Providing Route Instructions with Speech

goldbashedΤεχνίτη Νοημοσύνη και Ρομποτική

15 Νοε 2013 (πριν από 3 χρόνια και 4 μήνες)

53 εμφανίσεις













Providing Route Instructions with Speech



Kaj Mäkelä & Perttu Prusi


kaj@cs.uta.fi

prusi@cs.uta.fi


Project paper for

Speech Interface Design course

Autumn 2001


Department of Computer Science

The University of Tampere













2

Index


Keywords

________________________________
_________________________

3

Introduction

________________________________
_______________________

3

Definitions

________________________________
________________________

3

Wayfinding

________________________________
______________________

3

Navigation

________________________________
_______________________

3

Orientation

________________________________
______________________

4

Route

________________________________
___________________________

4

Landmark

________________________________
_______________________

4

Types of route guiding

________________________________
______________

5

In
-
vehicle route guidance and information systems

_______________________

5

Blind and visually impaired travellers

________________________________
_

5

Office environment

________________________________
________________

6

Distance vs. land
marks

________________________________
_____________

6

Characteristics of speech in route guiding

______________________________

7

Speech and guiding

________________________________
________________

7

Pr
os

________________________________
____________________________

7

Cons

________________________________
___________________________

8

Current applications

________________________________
________________

8

The Doorman

________________________________
____________________

8

VODIS

________________________________
_________________________

9

MoBIC system

________________________________
___________________

9

User Requirements

________________________________
_____________

10

Personal Guidance System

________________________________
_________

10

Route guiding in Personal Guiding System

__________________________

10

User interface of the Personal Guidance Device

______________________

11

Constructing the route information

________________________________
__

11

Accuracy

________________________________
_______________________

11

Scaling distances

________________________________
_________________

12

Scaling directions

________________________________
________________

12

Evaluating landmarks

________________________________
_____________

13

Conclusion

________________________________
_______________________

13

References

________________________________
_______________________

14



3

Keywords

Wayfinding, route information, spatial data, speech user interface, navigation,
orientation, route, landmark.


Introduction

Users often need guidance in unfamiliar environment
. They may know only the
street address or the city of the destination or the name of the person they are
visiting. To be able to automatically plan the route for the user the system
needs to know available and suitable routes and conditions on them, the
s
tarting point and the destination of the route. The user needs to provide the
information on the destination and special needs for the route. After finding a
route to the destination it needs to be conveyed to the user in the memorable
and simple form that

is suitable for conditions. The route can be presented to
the user with different modalities, for example, visually or aurally. [Streeter et
al. 1985.]


Our own motivation to this study is the development of the receptionist system
that we develop in the
premises of our faculty. This Doorman system lets
visitors and staff in our units’ premises and guides visitors to their destinations.
During our study we have found out that there seem to be few papers written
from this sort of indoor route guiding system
s that do not require any mobile
devices or electronic badges carried by the user.


Definitions

There are some basic concepts concerning the route information that will be used
frequently. In this paragraph these concepts will be defined.


Wayfinding

Wayfi
nding

is the cognitive element of navigation. It does not involve movement of
any kind, but only the tactical and strategic parts that guide movement. Wayfinding
and motion are intimately tied together in a complex negotiation that is navigation. An
essent
ial part of wayfinding is the development and use of cognitive map, also
referred as a mental map. Still poorly understood, a cognitive map is a mental
representation of an environment. This “picture in the head “ has been proven to have
a symbolic quality
, in addition to imaginary quality.


Navigation

Navigation

is the aggregate task of wayfinding and motion. It inherently must have
both the cognitive element (wayfinding), and the motoric element (motion).
Navigation tasks are essential to any environment
that demands movement over large
spaces. [Darken, Peterson 2001].



4

Orientation

Orientation
is our awareness of the space around us, including the location of
important objects in the environment. Orientation in space is crucial for finding one’s
way (or wa
yfinding) from one location to another. [Hunt, Waller 1999.]


