* Show IT Crowd Clip
wedge Place-Lab summary
wedge Implement location a location system which is
* ubiqutious
* cheap
* listening for beacons
* referencing their position
* in a cached database
wedge Why current location systems don't work?
* They don't work where people spend their time
wedge systems only work
* outdoors
* in a particular building
* on a particular campus with unique infrastructure
wedge Existing systems
wedge High cost to entry for users and developers
* Expensive infrastructure
* Exhaustive calibration
* Special tags
* Special beacons
* Special sensors
* Privacy
wedge Cycle of failure
* Users need location aware applications
* Location aware applications need location sensing
* Location sensing isn't built because there are no users
wedge Place Lab goals
* coverage
* low barrier to entry
* Privacy
wedge Secondary
* precision
wedge Related Work
wedge Compare the model of place lab to the model of GPS
* GPS meets low barrier of entry, but not coverage
wedge Compare the model of place lab to the model of cell phones
* In terms of Place Lab coverage model
* driven by E-911 and E-112
* cell phones meet coverage, but not low cost of entry
wedge Fingerprint beacon based solutions
* RADAR and Ekahau
wedge Don't work out of the box
* high cost of entry
* But, more accurate than Place Lab
wedge Lots of indoor systems
* optimized for accuracy
* use ultrasonic/infrared/ultra-wideband radio
* need infrastructure support
* limited coverage rules out many personal and social applications
wedge Place Lab Architecture
wedge Beacons
wedge Listening for "radio beacons"
* unique ID on each beacon
* 802.11
* bluetooth
* GSM
wedge Interaction with beacons
* theoretically passive
* in practice 802.11 is passive
* bluetooth API requires a MAC revealing scan to find beacons
* GSM API requires associating with a tower to get Cell ID
wedge Databases
* Place Lab is agnostic about databases
* from administrators
wedge from wardriving
* explain war driving
* network access
* show graphic
* discuss uncertainty in war-driving
wedge Clients
* Spotters
* Mappers
wedge Trackers
* simple average
* more elaborate techniques
wedge Privacy
* Passive system
* Cached Database
wedge issues
* downloading database reveals location
* war-drivers reveal their location
wedge beacon owners might not want to reveal their location
* hashed database
wedge Implementation
* Show slide
* Direct Linking to libraries
* Daemon
* Web proxy
* JSR 179
* NMEA 0183
wedge Accuracy
wedge Hard to characterize
wedge Better in urban areas
* beacons cover population
* Depends on beacon mix
* Experiment