant
         Introductory stage geared towards familiarising participants with the controls        Large arrows
         and principles of the application.
         Cues are displayed immediately before the manoeuvre is required                       Random
   2     (before a turn, but not when the participant is already at it). A relatively small
         number of navigation directions.
         Guidance is displayed directly at a turn, with additional confirmation via            Random
         the message “Continue straight ahead” when heading in the correct direction.
         A stage in which navigational directions are displayed along the entire route.        Random
   4     At any point along the way, while going in the right direction, the participant
         sees messages confirming further directions.
         A route analogous to Stage 4, but with a minimum number of navigation                 Random
         directions; i.e. only before places requiring a manoeuvre.
         A complicated and elaborate stage using only a few navigation directions, and         Random
   6     starting in a different place than the earlier stages. Due to the small number of
         navigation directions, situations having ambiguous further directions are possible.
         A special stage using extended text navigation directions. This stage requires        Text navigation
   7     the participant to stop, read a longer message, and then go to the next location      directions based
         based on the memorised information.                                                   on landmarks

taining a set of data stored in the key-value                3. A Google Sheets spreadsheet (Table 3),
form. This element was pre-populated with the              whose task was to collect the data on the im-
data collected during the experiment.                      plementation of each stage in a structured way.
    2. A cloud-based application based on the
Google Apps Script platform. The task of this              7. Preliminary results of the pilot study
application was to receive and process the raw
data sent by the VaRest add-on and then save                 Students and acquaintances of the study’s
it to a spreadsheet.                                       authors, a total of 36 people, were invited to
118                                   Jakub Łobodecki, Dariusz Gotlib

Table 2. Summary of data collected by the research application on completion of each stage

      Information collected                                       Description
Type of guidance                  Types of randomly selected navigation directions.
Duration                          The duration of the experiment in seconds. Duration for each stage and
                                  the total time of participation in the study.
Number of wrong turns             This number indicates how many times the participant went in the wrong
                                  direction for a given stage.
Number of map views               This number indicates how many times the participant needed to use
                                  the 2D map display for a given stage.
Time spent looking at the map     The duration in seconds the participant spent analysing the 2D map
                                  in search of the proper route.
Coordinates of the map display    Coordinates in the local XYZ system indicate where the participant
                                  displayed the 2D map within the building.

Table 3. Excerpt from the spreadsheet collecting the results of the experiment

                         Related                   Guid­                                      Was
                                 Guid­                    No of No of                                  navia­
                             to           Total    ance                    Coordinates       in the
 No Gender      Age              ance                      mis- map                                   gation
                         reseach          time    contact                  of map views      build­
                                 type                     takes views                                  apps
                           field                   time                                        ing
  1   woman    0 – 18      no     sign    55.82    25.64      0      0      3191.268799,      yes       yes
                                                                          1412.161987] …
  2   woman 40 – 60        no     sign   109.22    73.12      1      2      1638.719971,      yes       yes
  3    man     19 – 25     no     sign    29.68     2.51      0      5      2320.653564,      yes       yes
                                                                          1496.961914] …
  4   woman 19 – 25        no     sign    95.52    14.73      2      1      2674.232666,      no        no
                                                                          1412.161987] …
  5   woman 19 – 25        yes    sign    58.35     5.15      0      3      2962.134033,      no        yes
                                                                          1432.161987] …
  6    man     19 – 25     yes    sign    35.16        0      0      0           []           yes       no

  7   woman 19 – 25        yes    sign    40.17        0      0      0           []           no        yes
  8    man     19 – 25     yes    sign    90.40     8.09      4      3      2741.678223,      yes       no
                                                                          1432.201416] …
  9   woman 25 – 40        yes    sign   153.47     9.07      0      4      3242.214111,      yes       no
                                                                          1432.161987] …
 10    man     25 – 40     no     sign    93.55    34.96      4      11     3242.214111,      yes       yes
                                                                          1432.161987] …
                Developing a simulator of a mobile indoor navigation application as a tool...          119

