Sensor Fusion: A Technique To Improve Location Systems E-NEXT WG1 - PowerPoint PPT Presentation

Sensor Fusion: A Technique To Improve Location Systems E-NEXT WG1 TF Thomas King University of Mannheim king@informatik.uni-mannheim.de Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 1/14

Motivation Introduction ■ Various applications require location information ● Motivation ◆ mobile ad-hoc routing: position based routing ● Outline ● Categorization ◆ mobile business: context-aware applications Existing Work Sensor Fusion Conclusion ■ various location systems are around ■ no highly accurate, easy-to-use indoor location systems is available Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 2/14

Motivation Introduction ■ Various applications require location information ● Motivation ◆ mobile ad-hoc routing: position based routing ● Outline ● Categorization ◆ mobile business: context-aware applications Existing Work Sensor Fusion Conclusion ■ various location systems are around ■ no highly accurate, easy-to-use indoor location systems is available ⇒ we investigate sensor fusion based location systems Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 2/14

Outline Introduction ■ Categorization ● Motivation ● Outline ■ Existing Work ● Categorization Existing Work ■ Sensor Fusion Sensor Fusion Conclusion ■ Conclusion Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 3/14

Categorization Introduction ● Motivation Location Systems ● Outline ● Categorization Use Existing Infrastructure Use Dedicated Infrastructure Existing Work Sensor Fusion Conclusion Satellite Based Satellite Independent Televison Bluetooth Wireless LAN FM Radio Infrared Ultrasonic Rosum RADAR RightSPOT GPS Active Badge Cricket Sensor Fusion Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 4/14

Global Positioning System - Overview Introduction ■ designed for the U.S. military Existing Work ● Global Positioning System - ■ funded and operated by the Department of Overview ● Global Positioning System - Defense Flaws ● RADAR - Overview ● RADAR - Database ■ operational since 1995 ● RADAR - Positioning ● Bluetooth ■ 24 satellites orbit the earth at a height of 20.000 Sensor Fusion Conclusion km ■ provides an accuracy of 5-20 meters in 95 percent of all cases Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 5/14

Global Positioning System - Flaws Introduction ■ a line of sight to at least 4 satellites is required Existing Work ● Global Positioning System - ■ the GPS signals are blocked by obstacles (walls, Overview ● Global Positioning System - foilage, ...) Flaws ● RADAR - Overview ● RADAR - Database ● RADAR - Positioning ● Bluetooth Sensor Fusion Conclusion Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 6/14

Global Positioning System - Flaws Introduction ■ a line of sight to at least 4 satellites is required Existing Work ● Global Positioning System - ■ the GPS signals are blocked by obstacles (walls, Overview ● Global Positioning System - foilage, ...) Flaws ● RADAR - Overview ● RADAR - Database ● RADAR - Positioning ● Bluetooth Sensor Fusion Conclusion Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 6/14

RADAR - Overview Introduction ■ wireless LAN access points are used to determine Existing Work the position of mobile devices ● Global Positioning System - Overview ● Global Positioning System - ■ two step approach: Flaws ● RADAR - Overview ◆ trainging phase: a database with signal strength ● RADAR - Database ● RADAR - Positioning values of the operation area have to be created ● Bluetooth Sensor Fusion ◆ location phase: uses this database Conclusion ■ median distance error 2.5 meters Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 7/14

RADAR - Database Introduction ■ a dense grid of measurement points is required Existing Work ● Global Positioning System - Overview ● Global Positioning System - Flaws ● RADAR - Overview ● RADAR - Database ● RADAR - Positioning ● Bluetooth Sensor Fusion Conclusion Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 8/14

RADAR - Positioning Introduction ■ the mobile device measures the signal strength of Existing Work the access points in communication range ● Global Positioning System - Overview ● Global Positioning System - ■ the mobile device compares this sample with the Flaws ● RADAR - Overview values stored in the database ● RADAR - Database ● RADAR - Positioning ● Bluetooth Sensor Fusion Conclusion Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 9/14

Bluetooth Introduction ■ invented to replace low-bandwidth cabling e.g. Existing Work computer peripherals ● Global Positioning System - Overview ● Global Positioning System - ■ communication range up to 10 meters Flaws ● RADAR - Overview ● RADAR - Database ■ proximity based location determination ● RADAR - Positioning ● Bluetooth Sensor Fusion Conclusion Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 10/14

Bluetooth Introduction ■ invented to replace low-bandwidth cabling e.g. Existing Work computer peripherals ● Global Positioning System - Overview ● Global Positioning System - ■ communication range up to 10 meters Flaws ● RADAR - Overview ● RADAR - Database ■ proximity based location determination ● RADAR - Positioning ● Bluetooth Sensor Fusion Conclusion Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 10/14

Sensor Fusion - Overview Introduction ■ modern mobile devices contain a multiple of Existing Work communication and sensor interfaces (e.g. Sensor Fusion Wireless LAN, Bluetooth, GSM, ...) ● Sensor Fusion - Overview ● Sensor Fusion - Challenges ● Generic Location System Architecture Conclusion Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 11/14

Sensor Fusion - Overview Introduction ■ modern mobile devices contain a multiple of Existing Work communication and sensor interfaces (e.g. Sensor Fusion Wireless LAN, Bluetooth, GSM, ...) ● Sensor Fusion - Overview ● Sensor Fusion - Challenges ● Generic Location System Architecture ⇒ exploit the correlation between sensed Conclusion parameters to increase the positioning accuracy and availability Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 11/14

Sensor Fusion - Overview Introduction ■ modern mobile devices contain a multiple of Existing Work communication and sensor interfaces (e.g. Sensor Fusion Wireless LAN, Bluetooth, GSM, ...) ● Sensor Fusion - Overview ● Sensor Fusion - Challenges ● Generic Location System Architecture ⇒ exploit the correlation between sensed Conclusion parameters to increase the positioning accuracy and availability Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 11/14

Sensor Fusion - Challenges Introduction ■ challenges: Existing Work ◆ different sensor types provide different accuracy Sensor Fusion and precision ● Sensor Fusion - Overview ● Sensor Fusion - Challenges ● Generic Location System ◆ sensor data is often noisy Architecture Conclusion ◆ position estimates from different sensors may conflict with each other Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 12/14

Generic Location System Architecture Introduction Existing Work << interface >> SensorFusion Sensor Fusion + getPosition ():Position ● Sensor Fusion - Overview ● Sensor Fusion - Challenges BayesianNetwork ● Generic Location System Architecture + getPosition ():Position Conclusion GPS + getPosition ():Position KNearestNeighbor Viterbi << interface >> LocationDeterminationAlgorithm + getPosition ():Position + getPosition ():Position + getPosition ():Position BluetoothSensor WirelessLANSensor << interface >> Sensor + getSignalStrengthVector ():Vector + getSignalStrengthVector ():Vector + getSignalStrengthVector ():Vector Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 13/14

Conclusion Introduction ■ Categorization of location systems Existing Work Sensor Fusion ■ Existing Work Conclusion ◆ GPS ● Conclusion ◆ RADAR ◆ Bluetooth ■ Sensor Fusion Thomas King, 05/19/2005 Sensor Fusion: A Technique To Improve Location Systems - p. 14/14