Transcript Document
User responses to GPS
positioning information on a
2D digital map
Will Seager,
Department of Computer Science,
University College London, UK
Danae Stanton Fraser,
Department of Psychology
University of Bath, UK
Equator IRC
Talk overview
• Describe the rationale, method, & results from a
field experiment which examined user
responses to GPS positioning information on a
2D plan-view digital map + draw some design
lessons from the study
2D plan-view maps + GPS
• Common format used in some
research prototypes and most
commercial products
• Typically, the user is free to
manipulate map using pan and
zoom controls
• Many incorporate positioning
information (in outdoor contexts,
GPS)
• GPS is displayed as an icon
overlaying the map; tracks the
user’s changing position
2D plan-view maps + GPS
• In theory, positional information
should simplify the cognitive work
required to “self-locate”
– With traditional maps, users
self-locate by matching map &
environment landmarks
• Do this at the beginning, at the
end, when lost, and on route to
guide re-orientation decisions
& for reassurance
– Positioning information on
mobile 2D digital maps should
simplify the cognitive task:
users simply need to perceive
the position of the icon on the
map.
2D plan-view maps + GPS
• However, GPS is notoriously
inaccurate
– Factors such as tall
buildings and atmospheric
conditions can increase the
amount of error
=>Is positioning information
supplied by GPS too
inaccurate to be useful?
=>Could the inaccuracy
confuse users?
Field experiment
• Carried out a field experiment
• 24 participants were asked to navigate using a
map-based application with and without GPS
positioning information
• Does positioning information lead to more effective
navigation with respect to time taken and number
of wrong turns?
• What are participant’s subjective impressions of
the information?
Map application
• Tomtom Citymaps
software
• Main controls…
– Pan by dragging stylus
across screen
– Change scale using a
scroll bar
• Application installed on
an iPAQ 3630 handheld
computer
Positioning information
• Positioning information
supplied by a Garmin
eTREX GPS unit
• Displayed as either…
– A crosshair
– An arrow
– When “no fix”, last known
estimate displayed as a clear
circle
• Map matching algorithm
– Placed symbol on part of the
street network nearest
estimate
Positioning information
• Positioning information
supplied by a Garmin
eTREX GPS unit
• Displayed as either…
– A crosshair
– An arrow
– When “no fix”, last known
estimate displayed as a clear
circle
• Map matching algorithm
– Placed symbol on part of the
street network nearest
estimate
Positioning information
• Positioning information
supplied by a Garmin
eTREX GPS unit
• Displayed as either…
– A crosshair
– An arrow
– When “no fix”, last known
estimate displayed as a clear
circle
• Map matching algorithm
– Placed symbol on part of the
street network nearest
estimate
Automatic centring on GPS
• Application included an
automatic scrolling feature
– Kept the map centred on the
GPS icon
• In a pilot study, participants
found this feature confusing
– The map moved when they
weren’t looking
– Interfered with how they
wanted to position the map
Automatic centring on GPS
• Interference with map
positioning,,,
Participants generally
wanted to position the map
so that two route markers
were visible (e.g. the two
waypoints)
Automatic centring on GPS
• Interference with map
positioning,,,
With automatic centring on
GPS, the map would
“snapback” pushing one of
the markers off-screen
Automatic centring on GPS
• Interference with map
positioning,,,
To get both landmarks back
on screen they had to zoom
out to a lower scale
Meant that map detail was
harder to see
Automatic centring on GPS
• Interference with map
positioning,,,
To get both landmarks back
on screen they had to zoom
out to a lower scale
Meant that map detail was
harder to see
=> Automatic centring on GPS
switched off
Experimental design
• Within-subjects design: each participant experienced
all conditions
• Conditions
– Map software only
– Map software + GPS positional information
– Baseline paper map condition
• Three routes
• Conditions and routes counterbalanced
– Experienced three routes in the same order but order of
conditions varied (should control for route and order effects)
• Dependents: timings & errors
Participants & routes
•
•
•
•
24 participants (12 male and 12 female)
Unfamiliar with the area surrounding the routes
Routes in an urban area; no “urban canyons”
Each route…
– Approximately 1 km in length
– Included two waypoints
– Successful navigation required 10 correct decisions
• Routes indicated on the map using a markers for the
starting position, the two waypoints and the
destination
Procedure
• Practice using zoom and pan controls
• Shown symbols used for positioning information
(crosshair, arrow & circle); told information could be
inaccurate (lag & arrow direction)
• Orientated (shown current position & how the map
aligned with the environment)
• Task: to navigate to the destination via the two
waypoints, taking the shortest possible route
• Timings; errors counted
• Unstructured observations; comments on route;
semi-structured interview
Timings
18.00
Timings (mins) s
16.00
14.00
12.00
10.00
8.00
6.00
4.00
2.00
0.00
pda
gps
Navigation aid
paper
• Main effect for timings
• Post-hoc analyses
revealed two significant
contrasts: paper map
versus other two
conditions
• So no evidence that
positional information
helped participants
navigate more quickly
Errors
2
s
1.6
Number of errors
1.8
1.4
1.2
1
0.8
0.6
0.4
0.2
0
pda
gps
Navigation aid
paper
• Main effect for errors
• Post-hoc analyses
revealed two significant
contrasts: paper map
versus other two
conditions
• So no evidence that
positional information
helped participants
navigate with fewer
errors
Qualitative results
• Subjective impression of redundancy:
• A large proportion of the participants said that the
positional information provided by GPS was too
inaccurate to be helpful.