participate in the experiment. As in any survey,        Table 4. Summary of the most common comments
information about the participants’ characteris-        from the questionnaire at the end of the study
tics was collected first (gender, age, associa-                               Comment
tion with the field of surveying and mapping,            Mouse and keyboard control, and lack
familiarity with the building in which the survey        of experience with computer games, makes
was being conducted, regularity of use of navi-          the survey very difficult.
gation applications, etc.); then participants            Suggestion for introducing up or down stairs
were given several tasks which consisted of              guidance messages.
navigating a virtual building and reaching a de-         At some stages the guidance messages were
stination, during which the app monitored their          displayed for too short a time.
behaviour. At the end of the experiment, the             There was a clear indication of the predominance
participants were asked what they felt was the           of graphic guidance messages over text and
best way to provide navigational information             audio messages.
out of all those tested.                                 Noting the advantage of voice guidance when
   In the final phase of the pilot study, the parti-     it was inconvenient to look at the phone, and also
cipants were given the opportunity to make               the usefulness of this message type for the blind
their own comments and observations about                people.
the study. Twenty out of the 36 participants
who reached this stage took this opportunity.
The most common comments are summarised                 parative tests of wayfinding tasks, in which the
in the table 4.                                         virtual environment was directly comparable
   The results given here are fragmentary and           with the real environment, indicate that both
should only be regarded as an example of the            environments can be comparable across all
possibilities of the created test environment.          important aspects (Stachoň et al., 2022).
Therefore, the full results of the experimental            Analysis of the questionnaires completed by
study of the users’ perceptions are intentionally       the users showed that the main drawback of
not described in this article. Further research         the developed test environment was the com-
requires the collection of an appropriately se-         plexity of the controls, which may make it difficult
lected group of test participants, the refinement       for people without gaming experience to parti-
of the application based on the conclusions of          cipate in such a study. Improvements in this
the first stage of research, the methodological         area could be achieved, among other ways,
preparation of the survey, the inclusion of addi-       through virtual reality technology, for example,
tional navigation cue variants, etc. Experiments        VR goggles and multi-directional treadmills,
of this type will be carried out in subsequent          and additional training of individuals before the
stages of the research.                                 start of a study. A lack of skills in the gaming
   For additional familiarisation with the devel-       environment may also affect the study’s outcome,
oped prototype research tool described here, it         which should be considered when formulating
is possible to download the zipped application          further study methodology.
files located at the following address:                    The prototype application to simulate the            navigation process presented in this thesis is
V3DtnvxIRq5HKIiTnTkb6.                                  not yet a complete solution. More reliable and
                                                        extended analyses will be possible with the
8. Conclusions and further work                         further development of the proposed research
                                                        methodology and the precise preparation of
   The research and the technology has shown            test routes and navigation messages. The
that it is possible to create an environment            experiences gathered during the prototype’s
simulating an indoor navigation application in          development, the observations of the partici-
a relatively simple way. The use of game tech-          pants, and the feedback collected from them,
nology in combination with geospatial data              as well as the review of related literature, allow
opens up new research opportunities in the              us to outline the future direction of research
field of cartography, particularly in the field of      on the mapping aspects of indoor navigation
studies on the user perception of maps and              application design, in particular research on
other geo-information products. The first com-          the perception of navigation messages. In the
120                                   Jakub Łobodecki, Dariusz Gotlib

context of the development of the proposed si-          planned route. In addition, it is advisable to add
mulation environment, the possibility of changing       the functionality of a lift ride, which is an integral
the scale of the displayed 2D map within the            part of navigating large buildings.
simulator, the display of 3D maps, the use of              In conclusion, it can be said that the conducted
different colours for navigation maps, different        study of the prototype original test environ-
object signatures, etc., should be investigated.        ment showed that, despite some limitations,
Furthermore, the presentation of an entire              game technology could successfully be used
planned route on a map could significantly              at the stage of designing and testing cartogra-
impact the level of orientation within the building     phic aspects of navigation systems.
space when following subsequent directions.                The developed prototype test environment
The surroundings of the building should also            was built in such a way as to allow its further
be added, so navigation outside and inside the          development until a fully comprehensive simu-
building can be integrated.                             lation test system for navigation products is
   Another essential element of the simulator           achieved. The described preliminary research
is the realism of the rendered 3D graphics.             proved the feasibility of the task and allows for
Among other things, a check should be made              the formulation of assumptions for subsequent
in subsequent tests as to whether the schematic
                                                        research stages.
model of the building shown in the simulator is
sufficient to maintain the tests’ reliability, and
                                                           Participation of authors. D. Gotlib: general
what effect using a photorealistic model would
                                                        conception of the research and testing envi-
have on the perception of the participants. In
                                                        ronment, supervision of the work, testing of the
addition, the model should be equipped with
                                                        system, editing of the article’s content. J. Ło-
spatial orientation aids analogous to those
found in the real world, such as directional            bodecki: development of the system prototype,
signs, door numbers, and floor identifiers.             system tests, conducting test studies, editing
   In order to make the simulation more immer-          the article’s content.
sive, it would be worth adding additional ele-             Acknowledgements and additional infor­
ments that can influence the users’ perception.         mation. The article describes, in its main part,
Among other things, it may be essential to si-          work carried out as part of a master’s thesis by
mulate the sounds of the surroundings as well           Jakub Łobodecki, M.Sc. (2020), conducted at
as simulate other people in the building to             the Department of Cartography of the Warsaw
study their impact on the user’s reactions and          University of Technology, according to the con-
feelings.                                               cept by, and under the supervision of Professor
   The mechanics of moving from floor to floor          Dariusz Gotlib. The work was realised using
also require special attention. It is necessary         a building model and other data developed as
to provide a specific type of message for this          part of various works carried out at the Depart-
type of manoeuvre and to exclude the risk of            ment of Cartography of the Warsaw University
“falling” down the stairs, which in the current         of Technology. The authors would like to thank,
version of the application could occur and,             in particular, Miłosz Gnat, M.Sc. and Hubert
thus, prolong the time it takes to travel the           Janicki, M.Sc.