Route

The
route

is a central concept in the process of finding a place. A route is a sequence
of nodes (locations) where a wayfinder selects a new bearing. In between nodes there
are segments th
at have to be travelled [Hunt, Waller 1999]. These nodes and segments
are the elements that form a route. The region where a person is traversing is bounded
by
edges
. Indoors these edges can be the outside walls of the building and outdoors
they can be riv
ers for example. An interesting fact is that classifying some city object
as one element or another does not preclude it being classified as something else in
another context. To a pedestrian, a walking path is a route while highways and
railroads are edge
s. To a driver, the roads are the routes and everything else is an
edge. Furthermore, the mode of travel also effects what gets encoded, not just how it
gets encoded. [Darken, Peterson 2001.]


Route instructions
,
route information
-

or
route guiding
-
message
s are usually
combinations of multiple subroutes. A smallest scale instruction contains the
information of the two nodes and the arc that combines them. People may use
different methods, or mixture of methods, to find their way.


Landmark

In environment th
ere can be virtual cues that people easily perceive. The most salient
cues in any environment are
landmarks a.k.a. point of interest
. The route may make
use of landmark knowledge that is which information about the visual details of a
specific location [Da
rken, Sibert 1996]. In an ideal case a person who traverses a
route, just moves towards the desired landmark and when it is reached person sees the
next landmark and starts to move towards. This way he/she needs not to think about
the directions and the di
stances that are easy to forget or misunderstand. All that the
user needs to remember is the list of landmarks in the right order.


The role of landmarks in route guiding varies. In an ideal case the landmarks could
alone guide users to their destinations.

The guiding system would just give the proper
landmarks in the right order and all that users should do would be to go to the first
landmark and then see the next one and go to it until they have reached their
destination. [Darken, Peterson 2001].


One ot
her way to use landmarks in in
-
vehicle systems would be just to give them to
‘spice’ the instructions. The guiding would consist of directions and distances
including street names. Street names alone however are not suitable cues for
wayfinding, because th
e name signs are easily missed in crossings because of their
need for users full attention. This is why Streeter et al. used landmarks to indicate
places of turn. The landmarks were generic, for example public buildings that were
located on crossing area.

Also landmarks can be used on the route to assure the driver
is on right track. [Streeter et al. 1985].


5


Types of route guiding

All people have experiences of being lost in some public place. There are some
occasions where reliable route guiding might be
hard or impossible to get. There are
different route information needs for different situations and different user groups. In
the following few speech based route information system categories will be
introduced.


In
-
vehicle route guidance and information
systems

Electronic vehicle navigation systems aim to support drivers when travelling through
unfamiliar areas. Above and beyond these systems aim to help users to efficiently use
road network; avoid road sections that have heavy traffic load, offer alterna
tive routes
based on different needs like, etc. These systems are called IVRGIS (In
-
Vehicle
Route Guidance and Information System).


There are some elements that make the designing of IVRGIS differ from similar
systems designed for pedestrians. Car driver
’s safety plays a significant role while
designing the dialogue between the user and the system: Driver must stay focused on
the main task, driving, all times. So timing of the instructions is to be taken into
consider. Instructions coming too late may cau
se the driver to try some dangerous
manoeuvres. Some IVRGIS systems try to offer route instructions in small fragments
as the journey passes [Streeter et al. 1985]. This can be done because the IVRGIS may
be able to communicate with car’s speed meter, mile

meter and possible GPS
-
system
and thus stay aware of the location of the user.


One other matter concerning IVRGIS is the finality of the choices. Turning back and
stopping may be time consuming, dangerous. Turning from the wrong junction may
cause the u
ser to get even more lost.


Blind and visually impaired travellers

The people lacking visual sense have great difficulties on connecting their current
location to general geographic orientation. They are not able to use map to plan route
and find opti
mal route to the destination. When underway, they are not able to receive
cues on distant landmarks for maintaining goal orientation and on near landmarks for
checking if they are on route. Locally, they have also difficulties on keeping on
selected path a
nd detecting and avoiding hazards and obstacles. Because of these,
there is always a risk of disorientation and getting lost, which may cause stress and
fear, or even panic.
[Loomis et al., 1993.]


Blind and visually impaired people can use long
-
cane and s
eeing
-
eye dogs when
navigating in environment. They help with local aspects of navigation, such as
detecting stationary and moving obstacles, but not on remote aspects and geographic
orientation.
[Loomis et al., 1993.]


One way used to present spatial data

and route instructions to visually impaired
people is stereo sound. Sounds are presented through headphones, but they appear to

6

be within the auditory space of the user. One application, The personal guidance
system, using this technique will be presented

later on in this paper.


