Location Based Services On the Road to Context-Aware Systems

Transcription

1 University of Stuttgart Institute of Parallel and Distributed Systems () Universitätsstraße 38 D Stuttgart Location Based Services On the Road to Context-Aware Systems Kurt Rothermel June 2, 2004

2 Overview Technology trends: Sensors Context-aware applications Scenario and definition Examples DFG Sonderforschungsbereich Nexus Context-aware communication Summary

3 Trends: Sensors Multitude of different sensors mechanical: position, velocity, acceleration symbolic: identification, localization thermal, optical, acoustic, chemical,... GPS receiver (RoyalTek) Miniaturization and price decline Example: positioning systems outdoor: GPS: 25m, DGPS: < 10 m, Galileo (2008): < 4m, indoor: ultrasound-based (e.g. Intersense): < 1 cm, WLAN-based (e.g. Ekahau): < 5 m

4 Trends: Sensors Multitude of different sensors mechanical: position, velocity, acceleration symbolic: identification, localization thermal, optical, acoustic, chemical,... Miniaturization and price decline Example: pressure sensor Siemens membrane: 120 µm (human hair: 100 µm) membrane Fraunhofer Institut

5 Trends: Sensor Platforms Embedded systems Integration into everyday objects: smart things Capture, process, communicate context of things Example: Smart-It Sensor networks TecO Smart-It 15x40 mm Cheap nodes at size of grain of sand Nodes aggregate and communicate sensor data cooperatively Examples: Smart Dust, ARTMIS (architecture, communication protocols, information management algorithms,...) Source: UC Berkeley

6 Trends: Conclusion Technology trends Miniaturization Price decline Integration: smart things, sensor networks Proliferation of sensor systems Context-aware systems Position, direction, velocity Identity Objects in vicinity Temperature,... real world digital world

7 Scenario: John Visits a Fair When John enters a booth, information about the exhibits is automatically displayed on his PDA. John leaves a virtual note for his colleagues at this booth indicating that he already made contact. John made an appointment with Carol in Hall 8 between 10 and 11 o clock. The system reserves a meeting room and navigates both to the room when both have entered Hall 8. The meeting room is equipped with a big screen. When John and Carol enter the room, their PDAs output is automatically routed to this screen. Messe Stuttgart International

8 Definition: Context-awareness Context is any information that can be used to characterize the situation of an entity, where an entity is a person, place, or object that is considered relevant to the application. Example: Fair information system Entities: Visitors, meeting rooms, halls, Context of a visitor: Identity, location, other persons in vicinity,... An application is context-aware if its behavior depends on context information.

9 Location-based Services: Today Traffic information and navigation services traffic information captured by traffic sensors along highways, Floating Car Data location-based access to traffic information examples: Passo (Vodafone), ADAC traffic information,... navigation based on current traffic situation: example: T-Traffic (T-Mobile),... Personal information services various applications: tourist info, shop finder, night guides (bars, cinemas, etc.) friend finder, child watch, based on maps plus points of interest (spatial DB) examples: Vodafone, T-Mobile, Jamba, Wapme,

10 Context-aware Systems: In the Near Future Smart factory (SFB 627) Detailed model of factory Locations of objects Static objects: machines, installations,... Mobile objects: tools, work pieces, measurement equipment, transport systems State of objects: machinery settings, configuration and wear of tools,... Various context-aware applications tool management quality assurance production planning and controlling

11 Context-aware Systems: In the Near Future Vergina: Virtual Environment Research in Hapto-visual Gesturerecognition Interfaces Project Leader: Nicolas D. Georganas One Goal: Context-based, dynamic gesture recognition Recognizing natural human gestures according to the context of the situation Example: wrist rotation (e.g., sensed by data glove) in front of a door, should actuate turning the door knob and opening the door in front of a kitchen stove, should actuate switching on/off the stove

12 Vision: Federated World Models Navigation Smart Factory... City Guide Applications Federation Data from World Models billions of sensors... Information Spaces WWW Digital Libraries

13 DFG Sonderforschungsbereich Nexus Goal: Platform for context-aware applications Federation of heterogeneous world models Connect information spaces with real world objects Allow applications to interact with real world objects (Push/Pull) Physical World Nexus World Model Project: Started January research staff members 9 research groups of University of Stuttgart Applications System Functions Information Spaces

14 Context-aware Communication: Hoarding Goal: Information Anytime Anyplace Assumption: Hotspots scattered in a given region Infostation Wireless LAN Approach: While in WLAN hoard information items that might be requested in hoarding area outside WLAN: accesses are satisfied from hoard cache if possible Major Problem: Selection of items to hoard Mobile device Hotspot Transmission Area Hoarding Area

15 Hoarding: Exploit Location Information Hoarding Area Principle: [ACM MobiCom 2001] Hoarding area divided into zones Mobile devices record access to information items (Id, Location) upload log to infostation Infostation computes v(z): visit probability a(i,z): access probability Hoarding decision: items ranked by their access probability p ( i) = 1 (1 v( z) a( i, z)) z Z

16 Hoarding: Exploit Location Information Hoarding Area Principle: [ACM MobiCom 2001] Hoarding area divided into zones Hit ratio Enhanced approach Basic approach Mobile devices record access to information items (Id, Location) upload log to infostation Infostation computes v(z): visit probability a(i,z): access probability Hoard-size [number of objects] Hoarding decision: items ranked by their access probability p ( i) = 1 (1 v( z) a( i, z)) z Z

17 Hoarding: Context Beyond Location User preferences access behavior depends on user s interest (sports, culture,...) approach: record access behavior for individual user profiles Semantic distance of information objects hoard objects that belong together approach for semi-structured data (WWW) [MDM 2004] monitor how users navigate the web compute probabilities for link usage further improvements in hit ratio ~ 25% Others... 0,7 g s 1 s 2 0,4 0,25 0,26 0,75 0,46 Root Vertex 0,15 v 10 v 5 v 4 Session Start Vertices

18 Summary Trend: Proliferation of sensor technology smart things, sensor networks, wearable computers,... Most future applications will be context-aware System mechanisms will exploit context-information Example: Hoarding Promising field of research & development many research activities on various aspects one of it: Sonderforschungsbereich Nexus

19 Congratulations!