Çamcı, A., & Hamilton, R. (2020). Audio-first VR:         An evaluation of three approaches. Proceedings of
  New perspectives on musical experiences in virtu-       the 7th International Conference on Human Com­
  al environments. Journal of New Music Research,         puter Interaction with Mobile Devices & Ser­vices,
  49(1), 1–7.        107–114.
  9.1707234                                             Christopoulou, E., & Xinogalos, S. (2017). Overview
Chen, J., & Clarke, K. C. (2020). Indoor cartography.     and comparative analysis of game engines for
  Cartography and Geographic Information Science,         desktop and mobile devices. International Journal
  47(2), 95–109.        of Serious Games, 4(4), Article 4.
  2019.1619482                                            10.17083/ijsg.v4i4.194
Chittaro, L., & Burigat, S. (2005). Augmenting audio    Ciekanowska, A., Gliński, A. K., & Dziedzic, K. (2021).
  messages with visual directions in mobile guides:       Comparative analysis of Unity and Unreal Engine
                  Developing a simulator of a mobile indoor navigation application as a tool...                 121

   efficiency in creating virtual exhibitions of 3D          Janicki, H. (2020). Visualization of the Main Building
   scanned models. Journal of Computer Sciences                of Warsaw University of Technology using a se­
   Institute, 20, 247–253.           lected game engine [Bachelor’s thesis]. Warsaw
   jcsi.2698                                                   University of Technology, Faculty of Geodesy and
Czajewski, J. (1991). Locja śródlądowa morska. Wy-             Cartography.
   dawnictwa Komunikacji i Łączności.                        Jelfs, A., & Whitelock, D. (2000). The notion of pre-
De Cock, L., Ooms, K., Van de Weghe, N., Vanhae­               sence in virtual learning environments: What
   ren, N., & De Maeyer, P. (2019). User preferences           makes the environment “real”. British Journal of
   on route instruction types for mobile indoor route          Educational Technology, 31(2), 145–152. https://
   guidance. ISPRS International Journal of Geo-In­  
   formation, 8(11), 482.           Khan, N., & Rahman, A. U. (2018). Rethinking the
   ijgi8110482                                                 mini-map: a navigational aid to support spatial
De Cock, L., Van de Weghe, N., Ooms, K., Saenen, I.,           learning in urban game environments. Internatio­
   Van Kets, N., Van Wallendael, G., Lambert, P.,              nal Journal of Human-Computer Interaction,
   & De Maeyer, P. (2022). Linking the cognitive load          34(12), 1135–1147.
   induced by route instruction types and building             318.2017.1418804
   configuration during indoor route guidance, a usa-        Kinateder, M., Ronchi, E., Gromer, D., Müller, M.,
   bility study in VR. International Journal of Geogra­        Jost, M., Nehfischer, M., Mühlberger, A., & Pauli, P.
   phical Information Science, 1–31.         (2014a). Social influence on route choice in a vir-
   0.1080/13658816.2022.2032080                                tual reality tunnel fire. Transportation Research
Google Trends. (2022, August 23). Comparision of               Part F: Traffic Psychology and Behaviour, 26, Part A,
   phrases: “unreal engine” and “unity”. https://trends.       116–125.,unity          Kinateder, M., Ronchi, E., Nilsson, D., Kobes, M., Mül-
Gotlib, D., & Gnat, M. (2018). Koncepcja i prototyp            ler, M., Pauli, P., & Mühlberger, A. (2014b). Virtual
   wielofunkcyjnego systemu informacji przestrzen-             reality for fire evacuation research. Proceedings
   nej wspomagającego zarządzanie i użytkowanie                of the 2014 Federated Conference on Computer
   nieruchomości Politechniki Warszawskiej. Rocz­              Science and Information Systems, 313–321. https://
