Continuous Scanning with Mobile Reader in RFID Systems : an Experimental Study

1. Continuous Scanning with Mobile Reader in RFID Systems: an Experimental Study Author：Lei Xie, Qun Li, Xi Chen, Sanglu Lu, Daoxu Chen Presenter：Dr. Lei Xie Associate Professor，Nanjing University, China lxie@nju.edu.cn

2. Outline Background and Motivation Problem Formulation Studies from Realistic Experiments Continuous Scanning with Mobile Reader Performance Evaluation Conclusion

3. Background and Motivation • Identify Tags in Passive RFID System RFIDReader PC Passive RFIDTag RFIDAntenna continuous wave backscatter

4. Background and Motivation • Application scenarios • How to scan books in a library or a bookstore？

5. Background and Motivation • Application scenarios • Howto make an inventory of merchandises in a store?

6. Background and Motivation • These requirements are very common for RFID-based applications. • Solutions • Solution 1: Scan the tags with a fixed reader. • Solution 2: Continuously scan the tags with a mobile reader.

7. Background and Motivation • Solution 1: Scan the tags with a fixed reader. Evaluation: Limited scanning range even with the maximum power

8. Background and Motivation • Solution 2: Continuously scan the tags with a mobile reader • Evaluation: the right way for efficient tag identification, effectively compensating for its limited reading range.

9. Background and Motivation • Challenges for Continuous Scanning with Mobile Reader • No realistic model to characterize the performance for mobile RFID reading in a large scale setting. • The factors that affect the mobile reading performance are very complicated. • Reader’s power, • Tag density. • …

10. Background and Motivation • Previous studies: • experiments in a small scale (< 20 tags). • reading performance in a close to free space scenario. • didn't consider important factors (e.g., reader power, tag density) • Previous work does not give a model for RFID reading process in a realistic and large scale setting • it does not include the power and tag density.

11. Background and Motivation Our Contributions • Model the reading performance based on • large experiments (240 tags, 90 tags per square meter) • model the relationship between reading performance and reading power & tag density • Algorithm design to optimize performance based on the model

12. Problem Formulation • Typical scenario: • Continuous scanning in realistic settings, i.e., using a mobile reader to identify a large volume of tags deployed over a wide area. • The tags are continuously placed with a uniform or nonuniformdensity. • Performance metrics: • Time efficiency • Energy efficiency • Coverage ratio

13. Problem Formulation

14. Experiment Settings The RFID reader is statically deployed by facing its antenna towards the book shelf. We set the distance to 1.5m to guarantee the reading performance. We use the Alien-9900 reader and Alien-9611 linear antenna with a directional gain of 6dB. The tags used are Alien 9640 general-purpose tags. We attach the RFID tags onto the books which are placed in a large bookshelf. The bookshelf is composed of 12 grids with 4 columns and 3 rows. This setting is close to a typical noisy condition, which is distinct from the free space condition.

16. Studies from Realistic Experiments • Probabilistic backscattering During the query cycles, each tag responds to the reader with a certain probability between 0 and 1. (b) Probability density functions (a) Histogram of read ratio

17. Studies from Realistic Experiments • Major vs minor detection region (c) Histogram of read ratio (ρ = 10) (d) Histogram of read ratio (ρ = 20)

18. Studies from Realistic Experiments • Major vs minor detection region Two distinct regions: the major detection region where the tags can be identified with high probability, and the minor detection region where the tags can be identified with low probability.. (e) Histogram of read ratio (ρ = 30) (f) Histogram of read ratio (ρ = 40)

19. Studies from Realistic Experiments • Marginal decreasing effect (g) Width of major detection region (h) The detection probability in major detection region

20. Studies from Realistic Experiments • Marginal decreasing effect As the reader’s power is increasing, the scanning range, the detection probability, as well as the number of identified tags, are not increasing equally with the power. (i) Overall number of tags identified after 50 query cycles

21. Studies from Realistic Experiments • Query cycle duration vs the number of identified tags per cycle (j) Query cycle duration (k) The number of identified tags per cycle

22. Studies from Realistic Experiments • Query cycle duration vs the number of identified tags per cycle As the reader’s power increases, the query cycle duration does not increase linearly with the number of identified tags per cycle, causing the variation of the throughput. (l) Throughput

23. Studies from Realistic Experiments • Deriving a Model from Realistic Experiments • A model of the effective scanning window to evaluate the reading performance over multiple tags. Within the effective scanning window, each tag has a probability to be detected for each query cycle. In order to guarantee the coverage constraint, multiple query cycles should be issued over each tag while it is within the effective scanning window. Figure 3: The model of effective scanning window

24. Studies from Realistic Experiments • Effective scanning window Assume the tags are uniformly deployed • The number of tags within the scanning window is always constant. • The number of tags in a query cycle mainly remains constant. If the mobile reader is set to a constant power and a constant moving speed • After multiple query cycles, each tag has a uniform probability to be detected.

25. Studies from Realistic Experiments • Effective scanning window • If we use p’ to denote the uniform detection probability, then, the probability for an arbitrary tag to be identified at least once is • The value of m is equal to τw/τc, here τw is the duration in the effective scanning window, and τc is the average duration of a query cycle. Moreover, τw is equal to the ratio of the window width w to the moving speed v, hence m = w/(v·τc).

