Panda Poster

1
The Smart Party
A Personalized Location-Aware Multimedia
Experience
V. Ramakrishna
Kevin Eustice
Nam Nguyen
Dr. Leonard Kleinrock
Dr. Peter Reiher
A ubicomp application to dynamically customize
the music played at a party to the tastes of the
party-goers
Dynamic Playlists and Music Transfers
Device Configuration Collaboration
Party guests carry mobile devices with songs and
preferences

• Iterative Context Refinement
• Mapping observed network to party location
• Trigger appropriate application that runs on
  guests mobile devices

• Each Room
• Maintains a dynamic playlist
• Plays music that represents the collective
  preference of its current occupants

• Digital Credentials
• Host provides vouchers to guests as invitations
  prior to the party

I can see the Party environment!

• Goal Maximize user satisfaction through
  collaboration
• Criteria for music selection song,
  artist, genre, album, period

• Device activates appropriate voucher based on the
  detected environment
• Voucher indicates which network to join

VOUCHER
Looks like theres a Party going on!
Found a Party voucher!

• Dynamic Collaboration
• Negotiation between guest device and server
  allows user into the party
• Deliver map of to guest devices
• Switch rooms based on the map as guest moves
  around
• Collaborates with room servers and other devices
  for musical experience

• Protocol
• Room requests song suggestions from users
• Room makes selection requests song delivery
• User device possessing that song delivers it to
  the playback device
• Songs are cached for efficient bandwidth use

OUTSIDE THE PARTY ENVIRONMENT
Ive joined the Party!
INSIDE THE PARTY ENVIRONMENT
SONG PLAYLIST
Delivering Map of Party Locations
SONG PLAYLIST
ROCK
ROCK
JAZZ
HIP-HOP
I cant stand the music here! The dining room is
better suited to my tastes.

• Semantic Maps
• Associate arbitrarily sized semantic regions with
  observed wireless networks
• Semantic regions Family, Living and Dining Rooms

• Rich User Experience
• Intersection of social group and location contexts

HIP-HOP
JAZZ
LIVING ROOM
FAMILY ROOM

• Dynamic Interactions
• Showcases different models cooperation,
  competition, even active opposition

• Localization Scheme
• A combination of scene analysis and attenuation
  monitoring

ROCK

• Reconciling Guest Preferences and Selecting Songs
• Basic round-robin selection
• Dynamic user voting
• Leader-based
• Democratic model

JAZZ

• Observed signal strengths from nearby 802.11
  wireless access points are characteristic to a
  semantic region

JAZZ

• Need to build map through observations prior to
  inference

SONG PLAYLIST
DINING ROOM

• Important Metrics of Success
• Fairness give everyone an equal chance to
  suggest a song of his choice
• Satisfaction provide better average user
  experience over a period of time
• Occasional tradeoff required between these metrics

• Advantages of Localization Technology
• Requires no deployment of custom hardware
• Can distinguish semantic regions on the basis of
  altitude, unlike GPS

Inference of Location Context
Socially Aware Technology
The Panoply Middleware

• Functions
• Secure sphere management discovery, formation
  and reconfiguration
• Scoping event interest and dissemination
• Providing primitives for sphere collaboration
• Dynamic management of policies and system state
• Inter-sphere interaction and dynamic access
  control through negotiation
• Management of a database of vouchers electronic
  credentials that certify arbitrary properties of
  a sphere
• Provide primitives for the design of social- and
  location-sensitive applications like the Smart
  Party

• Spheres of Influence
• Automatically discovers and organizes devices
  into groups, or spheres, based on social groups,
  physical location or proximity, task, network
• Manages network connectivity, service discovery,
  location and social context

Other Supported Location-Sensitive Social
Applications
Our Vision

• Produce better user experience by making mobile
  consumer devices cooperate when in proximity
• Bring people together dont keep them apart
• Leverage preferences and desires of individual
  users
• Use groups and Panoply for organization
  management
• Mobile ubiquitous environment is complex and
  unpredictable, requiring better tools
• That allow simplifying generalizations, like
  spheres of influence
• For debugging and evaluation of complex
  cooperation and competition scenarios

Location-Driven Interactive Fiction
Intelligent Museum Gallery
Customized Location and Group-Aware Experience
for Individual Visitors and Tour Groups
Team-driven Non-Linear Narrative
Dynamically Affected by User
Actions at Designated Locations
This research was supported by the National
Science Foundation (NSF) Grant CNS 0427748