   niki Geomatyki - Annals of Geomatics, 16(4(83)),  
   299–318.                                                  Kraków Technology Park, & The Polish Gamers Ob-
Gotlib, D., Wyszomirski, M., & Gnat, M. (2020). A sim-         servatory. (2020). The state of Polish Video games
   plified method of cartographic visualisation of             industry 2020.
   buildings’ interiors (2D+) for navigation applications.     tent/uploads/2021/03/raport_2021_long_final_
   ISPRS International Journal of Geo-Information,             en_web.pdf
   9(407), 407.          Łobodecki, J. (2020). Development of a test environ­
Health IT Outcomes. (2015, June 30). Robert Wood               ment for analysing the perception of users of in­
   Johnson University Hospital launches Connexient’s           door navigation applications using game technology
   Medinav Navigator Edition 2.0 – first true indoor           [Master’s thesis]. Warsaw University of Technology,
   navigation in a hospital in North America. https://         Faculty of Geodesy and Cartography.                 Marciniak, J. (2018). Using cartographic models of build­
   johnson-launches-medinav-navigator-navigation-              ing interiors to improve the quality of the navigation
   hospital-north-america-0001                                 process [Doctoral dissertation]. Warsaw University of
Huang, H., Gartner, G., Krisp, J. M., Raubal, M., & Weg­       Technology, Faculty of Geodesy and Cartography.
   he, N. V. de. (2018). Location based services: On-        Sakpere, W., Oshin, M. A., & Mlitwa, N. B. (2017).
   going evolution and research agenda. Journal of             A state-of-the-art survey of indoor positioning and
   Location Based Services, 12(2), 63–93. https://             navigation systems and technologies. South African                       Computer Journal = Suid-Afrikaanse Rekenaar­
InMapz. (2022, September 12). About InMapz. https://           tydskrif, 29(3), 145–197.                              sacj.v29i3.452
Ishiko, H., Nakano, K., Zheng, R., Hagita, K., Kihira, M.,   Šašinka, Č., Stachoň, Z., Sedlák, M., Chmelík, J.,
   Yokozeki, T., Takayanagi, M., & Yano, K. (2014).            Herman, L., Kubíček, P., Šašinková, A., Doležal, M.,
   Evaluation of safety of navigation systems using            Tejkl, H., Urbánek, T., Svatoňová, H., Ugwitz, P.,
   a driving simulator. Seisan Kenkyu, 66(2), 173–178.         & Juřík, V. (2018). Collaborative immersive virtual                environments for education in geography. ISPRS
Isikdag, U., Zlatanova, S., & Underwood, J. (2013).            International Journal of Geo-Information, 8(1), 3.
   A BIM-Oriented Model for supporting indoor navi-  
   gation requirements. Computers, Environment               Sattarian, M., Rezazadeh, J., Farahbakhsh, R., & Bag­
   and Urban Systems, 41, 112–123.            heri, A. (2019). Indoor navigation systems based
   10.1016/j.compenvurbsys.2013.05.001                         on data mining techniques in internet of things:
122                                      Jakub Łobodecki, Dariusz Gotlib

  A survey. Wireless Networks, 25(3), 1385–1402.            Stylized Station. (2021, April 26). Should I learn Unity                   or Unreal Engine?
Stachoň, Z., Jochecová, K., Kvarda, O., Snopková, D.,         watch?v=OH2sYJw52BQ
  Ugwitz, P., Šašinková, A., Ježek, S., Kubíček, P.,        Toftedahl, M. (2019, September 30). Which are the
  Juřík, V., Švedová, H., & Sasinka, C. (2022). The           most commonly used Game Engines? https://
  possibilities of using virtual environments in research
  on wayfinding [Preprint]. In Review. https://doi.           -the-most-commonly-used-game-engines-

Developing a simulator of a mobile indoor navigation application as a tool for cartographic research

Authors Dariusz Gotlib Jakub Łobodecki

License CC-BY-4.0