Adaptive Accurate Indoor-Localization Using Passive RFID

1. Adaptive Accurate Indoor-Localization Using Passive RFID Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology Nanjing University, China Presenter: Dr. Lei Xie, Associate Professor

2. Adaptive Accurate Indoor-Localization Using Passive RFID Outline 1 3 2 4 Motivation & Observation Our Solution Localization Model Performance Evaluation Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

3. Adaptive Accurate Indoor-Localization Using Passive RFID Scenario Library Management Warehouse Management Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

4. Adaptive Accurate Indoor-Localization Using Passive RFID Observation Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

5. Adaptive Accurate Indoor-Localization Using Passive RFID Observation More unstable than other wireless devices More unstable than other wireless devices Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

6. Adaptive Accurate Indoor-Localization Using Passive RFID Observation NOT monotonically decreasing with the increasing distance Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

7. Adaptive Accurate Indoor-Localization Using Passive RFID Observation arrived in its saturation value Increasing rapidly with increased power CAN NOT be detected Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

8. Adaptive Accurate Indoor-Localization Using Passive RFID Motivation • The traditional methods, like LANDMARC-based methods, are not suitable for adaptive indoor-localization using passive RFID technology: • Multi-path effect and noise in the realistic indoor-environment; • Unstable RSSI value and special reading phenomenon of passive tag. • Challenges • The RSSI value is not linearly increasing while the reader’s power increases. • The RSSI value of passive RFID tag is more unstable than the active tag. Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

9. Adaptive Accurate Indoor-Localization Using Passive RFID Localization Model and Scenario RSSIcan’t be detected RSSI increases sharply with large standard deviation RSSI converges to a saturated value with small standard deviation Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

10. Adaptive Accurate Indoor-Localization Using Passive RFID Localization Model and Scenario Localization areas like shelf in library, supermarket • Accuracy: the average error should be less than a certain threshold , e.g. 50cm; • Time-delay: the time duration for localization should be less than a certain threshold , e.g. 5 seconds in realistic application. Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

11. Adaptive Accurate Indoor-Localization Using Passive RFID System framework Calibrate the unstable reading of tags by using auto-feedback data Measuring tags in the stable region by using appropriate power Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

12. Adaptive Accurate Indoor-Localization Using Passive RFID Adaptive Power Stepping ---- Motivation Only the target tag and the reference tags close to target tag can be activated Interference from the reference tag far away. Appropriate power Too large power Too small power Can not be detected : Target tag : Reference tag Antenna Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

13. Adaptive Accurate Indoor-Localization Using Passive RFID Adaptive Power Stepping Different reading results between maximum power and smaller power Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

14. Adaptive Accurate Indoor-Localization Using Passive RFID Adaptive Power Stepping Finding the appropriate power for the target tag < p1, p2, p3, p4 > Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

15. Adaptive Accurate Indoor-Localization Using Passive RFID Adaptive Power Stepping If the RSSI equals to 0, it then cannot effectively discriminate the nearby tags from the distant tags in regard to distance < r1, 0, r3, 0 > < r1, r2, r3, r4 > < r1, 0, r3, 0 > Filter the reference tags when it can not be detected by enough number of antennas

16. Adaptive Accurate Indoor-Localization Using Passive RFID Adaptive Power Stepping Filtering the reference tags by comparing the activating power For example: Reference tag Target tag (25.7, 18.7, 30.7, 29.7) (28.7, 29.7, 24.7, 22.7) (3, 11, 6, 7) If the threshold is 5, Then Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

17. Adaptive Accurate Indoor-Localization Using Passive RFID Adaptive Power Stepping Adaptive When the number of reference tags are too small to support localization Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

18. Adaptive Accurate Indoor-Localization Using Passive RFID Adaptive Grid-based Calibration ---- Motivation Auto-detection technologies can have feedbacks for result The right grid with exact position, detected by feedbacks. Leverage these data for the following localization Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

19. Adaptive Accurate Indoor-Localization Using Passive RFID Adaptive Grid-based Calibration Rule Checking —— Generation For each grid, we can have a rule set L {li} For one certain grid: S1 – S2 (3000, 1500, 1000, 500) 1500 S1 – S2 >1500 as one rule for the grid. (3100, 1400, 980, 450) 1700 … … … … Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

20. Adaptive Accurate Indoor-Localization Using Passive RFID Adaptive Grid-based Calibration Rule Checking —— Checking Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

21. Adaptive Accurate Indoor-Localization Using Passive RFID Integrated Method Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

22. Adaptive Accurate Indoor-Localization Using Passive RFID Evaluation Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

23. Adaptive Accurate Indoor-Localization Using Passive RFID Evaluation More feedback fingerprints can provide higher success ratio Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

24. Adaptive Accurate Indoor-Localization Using Passive RFID Evaluation Both methods have much better performance than the baseline method AGC method can use the automatic feedback fingerprints to calibrate negative impact APS method can effectively reduce the minimum error Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

25. Adaptive Accurate Indoor-Localization Using Passive RFID Evaluation Maximum error in APS method is much larger than AGC method Minimum error in APS method is smaller than AGC method. Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

26. Adaptive Accurate Indoor-Localization Using Passive RFID Evaluation average error is smaller by more number of reference tags Maximum error by 10 reference tags is smaller than it by 15 reference tags More interference existed by more reference tags Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

27. Adaptive Accurate Indoor-Localization Using Passive RFID Evaluation • Time-delay of grid-based calibration procedure is fairlysmall. • The APS method and Integrated method have an average delay of 2.5 seconds. • It is caused by the power stepping to find the appropriate transmitting power. Xi Chen, Lei Xie, Chuyu Wang, Sanglu Lu State Key Laboratory for Novel Software Technology, Nanjing University, China

28. Adaptive Accurate Indoor-Localization Using Passive RFID Thank you! Q&A