Office environment

Some public places may not offer any kind of route guiding for visitors. Such places
may be offices that do not have any receptionists or maps to tell visitors where their
target is located. It may be that the o
nly way to get some sort of guiding is to pop into
someone’s room and ask. Interfering other (unfamiliar) persons work is something
that people commonly want to avoid. Besides it has been shown that people often give
poor route instructions from spaces wel
l known to them [Streeter et al. 1985]. Office
facilities need a system that guides visitors to right target without the user needing to
wander around trying to find what he/she is looking for.


Contrary to an IVRGIS this sort of systems can grab user’s c
oncentration freely. This
way more complex messages can be conveyed with speech.


Distance vs. landmarks

With IVRGIS and indoor route information systems there seem to be two categories
for offering guiding with speech: With directions and distances, or
directions and
landmarks. Many IVRGIS developers [Krahmer et al.1998] and researchers [Streeter
et al. 1985; Jackson 1998; Burnett 1998] have done tests comparing these categories
and it would seem that better navigation results come with landmarks rather
than with
distances.


Presenting distance in metrical information is the intuitive way with vehicles, while
most of them contain some sort of milometer.

Researches have shown though
[Streeter et al. 1985] that people use topological information instead of
metrical, if
they don’t have any assistance meters. This is why the route instructions to
pedestrians should be made from a different point of view. Hunt and Waller (1999)
presented that when people give route instructions to each other, they tend to use
l
andmarks. To be able to decide whether some object is ‘qualified’ to be a landmark,
different factors affecting the process of comprehension have to be found and
evaluated. This matter will be discussed later in this paper.


There seem to be other fact
ors that cause landmark
-
based guiding to be more
desirable for many people. Burnett (1998) has tested car drivers with both types of
guiding. He found out that it takes a less time for a person to comprehend landmarks
than directions and distances.


Route

directions given by people are often ambiguous and incorrect. They rely on
using landmarks rather than distances when navigating and giving route guidance.
Anyhow the environmental cues referred by the person giving instructions are not
necessarily famili
ar or suitable for the person being instructed. The distances are
difficult to approximate, but in in
-
vehicle systems they can be used because of the
availability of distance information. Environmental elements such as traffic lights can
be used to indicat
e distances. However, people are not good in remembering number
of elements, so number of elements should be avoided as a navigational cue. Also the

7

notion of concepts, such as block may differ from person to person and cause
confusion. [Streeter et al, 19
85.]


Characteristics of speech in route guiding

Route guiding is good area for audio user interfaces, because of the user groups that
have special needs for the navigational assistance. For car drivers speech user
interface offers the possibility to relea
se the visual sense. This is important to
guarantee the safety of the driver.


One other user group that benefits the auditory interface is visually impaired people.
In addition to speech applications of this area often use features of sound to help the
us
er with orientation and wayfinding. Sounds are presented coming from a certain
direction in the stereo field.


Often happens that some sort of visual route information is presented in environment,
but the user does not discover it. Sound is spatially more

independent than vision. The
pervasive quality of sound can be used to make sure that user’s focus is captured. This
quality is often used in situations where the actual information is presented in some
other modality, but the attention has to be caught b
y a sound. A good example of
audio supporting visual modality is Microsoft Windows operating systems that use
different sounds to inform the user of different notifications of the system.


Speech and guiding

Streeter et al. (1985) have suggested, that the

vocal directions should include
information on the distance to the location where they need to turn, the direction they
need to turn to and what is the name of the street they are turning to. In in
-
vehicle
systems the user should be notified when the turn

is approaching so that the user has
time to prepare and react. Route instructions should also include information that
helps the user to determine if they are on right track, for example, landmarks on the
way, that are not necessarily needed for finding p
laces to turn.


Pros

Route information is often needed in motion, for example while driving. In motion
reading the map is difficult or even dangerous. Vocal directions enable the user to
drive and navigate at the same to, whereas visual maps demand visual

focus on map
[Streeter et al, 1985]. This is the case also in walking: voice instructions release the
visual sense to concentrate on observing the environment and finding navigational
cues. Users may have poor map reading skills, but well formed vocal dir
ections can
be more easily understood. [Streeter et al. 1985]


Streeter et al. (1985) compared giving route instructions by using taped spoken route
instructions, visual map and combination of these. It was shown that voice only
directions gave better resu
lts than visual maps and the combination of map and voice.
The test subjects, that had directions given with voice only made 70% less errors,
drove less and used less time to find their target in unfamiliar environment than those
who used visual map or bot
h voice and map.



