Smart Space - An Indoor Positioning Framework

Transcription

1 Smart Space - An Indoor Positioning Framework Droidcon 09 Berlin, Stephan Linzner, Daniel Kersting, Dr. Christian Hoene Universität Tübingen Research Group on Interactive Communication Systems (ICS)

2 Agenda 1. Location Matters! 2. Indoor Localisation Techniques 3. SmartSpace Framework (SSF) 4. Opportunities 2 von 24

3 Location Matters! Location changes everything. This one input our coordinates has the potential to change all the outputs. Where we shop, who we talk to, what we read, what we search for, where we go they all change once we merge location and the Web. Mathew Honan, WIRED magazine, von 24

4 BUT Location for outdoor environments is provided by the Global Positioning System (GPS) 4 von 24

5 BUT Location for outdoor environments is provided by the Global Positioning System (GPS). there is no solution for indoors yet! 4 von 24

6 AND New generation of mobile devices with multiple sensors WLAN, GSM/CDMA, IMUs and Camera 880 million global smartphone users by 2013 (Morgan Stanley) Omnipresent RFinfrastructure in urban areas WLAN: 319 million chipsets sold in 2008, with 862 million installed base (Morgan Stanley) GSM/CDMA 5 von 24

7 AND New generation of mobile devices with multiple sensors WLAN, GSM/CDMA, IMUs and Camera 880 million global smartphone users by 2013 (Morgan Stanley) Omnipresent RFinfrastructure in urban areas WLAN: 319 million chipsets sold in 2008, with 862 million installed base (Morgan Stanley) GSM/CDMA allows Fine grained sensor profiling of any environment 5 von 24

8 AND New generation of mobile devices with multiple sensors WLAN, GSM/CDMA, IMUs and Camera 880 million global smartphone users by 2013 (Morgan Stanley) Omnipresent RFinfrastructure in urban areas WLAN: 319 million chipsets sold in 2008, with 862 million installed base (Morgan Stanley) GSM/CDMA allows Fine grained sensor profiling of any environment enables Indoor Localisation 5 von 24

9 Agenda 1. Location Matters! 2. Indoor Localisation Techniques 3. SmartSpace Framework (SSF) 4. Opportunities 6 von 24

10 Indoor Localisation Techniques Location Fingerprinting (LFPT) Signal Strength Measurements WLAN RSSI GSM RSS Inertial Measuring Units (IMUs) Accelerometer Gyroscope Electric compass Marker based localisation QR-Codes Display von 24

11 LFPT WLAN Radio Propagation 8 von 24

12 LFPT WLAN Radio Propagation 9 von 24

13 LFPT - WLAN signal stability over time 10 von 24

14 LFPT - WLAN signal stability over time Awake 10 von 24

15 LFPT - WLAN signal stability over time Sleeping 10 von 24

16 LFPT = collect environmental sensor data GSM - 80 dbm, - 65 dbm - 45 dbm, - 57 dbm WLAN IMU Orientation Sensor Fusion Radio-Map 11 von 24

17 LFPT Sensor Fusion Synchronize sensors Measurement count Time Normalize data Variance Deviance Algorithms Fusionate data Kalman filters Particle filters Orientation Sensor Data GSM RSS WIFI RSS Location Fingerprint 12 von 24

18 LFPT How-to Testbed at University of Tübingen with 4 rooms + corridor: 1. Trainings-Phase Grid of Training-Points Record LFPTs Write to Radio-Map? 2. Realtime-Phase Record LFPT Alignment with Radio-Map Estimate position 13 von 24

19 LFPT How-to Testbed at University of Tübingen with 4 rooms + corridor: TP 1 TP 2 1. Trainings-Phase Grid of Training-Points Record LFPTs Write to Radio-Map TP 3 TP 4 2. Realtime-Phase Record LFPT Alignment with Radio-Map Estimate position TP 5 TP 7 TP 6 TP 8 13 von 24

20 LFPT How-to Testbed at University of Tübingen with 4 rooms + corridor: 1. Trainings-Phase Grid of Training-Points Record LFPTs Write to Radio-Map TP3 2. Realtime-Phase Record LFPT Alignment with Radio-Map Estimate position 13 von 24

21 IMU Localisation Dead Reckoning Current Position Estimate position based upon previous determined position Calculate Distance step length * total steps IMU Sensor Data Accelerometer, Orientation Step Detection Detect step patterns Filter Data Kalman Filter 14 von 24

22 IMU Localisation Step Detection 15 von 24

23 IMU Localisation Step Detection Swing Phase Stance Phase 15 von 24

24 IMU Localisation Step Detection Step Pattern 15 von 24

25 Agenda 1. Location Matters! 2. Indoor Localisation Techniques 3. SmartSpace Framework (SSF) 4. Opportunities 16 von 24

26 SSF - Architecture Custom Application Position API Motion Detector Core Framework Positioning Core Framework Persistance Sensor Synchronization Layer 17 von 24

27 SSF- Core Framework : Fixed Position : Motion LFPT Position GSM/WLAN/ X Marker Position + LFPT Training Inertial Position Tracking with IMUs 18 von 24

28 Sensor Data Condensation SSF - Sensor Synchronization Layer Core Framework Synchronization Strategy Sensor Fusion Sensor Data Processing Raw Sensor Data 19 von 24

29 Agenda 1. Location Matters! 2. Indoor Localisation Techniques 3. SmartSpace Framework (SSF) 4. Opportunities 20 von 24

30 New kinds of applications can be built Context aware Applications Localization of resources Automated service execution Implicit context recording, for later retrival Household Applications Room profiles Light applications Energy saving Shopping center Augmented (Reality) product search 21 von 24

31 SmartSpace Platform: Summary Extends the android platform by indoor localisation capabilities Uses WLAN, GSM/CDMA, IMUs and Camera Uses existing RF-infrastructure Sensor Fusion On-site positioning Autonomous Learning Trainings.apk Open Source Everybody can use it! 22 von 24

32 Thank you! Take everything you can get and cope with it! 23 von 24

33 Questions?? 24 von 24

34 Questions?? Stephan Daniel 24 von 24

35 Future of Localisation Mobile devices with multiple sensors Omnipresent RF- Infrastructure Localisation Cope with any available sensor data Seamless outdoor and indoor localisation