This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
8/7/2019 Unobtrusive Long-Range Detection of RFID Tag Motion
IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 55, NO. 1, FEBRUARY 2006 187
Unobtrusive Long-Range Detection
of Passive RFID Tag MotionBing Jiang, Student Member, IEEE , Kenneth P. Fishkin, Sumit Roy, Member, IEEE , and Matthai Philipose
Abstract—This paper presents a novel method for detectingthe motion of passive radio-frequency-identification (RFID) tagswithin the field of a detecting antenna. The method allows theunobtrusive detection of human interactions with RFID-taggedobjects without requiring any modifications to existing commu-nications protocols or RFID hardware. We use the response rate(a metric in lieu of the true received RF-signal intensity) at thereader to study the impact of tag translation, rotation, and cou-pling, as well as environmental effects. Performance is improvedby introducing the idea of multiple tags/readers. Movement-detection algorithms are developed and integrated into the RFID
monitoring system, and verified by experiments that demonstrateexcellent results.
rithms and protocols are required to disambiguate the
desired tag response from the multiple returns.
While these obstacles could be overcome, they would again
require new higher cost readers. We therefore investigate in-
stead techniques that are effective with existing tags and readers
without the need to modify either.
Instead of measuring the intensity at a signal tag directly, we
infer it indirectly. Existing readers all support a “poll” com-
mand, wherein the reader transmits a batch command signalto all tags (i.e., N polls within a period of 1 s) and counts
Fig. 2. Response rate at different locations, with N = 20 for a duration of 1 s.(a) Characteristics of response rate. (b) Power spectrum of response rate.
the number of responses received from that tag.1 We therefore
define a response rate α as
α =N rN
(1)
where N represents the number of polls and N r is the count
of responses to those polls. α is thus a dimensionless scalar
between the values of 0 and 1. Fig. 2(a) shows the response rate
of a tag at different locations in decreasing distance order from
the reader, for a fixed orientation; Fig. 2(b) shows frequency-
domain analysis based on the fast Fourier transform (FFT) of
the observed raw α set at each location.
The results reveal the following:
1) The farther the tag, the lower the response rate in Fig. 2(a)
as expected. This relationship is analogous to that of
1
The backscatter from the passive tag to the reader uses a load resistorconnected in parallel with the antenna that is switched on and off to modulatethe data stream [5].
8/7/2019 Unobtrusive Long-Range Detection of RFID Tag Motion
190 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 55, NO. 1, FEBRUARY 2006
Fig. 4. Relationship between the response rate and translation (915-MHzsystem). (a) Reponse rate versus translation: 915 MHz. (b) Relative standarddeviation versus translation: 915 MHz.
if the spatial variation of energy is to be used to detect tag
motion is given by
P th2 ≤ P r ≤ P th1 (8)
where P th1 and P th2 are determined by the physical
properties of the tag, the wavelength, as well as the
antenna/RF front-end circuit used. The response rate for
2.45 GHz does not show saturation, i.e., P r is always
smaller than P th1 as shown in Fig. 5, implying a much-
enhanced operating range.
2) The fields around an antenna are divided into two princi-
pal regions—the near field and the far field with boundary
given by [6]
r =2L2
λ(9)
where L is the aperture of the antenna. The calculated
Fig. 5. Relationship between the response rate and tag translation (2.45-GHzsystem). (a) Response rate versus translation: 2.45 GHz. (b) Response standarddeviation versus translation: 2.45 GHz.
r for 2.45 GHz (28 cm) conforms to the experimental
result (33 cm) well. The near field is not observed in the
915-MHz system, since the minimum received power of
the tag within the measurement range is higher than theupper threshold P th1. We note that αmean and αrms are
almost identical as expected and either can be used to
represent the response rate. αmean is used to represent the
response rate in the following section, because it is easier
to calculate.
3) In the reader antenna’s far field, the sensitivity of the
response rate with distance (i.e., the slope of the response
rate) varies significantly from 0.25 to 0.04/cm in Fig. 5,
but in general, is sufficient for motion detection. In the
reader antenna’s near field, this variation is much more
limited.
4) In the far field, the relative standard deviation increases
with distance and may also be used additionally to inferthe distance.