26. Studies from Realistic Experiments • Effective scanning window • In order to increase the detection probability p for an arbitrary tag, it is essential to • (1) increase the number of query cycles m as much as possible; • (2) increase the detection probability p’ as much as possible. p’ Reader’s power pw w andτc w increases much more slowly than τc m = w/(v·τc)

27. Studies from Realistic Experiments • Conclusion from Effective scanning window • As the moving speed v decreases, the value of m is monotonically increasing, while the value of p’ remains unchanged. • As the reader’s power pw increases, the value of p’is monotonically increasing, while the value of m is monotonically decreasing. • The value of pw should be appropriately selected to optimize the performance.

28. Continuous Scanning with Mobile Reader Solution for uniform tag density • We need to figure out the optimized value of pw and v such that the objective is achieved while the coverage constraints are satisfied. • In regard to the coverage constraint, we need to guarantee it is equivalent to ensure Then, the maximum allowable moving speed to satisfy the coverage constraint

29. Continuous Scanning with Mobile Reader Solution for uniform tag density • Considering the time-efficiency, in order to minimize T , it is equivalent to maximize v • Considering the energy constraint E ≤ α, the optimal value of power can be computed according to the following formulation:

30. Continuous Scanning with Mobile Reader Solution for uniform tag density • Considering the energy-efficiency, in order to minimize E, it is equivalent to minimize pw/v . • Considering the time constraint T ≤ γ, the optimal value of power can be computed according to the following formulation:

31. Continuous Scanning with Mobile Reader • Solution for uniform tag density Figure 4: Compute the value of yTand yEwith various values of pw

32. Continuous Scanning with Mobile Reader Solution for uniform tag density • In regard to various tag densities ρ, we pre-compute the optimal pairs of , and store them in a table. • When dealing with an arbitrary tag density, use the optimal pair of to achieve the time/energy efficiency.

33. Continuous Scanning with Mobile Reader Estimate the tag density: an important factor on the performance metrics. • Whenthe tag density cannot be pre-fetched or varies along the forwarding direction, it is essential to accurately estimate the current tag density, such that can be computed. • Challenges: • Due to the probabilistic backscattering, it is difficult to directly estimate the tag density. • Commercial RFID readers do not expose these low-level data to upper-layer applications. • It is essential to estimate the tag density in a more practical way.

34. Continuous Scanning with Mobile Reader • Estimate the tag density If the reader’s power pw is set to a certain value, the number of identified tags per cycle nc is varying as the tag density ρ varies, with a very small standard deviation. Due to the small variance of nc, there is a very stable pattern between nc and ρ that varies with pw.

35. Estimate the Tag Density • Algorithm Design Get nc in various power levels, assemble as a vector V. • Very practical and fully compatible with EPC C1G2 standard • Does not require to obtain any low-level parameters for RFID readers. Compare the similarity between V and the previously stored patterns. We use the k-nearest neighbor method to estimate the tag density based on k-nearest reference tag densities

36. Continuous Scanning for nonuniform tag density • In some applications, the tags are not uniformly deployed. During continuous scanning, the tag density always changes. • The constant moving speed and constant power for the mobile reader is no longer suitable. • The tag density changes slowly along the forward direction. • Close to uniform within the effective scanning window . • Reduce the situation with nonuniform tag density into multiple snapshots with fairly uniform tag density.

37. Continuous Scanning for nonuniform tag density • Solution for nonuniformtag density • In each query cycle, the mobile reader can be reset with the optimal pairs (pw,v) according to the nearby tag density. • Effectively improving reading performance by adjusting the reader’s power pw and the moving speed v. pw2，v2 pw3，v3 pw4，v4 pw1，v1 pw5，v5 pw6，v6

38. Experiment Settings We evaluate the performance in realistic settings. We use the Alien-9900 reader as the mobile reader to move forward for continuous scanning.

39. Performance Evaluation Both the 1-nearest neighbor method (1NN) and the 2-nearest neighbor method (2NN) achieve fairly good performance in terms of estimation accuracy. The standard deviation for 1NN: <5 tags/grid, the standard deviation for 2NN: < 3 tags/grid. Figure 5: Evaluate the accuracy of tag density estimation

40. Performance Evaluation While guaranteeing the coverage ratio, the optimal settings canachieve much better time-efficiency and energy-efficiency than other settings. Time-E saves more than 50% of the scanning time compared with Baseline, and Energy-E saves more than 83% of the energy consumption compared with Baseline. Figure 6: Experiment results in realistic settings

41. Conclusion • (1) The first to conduct an extensive experimental study over a relatively large number of tags (up to 240 tags) and a rather high tag density (up to 90 tags per square meter) in realistic settings • (2) The first work to give a framework of optimizing mobile reading performance based on experimental study • (3) A number of novel techniques in making our algorithms practical • (4) Being compatible with RFID standard, our solutions deliver significant performance gain. • Our practical solutions respectively reduce scanning time by 50% and energy consumption by 83% compared to the prior solutions.

42. Q & A Thanks for your attention!