“because it lagged, I didn’t bother with it at all”
“virtually no notice of the satellite thing. It’s too far behind. It’s
nowhere near where I am so there’s no point looking at it”
“not bothered about not having GPS…because it lags, it was
not really useful...you are already half way down the road
before it catches up with you”
Qualitative results
• Clear that participants were still using traditional
methods to self-locate e.g. comparing map and
environment street names
• Said it did not help to make navigation decisions or
even provide reassurance
• Could see the value of positioning information in
principle but said it would need to be much more
accurate (e.g. to within 5 meters)
Qualitative results
Inaccuracy led to confusion…
• Lag => participants thought they were further back
than they really were => missed turns.
• Failure of the map matching algorithm to place the
position symbol on the correct street
• Arrow misrepresenting actual heading i.e. pointing
the wrong way
Confusion despite warnings…
“predisposed to thinking that technology is correct in spite
of any warning that it might not be”
Qualitative results
An unexpected benefit of positioning information…
• Served as a useful “anchor point” when panning or
zooming e.g. users could pan forwards and then relocated their position more easily
• With map software only, users often got “lost on the
map” i.e. when panning or changing scale, they lost
track of their position
(with paper maps, the current location page has a
stable position with reference to the rectangular
frame of the page; with atlases, it also has a stable
page position)
Qualitative results
Participants noted
contexts where GPS
might help…
• In urban areas at the start of
a journey
• In urban areas when lost
• In rural, mountainous or
other areas where there
were no street signs
=> Helps find general location
“Good for mountains”
Conclusions
• Conclusion from pilot study: automatic centring over GPS icon
confuses users;
– the map moves when not looking
– it interferes with how they want to position the map
• GPS positional information does not lead to significant gains
with respect to improved journey timings or taking wrong turns
• Strong subjective impression of redundancy; too inaccurate to
be useful
• Confusion (lag, failure of map matching algorithm and arrow)
• Position icons help users to anchor their position when zooming
or panning
Overall conclusion:
GPS positioning information helps users to find their general
location but it’s not good enough during the navigational task
itself to replace traditional self-location methods (i.e.
map/environment landmark matching)
Re-design?
•
•
Dispense with the icon?
Only use GPS to position the
map (but only in certain
situations given problem with
automatic centring on GPS)?
E.g. The user switches the map
on and it automatically presents
the area of map surrounding
his/her current position
When lost, the user can press a
button and it automatically
centres on his/her position
Re-design?
•
•
Embody the level of accuracy
in the icon?
A wider transparent icon could
convey the level of accuracy
– Constant size: reflects the
average level accuracy of the
GPS unit in that particular
region
– Or the size varies in response
to the number of satellites,
atmospheric conditions,
urban/rural etc
•
•
Aids user to find general
location (e.g. journey start,
when lost, where no street
names) but less likely to
confuse
Still helps to anchor position
when zooming, panning or
looking away for long periods
Re-design?
•
•
Embody the level of accuracy
in the icon?
A wider transparent icon could
convey the level of accuracy
– Constant size: reflects the
average level accuracy of the
GPS unit in that particular
region
– Or the size varies in response
to the number of satellites,
atmospheric conditions,
urban/rural etc
•
•
Aids user to find general
location (e.g. journey start,
when lost, where no street
names) but less likely to
confuse
Still helps to anchor position
when zooming, panning or
looking away for long periods
Re-design?
•
•
Embody the level of accuracy
in the icon?
A wider transparent icon could
convey the level of accuracy
– Constant size: reflects the
average level accuracy of the
GPS unit in that particular
region
– Or the size varies in response
to the number of satellites,
atmospheric conditions,
urban/rural etc
•
•
Aids user to find general
location (e.g. journey start,
when lost, where no street
names) but less likely to
confuse
Still helps to anchor position
when zooming, panning or
looking away for long periods