8/7/2019 Unobtrusive Long-Range Detection of RFID Tag Motion
196 IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENT, VOL. 55, NO. 1, FEBRUARY 2006
by analyzing changes in an easily derived approximation of
signal-intensity level. System performance is greatly improved
by introducing a multitag/reader configuration. The effect of
tag movement on the response rate was categorized into four
cases, which allow prompt and accurate detection. Experi-
ments proved the validity of this cost-effective response-rate-
based passive RFID system as a sensor network for movementdetection.
REFERENCES
[1] V. Stanford, “Pervasive computing goes the last hundred feet with RFIDsystems,” IEEE Pervasive Comput., vol. 2, no. 2, pp. 9–14, Apr.–Jun.2003.
[2] H. Stochman, “Communication by means of reflected power,” Proc. IRE ,vol. 36, pp. 1196–1204, Oct. 1948.
[3] J. Landt, “Shrouds of time—The history of RFID,” in RFID Connections.Pittsburgh, PA: AIM Global, 2002, no. 2.
[4] M. Philipose, K. P. Fishkin, M. Perkowitz, D. Patterson, and D. Haehnel,“The probabilistic activity toolkit: Towards enabling activity—Awarecomputer interfaces,” Intel Research Seattle, Seattle, WA, Tech. Memo.IRS-TR-03-013, Dec. 2003.
[5] K. Finkenzeller, RFID Handbook: Fundamentals and Applications inContactless Smart Cards and Identification, 2nd ed. New York: Wiley,2003.
[6] J. Kraus, Antennas, 2nd ed. New York: McGraw-Hill, 1988.[7] Agilent Technologies Application Note 1089, Designing Detector for
RF/ID Tags, Palo Alto, CA: Agilent Corporate.[8] U. Karthaus and M. Fischer, “Fully integrated passive UHF RFID
transponder IC with 16.7 µW minimum RF input power,” IEEE J. Solid-State Circuits, vol. 38, no. 10, pp. 1602–1608, Oct. 2003.
[9] E. B. Mode, Elements of Probability and Statistics. Englewood Cliffs,NJ: Prentice-Hall, 1966.
[10] K. P. Fishkin, B. Jiang, M. Philipose, and S. Roy, “I sense a disturbancein the force: Long-range detection of interactions with RFID-tagged ob-
jects,” in Proc. 6th Int. Conf. Ubiquitous Computing, Nottingham, U.K.,2004, vol. 3205, Lecture Notes in Computer Science, pp. 268–282.
Bing Jiang (S’01) received the B.S. degree fromTianjin University, China, in 1995, and the M.S.E.E.degree from the University of Washington, Seattle, in2003. He is working toward the Ph.D. degree in theDepartment of Electrical Engineering, University of Washington.
He worked as an Intern at Intel Research, Seattle,from September 2003 to June 2004. He is interestedin research on radio-frequency identification (RFID),sensors, and robotics.
Kenneth P. Fishkin received the B.S. degree incomputer science and B.S. degree in mathematicsfrom the University of Wisconsin, Madison, and theM.S. degree in computer science from the Universityof California, Berkeley.
He was previously a Senior Researcher at IntelResearch, Seattle, when the work presented in thispaper was performed. He is currently a Software
Engineer at Google. He has published widely in thefields of radio-frequency identification (RFID), userinterfaces, and computer graphics.
Sumit Roy (S’96–M’98) received the B.Tech. de-gree in electrical engineering from the Indian Insti-tute of Technology, Kanpur, India, in 1983, and theM.S. and Ph.D. degrees in electrical engineering andthe M.A. degree in statistics and applied probabilityfrom the University of California, Santa Barbara, in1985, 1988, and 1988, respectively.
He is a Professor of electrical engineering at theUniversity of Washington, Seattle, and a VisitingFaculty Consultant to Intel Research, Seattle. Hisareas of technical interest involve wireless commu-
nication systems. He is currently exploring the use of radio-frequency identi-fication (RFID) and 802.11 wireless-local-area-network (WLAN) technologieswithin networked ubiquitous computing environments. He recently spent twoyearson academic leave at the Wireless Technology Development Group withinIntel Labs, Hillsboro, OR, working on next-generation WLAN and personal-area-network (PAN) systems.
Prof. Roy is an active member of the IEEE Communications Society.
Matthai Philipose received the B.S. degree fromCornell University, Ithaca, NY, in 1994, and the M.S.degree from the University of Washington, Seattle, in1996, both in computer science.
He is a Researcher at Intel Research, Seattle (IRS).His primary areas of interest are programming lan-guages and probabilistic reasoning. He is currentlyworking on human-activity inferencing at IRS.