8

This does not however mean that the vocal form alone is the solution for succeeding
guiding. The nature of the instructions has important role. The forming of instructions
needs to be carefully designed. The instructions used in the exper
iment were formed
based on previous studies on the area of navigation and psychology and iterative
design process was used to find suitable form. [Streeter et al. 1985.]


Cons

The disadvantage of using vocal directions is that they cannot give an overview
of the
route. The directions need to be memorized and therefore they need to be as simple as
possible. Repeating the directions and having space between repeats helps learning
and recalling. Also first and last element is best recalled. [Streeter et al, 19
85.]


Speech is sensitive for cultural differences. Place names in a written map are easier to
understand for a foreigner, than oral instructions. Speech understanding demands that
the foreigner has some skills in pronunciation of the language of the land

he/she is
visiting.


Current applications

In the following we present route guiding applications designed for vehicle drivers,
visually impaired people and visitors of public facilities. One other important
application group will be neglected: The mobile

device user. There are route
information systems made for different mobile devices, but their approach is quite
different than ours.


The Doorman

Many visitors come to our unit every day to participate in usability test or attend
meetings. Since our premi
ses are locked, the doorbell rings throughout the day
causing staff members to interrupt their work and go open the door. We have
developed the doorman system that lets visitors and staff member in and guides
visitors to their targets. In the future it wil
l also be able to convey spoken messages
from the user to staff members.


Doorman communicates with the user through speech dialogue. Outside the front door
system tries to find out the cause and target of the visit. If target is found from the
system’s da
tabase and it is recognised by system’s speech recognition module the
visitor is let in. There is also the possibility to bypass the system by pressing the
doorbell. System tries to recognise staff members from visitors speech inputs based on
language mode
ls and vocabulary.



In the future we are going to divide route instructions in several subinstructions
presented to the user as they advance in their route. Implementing more robots in the
facilities will do this with each giving instructions further on

the road. This way we
will be supporting the mobility of user in a fluent way with technology implemented
ubiquitously in the environment. Some barriers are still to be cleared. It is not that
easy for a system to stay conscious of user’s location withou
t him/her carrying any
device that could be used for locationing.


9


VODIS

VODIS is a European project, which aims at the development of a spoken interface
for a driver information system. This system uses LGM (Language Generation
Module).


For the first VO
DIS prototype, there are two sources of voice output: (i) the brief, pre
-
recorded route guidance messages (“turn left in 50 meters”), and (ii) the feedback
messages and prompts related to the ongoing dialogue between the user and the
system. The main purpo
se is to inform the user about the current state of the system,
thereby allowing the user to keep a good mental representation of the system.


VODIS gives the user the possibility to ask pre
-
trip route general descriptions of the
route additionally to inc
remental on
-
driving instructions. VODIS is also capable of
scaling instructions depending on are they inter
-
city or city
-
internal.


LGM was developed for the generation of spoken monologues on the basis of
structured, non
-
linguistic information. It aims a
t generating coherent monologues,
which allow a certain amount of variation. LGM is also capable of producing some
prosody to its monologues. This is done by evaluating the importance and meaning of
the words in the sentence it will present.


MoBIC system

MoBIC (Mobility of Blind and Elderly People Interacting with Computers) is a family
of orientational and navigational aids for blind and elderly people for route planning
and navigation. The aim is to increase independent mobility of visually disabled
peo
ple travelling in unknown places. The problems the visually impaired and old
people have is finding and accessing relevant information for planning the route. In
the planning stage visually impaired users need more detailed information on the
route and the

environment, because they cannot rely perceptual information gathered
in the situation. During the travel, they need support in knowing their current location
and orientation and keeping on the route.
[Strothotte et al. 1995.]


