Upload
aleesha-reeves
View
217
Download
0
Embed Size (px)
Citation preview
User-Driven Indoor Visibility Localization with Wayfinding
Aleksandar Kuzmanovic
http://networks.cs.northwestern.edu
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding2
Motivation - Indoor localization
GPS works outside, not inside
What if you are lost indoors?
Food Mart
?
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding3
Triangulation – Delay, Angle, RSSI
RF Signatures– Beacons– Impulse responseInfrastructure
Manual Labor
RF-derived approaches
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding4
Some challenges with RF
Food Mart
RF Here RF Here
Not the same!
$$$ - No Thanks
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding5
Avoid these obstacles
Leave the building owner out of it– Don’t install infrastructure– Don’t ask for detailed schematics
Cancel site surveys– Don’t build detailed RF maps manually
Use something simple that works– Don’t rely on signal processing
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding6
Observations
Many facilities publish free floor plans online
Use pre-computation to extract information People are good noise-filters
Hyatt Regency O’hare
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding7
Strategy: Relax constraints, use logic
Let users report what they see
Take advantage of relationships among identifiable features of a room
Absolute precision in x cm is not as important as a person comprehending where he is
OlivesOlives
Self-Serve CounterWine
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding8
Important Definitions
Isovist: The visible area from a location’s perspective – Vi,j = {…}, set of coordinates visible from point (i,j)– V’i,j = {…}, set of coordinates invisible from point (i,j)
Feature: An identifiable landmark, e.g. cash register, bathroom, elevator…Feature Vector: [f1 f2…fh], where fi in {0,1}
– fi == 1, fi is visible– fi == 0, fi is invisible
Region: Subset of coordinates sharing an identical feature vector
Isovists G, B, and R
G
B
R
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding9
System modeling
Feature: An identifiable landmark, e.g. cash register, bathroom, elevator…Region: Subset of coordinates sharing an identical feature vector
2h possible vectors, so locatable regions grow in O(2h)– Corollary: Average region size decreases as h increases
AB
C
D
AB
C
D
AB
C
D
AB
C
D
AB
C
D
AB
C
DA’B’B’
B’D’
D’
B’C’D’B’D’
C’C’D’
A’C’
A’A’C’
C’C’D’
D’
A’B’
A’B’C’ A’C’
User: I see A and B, but not C or D[A B C’ D’]
You are Here
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding10
A simple example: 3 features
If feature A in Vi,j, then point (i,j) in VA
– i.e. if you can see a feature, then that feature can see you.
In general, for all features fp reported visible, and all fq reported invisible:– Location
– 2h (or 2h-1) locatable regions, for h total features in the environment
A
BC
A
B’
BC
B’A’
A’
C’
B’C’
A’B’
C’
A’C’
User: I see A and B, but not C[A B C’]
You are Here
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding11
User error
Mistakes happen– Type I Error: Report an invisible feature is visible
• Possible, but rare• Example: confusing “stairs” with “escalator.”
– Type II Error: Report a visible feature is invisible• Not only possible, it’s probable• Our study revealed ~50% hit-rate on noticing features
Assume positive sightings are trustworthy, and negative sightings are completely untrustworthy– Sacrifice all info gained from unsighted features– All information comes from accumulation of positive sightings
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding12
Avg. located area vs. h features
Important result– The plausible range of operation (2:5 sightings) performs in line with
the unlimited range of operation
Exponential & Perfect
Accumulative full range
Accumulative up to 5 reports
Accumulative 2 to 5 reports
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding13
System in real-life
Internet
Food Mart
?
Company Website
NU Networks Group
What do you see? Check all that apply
Cash RegisterExitEscalator
Submit
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding14
Finding your way For some, learning location is enoughFor others, the real goal is a destination– Where is the {dairy section, shoe department, light bulbs}?
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding15
The processModel environment as a network
– nodes (features) and – links (intervisibilities)
At each hop, report what you see– app recommends a new next hop
Only adjacent subset of features are offered at each hop
– Makes the list much smaller
If you don’t sight a feature, the link is down
– Ineligible as a next hop
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding16
User/App interaction
Based on your input to Part I– I know where you are– More important –
I know what you see
Knowing what you see, I can tell you to walk over to it– Our application picks the best
“next hop” on the way to the destination
– Repeat this in a simple way until the user is within L.O.S. of destination
Application on HTC Android G1
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding17
How this is different
Traditional turn-by-turn– Google Maps, GPS, etc.
• “Continue two blocks, turn Right at Main St.” • “Walk down the hallway, turn Right at the Bookcase”
– User must understand semantic instructions• No assurance whether user sees/comprehends• A wrong turn can go unnoticed for a long time• Some people just don’t like thinking this way
Our approach uses collisions to navigate – No wrong turns, because each next-hop has already been
affirmatively identified by the user– User input is a simple Yes/No, rather than semantics (“I see a
red house”) requiring artificial intelligence.
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding18
Field Study
Large Hotel/Convention Center (with permission)
– h = 38 features– p = .283 edge density
10 volunteers with no prior knowledge
Test 1 – sighting features– Each subject tested from multiple
vantage points– Positive sightings rate
pv = 0.496, with 90% confidence pv > 0.46
Test 2 – usability– Wayfinding task A B– Tracked user experience
• Willingness to use the tool• Ability to use the tool • Feedback & suggestions
Aleksandar Kuzmanovic User-Driven Indoor Visibility Localization with Wayfinding19
Summary
Users are willing and able to interact with the tool– Requires very little training– Some users eager to find features, like a game– Some users gave cursory effort– The best next-hop sometimes does not feel right, and users
want to override, i.e. take the 2nd best option.
The overall way finding success rate > 95%
Key modeling result– p*pv > (ln h)/h gives good success rate