MoBIC is divided to two par
ts. MoPS (MoBIC Pre
-
Journey System) helps the user in
planning the route and building mental model of area. It provides timetables of public
transport and digital maps to support planning and studying route in advance in
indoors. It also contains tools for

example, route
-
calculation algorithms considering
ease of travel. MoODS (MoBIC Outdoor System) helps to execute plans by providing
users with orientation and navigation assistance during journey. It guides on route and
provides relevant additional informa
tion on the environment and route.
[Strothotte et
al. 1995.]


MoBIC is operated with notebook computer, GPS and electronic compass. Input is
given through hand
-
held keyboard, which has buttons for preformed questions such as
‘where am I’, ‘what is near me’

and so on. Output is given with synthesised speech
through open headphones. Route instructions were based on both distances and
landmarks. Clock metaphor was used to give instructions on directions.
[Strothotte et
al. 1995.]


10


MoBIC is secondary aid, it
is used to support the use of long
-
cane and guide dog
[Strothotte et al. 1995].

User Requirements

The Design of the system focuses on user requirements. Blind people participated
both specification and implementation. In designing stage it was investigate
d how
blind people find their way in complex environment, what problems they experience
in orientation and mobility and what info should be presented to solve experienced
problems. User requirements were gathered by analysing previous researches and by
int
erviewing visually impaired and elderly people and mobility instructors. Mock
-
ups
were tested by potential users in realistic situations.


Visually impaired users wanted to get information on locations of public transport,
useful landmarks and services, ro
ad works, obstacles on the ground and on head
height and suitable places to cross streets. Problem of using headphones to present
information to the users is that they block environmental sounds and possible
navigational cues. Users also wished that the a
ppearance of device would not look
conspicuous.



Personal Guidance System

The personal guidance system [Loomis et al, 1993] is navigation aid for visually
impaired people travelling independently. It assists the user on route planning before
travelling an
d on the way it provides information about the immediate surroundings to
support orientation and keeping on right track.


The Personal Guidance System indicates the positions of landmarks to blind travellers
by using speech and virtual acoustic display. T
he speech synthesis is used to present
labels and attributes of landmarks. Spatial sound indicates the correct location of the
landmark. The directional sound is created with binaural sound signal output through
stereo headphones.


System uses different
ial GPS (Global Positioning System) for positioning and getting
orientation the user and GIS (Geographic Information system) for mapping landmarks
and their information to the environment of the user. GIS is a database containing
both spatial and semantic
information on environment, functions and characteristics of
landmarks, for example locations, names and attributes of objects, such as buildings,
walkways and large obstacles on the road. GIS provides information supporting route
planning and can be searc
hed for environmental information.


The Personal Guidance System is used together with long
-
cane and guide dog. There
might be risk, that if the user becomes dependent on the aid, other navigational skills
might be forgotten and this could be cause naviga
tional impairment. Also, technical
aids like Personal Guidance System can encounter errors or they can work incorrectly,
which could be harmful in some situations.

Route guiding in Personal Guiding System

Route guiding can be done in two ways. When using

directional sound, the route is
presented by succession of virtual beacons. The user is guided towards one location,

11

which is presented with sound. After reaching the location, next locations is
presented. Sounds are presented through headphones, but they

appear to be within the
auditory space of the user. Other way is to use natural
-
language commands in route
guiding.
[Loomis et al., 1993]

User interface of the Personal Guidance Device

Personal Guidance system has two alternative display methods. The rou
te can be
presented conventionally by giving route instructions and describing surroundings
only in spoken form through headphones. The other possibility is to use virtual
acoustic display to give navigational cues on the direction of the current destinati
on
through binaural headphones.
[Loomis et al., 1993.]


Personal Guidance system is used with handheld keypad, which enables user to select
landmarks and destination, add landmarks to database, change display mode between
beacon based and spoken route ins
tructions and change display parameters. Also
speech input can be used with limited vocabulary.
[Loomis et al., 1993.]


Constructing the route information

To be able to construct proper route instructions, we have to know what information is
significant. T
his is the reason why we have to define the basic factors of spatial
orientation. The three major dimensions of spatial ability that are commonly
addressed are spatial orientation, spatial visualization and spatial relations. Spatial
orientation involves t
he ability to mentally move or transform stimuli, while retaining
their relationships. Spatial orientation also involves the mental manipulation of an
object using oneself for reference. Spatial visualization goes further, in that the person
can manipulate

the relationships within an object. The third dimension, spatial
relations, consists of the ability to imagine how an object will align from different
perspectives. [Satalich 1995.]


In the following we will discuss the accuracy of the route instructions
and try to find
the scalable elements in them. Last we try to find the factors in landmarks that help to
decide whether they are usable or not.


Accuracy

When constructing the route information, the accuracy of the guiding is an important
issue. Some res
earchers and system developers have introduced rules by which the
guiding message should be constructed.


Distance is one factor that creates a need to scale instructions. IVRGIS users need
different sort of information when they want to find a place coupl
e of blocks away
(city
-
internal
), than when they need instructions on how to get to another city (
inter
-
city
).
Krahmer et al.
(1998) found out when doing research for an IVRGIS that users
might also want more general route instructions instead of specific
instructions.
Krahmer et al. Proposed that inter
-
city route descriptions should contain following
information:


1.

Current location; city and street, plus crossing street (if entered)

2.

Distance in kilometres from current position to destination


12

3.

The expected t
ime of arrival

4.

Time in minutes until the next incremental guidance message

5.

A general route description, consisting of a sequence of subroutes, each in turn
consisting of the names of the motorways visited during the route, together
with begin
-

and end
-
poin
ts, distance between them in kilometres, and (if
available) points of interest, which mark the switch to a different motorway


For city
-
internal routes only items 1 to 4 are relevant.


Streeter et al. (1985) also made direction rules that present what info
rmation should
be presented and in which order. They also thought about two things: Error handling
and repeating to aid the memory.


1.

Critical direction:

This message contains the information of how long the
driver has to go one before turning, which way to

turn when necessary and to
which road.
Continue instruction:
These instructions tell the user the new
name of the road driven, if the road changes to another, the distance to the next
turn, the direction of the turn and the name of that road.

2.

When to turn

instruction:
If there is a known landmark available near the next
turning point, the following will be prompted: {Landmark} is on the {left
(corner)|right (corner)|straight ahead}. If there is no suitable landmark, the turn
will be defined by the road on

the same side of the road before the road to be
turned to: {street_before_name} is the street before {street_name}.

3.

Too far instruction:
These instruction help the users to identify when they are
lost. If you come to {landmark|major_street} you’ve gone to
o far.

4.

Summary instruction:
This instruction is the repeating of the first one.
Remember you it’s {X, X} miles to your {left|right} turn onto {street_name}.


Scaling distances

As discussed above in the accuracy paragraph the route instructions should be ta
ilored
to each and every occasion. Scaling is needed to guarantee that only necessary
information is given to the users, since their ability to comprehend and remember long
instructions have been shown to be poor.


The main factor to be scaled in guiding
prompts is the distance. In the accuracy
paragraph the first set of the description rules by Krahmer et al. (1998) left general
instructions out from city
-
internal route descriptions.


Scaling directions

In the Doorman system the directions are scaled. Th
e algorithm constructs two lines:
One from the point of origin to the centre of current place. The other is from the point
of origin to the centre of place next to be traversed. If this angle is less than 20
degrees the direction will be left out from the
prompt.


One other way to leave directions out of the route description or at least reduce the
amount of direction expressions, is to rely on landmarks so completely that users are
guided from a landmark to another without any directional aid [Darken, Pete
rson

13

2001]. This could be useful only in some special environment, but in most
environments, there are not enough reliable landmarks to enable neglecting the
directional elements from the route descriptions.


Evaluating landmarks

Earlier in this paper wa
s mentioned that landmarks are the most salient cues in an
environment. What are the factors that cause some object to pop up in person’s field
of vision? Defining these factors is the first step, if we want to decide whether some
object is qualified as a
landmark. Evaluating these factors is an important task in
forming a structured and generic procedure to select suitable landmarks. Burnett
(1998) has separated 11 attributes that define a ‘good’ landmark.


1.

Permanence:

The likelihood of the landmark being
present.

2.

Visibility:
Whether the landmark's size/shape means it can be seen clearly in
all conditions.

3.

Conspicuity:
The attention
-
grabbing nature of the landmark.

4.

Predictability in location:

Knowing where to look for the landmark.

5.

Openness:
The likelihood

of the landmark being obscured by other objects etc.

6.

Familiarity:
Whether the landmark is well known.

7.

Predictability in appearance:
Knowing what the landmark will look like.

8.

Uniqueness:
Whether the appearance of the landmark is such that it is unlikely
to

be mistaken for anything else.

9.

Degree of separation:
The extent to which examples of the landmark are ‘far
apart’.

10.

Usefulness of location:
Whether the landmark is located close to navigational
decision points.

11.

Compactness:
Whether the size of the landmark

allows it to be easily related
to a turning.


One thing left out the discussion here is the direction of approach affects the visibility
and comprehension of the landmark. For example a certain monument or building can
look quite different whether the per
son closes it from the front
-

or backside. This
factor should also be taken into count when measuring a landmark.


As we can see Burnett considered the users to be quite homogenous group. User’s
social and psychological background has a great effect on wh
at he/she see as large,
small, close or distant etc. Cultural background may effect on what the user expects to
see in some environment. Landmarks like traffic signs may be located differently in
different countries. This may cause route description prompt
s to be inadequate to a
foreign user even if he/she is familiar with the language.


Conclusion

During our study we have found out that quite different approach have to be taken
whether the route guidance system is designed for blind users, pedestrians in
common
or in
-
vehicle drivers. We have found out that time and system environment factors
differ indoor guiding from in
-
vehicle systems. For blind persons the description of the
environment is useless, while it can be significant to others. We have also bec
ome

14

aware of the fact that people are often poor giving route instructions concerning
environment known to them.


For future work it would be beneficial to orientate to landmarks and their features,
since they seem to be beneficial according to various res
earches discussed in this
paper.


One important thing we have encountered is the usefulness of the sound in route
guiding in some situations. The nature of audio releases user’s cognitive resources on
one hand, and on the other hand its pervasive nature g
ives the opportunity to give the
user the important information in the right moment.


It would be beneficial to try to find some basic principles by which route information
given by a speech user interface system could be developed. There would also seem t
o
be a need for tool for evaluating the reliability of landmarks.



References


[
Burnett 1998
] Burnett, G.E. (1998). "Turn Right at the King's Head". Drivers'
Requirements for Route Guidance Information. [PhD Thesis], Loughborough
University.


[Darken, Pet
erson 2001
] Darken, R. P., & Peterson P. (2001). Spatial Orientation,
Wayfinding, and Representation. Handbook of Virtual Environment Technology.
Stanney, K Ed.


[
Darken, Sibert 1996
] Darken, R.P., & Sibert, J.L. (1996). Navigating in Large
Virtual Worlds.

The International Journal of Human
-
Computer Interaction, 8(1], pp.
49
-
72.


[
Hunt, Waller 1999
] Hunt, E., & Waller, D. (1999). Orientation and wayfinding: A
review (ONR technical report N00014
-
96
-
0380]. Arlington, VA: Office of Naval
Research


[
Jackson 199
8
] Jackson, P.G. (1998). In search of better route guidance instructions.
Ergonomics, Vol. 41, 7, 1000
-
1013


[
Krahmer et al.1998
] Krahmer, E., Landsbergen, J., & Odijk, J. (1998).
A Guided
Tour Thro
ugh LGM. How to Generate Spoken Route Descriptions
. IPO Report No,
1182.


[
Loomis et al. 1993
] Loomis, J. M., Golledge, R. G., & Klatzky, R. L. (1993)
Personal guidance system for the visually impaired using GPS, GIS, and VR
technologies.
Proceedings of
the Conference on Virtual Reality and Persons with
Disabilities
, June 17
-
18, 1993, Millbrae, CA.


[
Satalich 1995
] Satalich, G. (1995). Navigation And Wayfinding In Virtual Reality:
Finding The Proper Tools And Cues To Enhance Navigational Awareness. Master
's
Thesis, University of Washington, Seattle.


15


[
Streeter et al. 1985
] Streeter, L.A., Vitello, D., & Wonsiewicz, S.A. (1985). How to
tell people where to go: Comparing navigational aids. International Journal of Man
-
Machine Studies, 22, 549
-
562.


[
Strothot
te et al. 1995
] Strothotte, T., Petrie, H., Johnson V., Reichert L. (1995).
MoBIC: user needs and preliminary design for a mobility aid for blind and elderly
travellers.