tagoram: real-time tracking of mobile rfid tags to high-precision accuracy using cots devices...

Tagoram: Real-Time Tracking of Mobile RFID Tags to High-Precision Accuracy

Using COTS Devices

Xiang-Yang Li

Lei Yang, Yekui Chen, Xiang-Yang Li, Chaowei Xiao, Mo Li, Yunhao Liu

Sep. 9, 2014

Tagoram: Real-Time Tracking of Mobile RFID Tags to High-Precision Accuracy

Using COTS Devices

Xiang-Yang Li

Lei Yang, Yekui Chen, Xiang-Yang Li, Chaowei Xiao, Mo Li, Yunhao Liu

Sep. 9, 2014

Movement with known track04.

Outline

State-of-the-art02.Overview03.

Implementation & Evaluation06.Movement with unknown track05.

Motivation01.

Pilot Study07.

Conclusion08.

Motivation1

Human-computer interface

Imagine you can localize RFIDs to within 0.1cm to 1 cm!

Baggage sortation in airport

Imagine you can localize RFIDs to within 0.1cm to 1 cm!

Goal-line technology

Imagine you can localize RFIDs to within 0.1cm to 1 cm!

Demonstration

http://young.tagsys.org/tracking/tagoram/youtube

High-Precision RFID Tracking Using COTS Devices

30cm40cm

Drawing in the Air

TrackPoint Deployed at AirportsReader

Industrial computer

State-of-the-art2

RFID Tracking & LocalizationState-of-the-art

2004 2014201320112010

[1] Lionel M. Ni, Yunhao Liu, etalLANDMARC: Indoor Location Sensing Using Active RFIDWireless Networks

1120mm

LANDMARC Lionel M. Ni

RFID Tracking & LocalizationState-of-the-art

2004 2014201320112010

[1] Lionel M. Ni, Yunhao Liu, etalLANDMARC: Indoor Location Sensing Using Active RFIDWireless Networks

1120mm

LANDMARC Lionel M. Ni

[3] P.Nikitin, etalPhase based spatial identification of uhf rfid tags.IEEE RFID 2010.

680mm

[2] C.Hekimian-Williams, eltaAccurate localization of rfid tags using phase difference. IEEE RFID 2010.

280mm

Diff basedC. H.Williams

Spatial basedP. Nikitin

RFID Tracking & LocalizationState-of-the-art

2004 2014201320112010

[1] Lionel M. Ni, Yunhao Liu, etalLANDMARC: Indoor Location Sensing Using Active RFIDWireless Networks

1120mm [3] P.Nikitin, etalPhase based spatial identification of uhf rfid tags.IEEE RFID 2010.

680mm

[2] C.Hekimian-Williams, eltaAccurate localization of rfid tags using phase difference. IEEE RFID 2010.

280mm

Diff basedC. H.Williams

Spatial basedP. Nikitin

[4] Salah Azzouzi, etalNew measurement results for the localization of uhf RFID transponders using an angle of arrival (aoa) approachIEEE RFID 2011.

300mm

210mm

[5] R. Miesen etal. Holographic localization of passive uhf rfid transponders. IEEE RFID 2011.

AOA basedS. Azzouzi

HolographicR. Miesen

RFID Tracking & LocalizationState-of-the-art

2004 2014201320112010

[1] Lionel M. Ni, Yunhao Liu, etalLANDMARC: Indoor Location Sensing Using Active RFIDWireless Networks

1120mm

[4] Salah Azzouzi, etalNew measurement results for the localization of uhf RFID transponders using an angle of arrival (aoa) approachIEEE RFID 2011.

300mm

210mm

[5] R. Miesen etal. Holographic localization of passive uhf rfid transponders. IEEE RFID 2011.

AOA basedS. Azzouzi

HolographicR. Miesen

[6] Jue Wang, etalDude, where's my card?: RFID positioning that works with multipath and non-line of sightACM SIGCOMM 2013

110mm

12.8mm

[7] Jue Wang, etalRF-compass: robot object manipulation using RFIDsACM MobiCom 2013

PinItJue Wang

RF-CompassJue Wang

680mm

[2] C.Hekimian-Williams, eltaAccurate localization of rfid tags using phase difference. IEEE RFID 2010.

280mm

[3] P.Nikitin, etalPhase based spatial identification of uhf rfid tags.IEEE RFID 2010.

RFID Tracking & LocalizationState-of-the-art

2004 2014201320112010

[1] Lionel M. Ni, Yunhao Liu, etalLANDMARC: Indoor Location Sensing Using Active RFIDWireless Networks

1120mm

[4] Salah Azzouzi, etalNew measurement results for the localization of uhf RFID transponders using an angle of arrival (aoa) approachIEEE RFID 2011.

300mm

210mm

[5] R. Miesen etal. Holographic localization of passive uhf rfid transponders. IEEE RFID 2011.

[6] Jue Wang, etalDude, where's my card?: RFID positioning that works with multipath and non-line of sightACM SIGCOMM 2013

110mm

27.6mm

[7] Jue Wang, etalRF-compass: robot object manipulation using RFIDsACM MobiCom 2013

[8] Tianci Liu, etal,Anchor-free Backscatter Positioning for RFID Tags with High AccuracyIEEE INFOCOM 2014

128mm

[9] Jue Wang, etal,RF-Idaw: Virtual Touch Screen in the Air Using RF SignalsIEEE INFOCOM 2014

37mm

BackPosTianci Liu

RF-IDRawJue Wang

[3] P.Nikitin, etalPhase based spatial identification of uhf rfid tags.IEEE RFID 2010.

680mm

[2] C.Hekimian-Williams, eltaAccurate localization of rfid tags using phase difference. IEEE RFID 2010.

280mm

PinItJue Wang

RF-CompassJue Wang

RFID Tracking & LocalizationState-of-the-art

2004 2014201320112010

[1] Lionel M. Ni, Yunhao Liu, etalLANDMARC: Indoor Location Sensing Using Active RFIDWireless Networks

1120mm

[4] Salah Azzouzi, etalNew measurement results for the localization of uhf RFID transponders using an angle of arrival (aoa) approachIEEE RFID 2011.

300mm

210mm

[5] R. Miesen etal. Holographic localization of passive uhf rfid transponders. IEEE RFID 2011.

[6] Jue Wang, etalDude, where's my card?: RFID positioning that works with multipath and non-line of sightACM SIGCOMM 2013

110mm

27.6mm

[7] Jue Wang, etalRF-compass: robot object manipulation using RFIDsACM MobiCom 2013

[8] Tianci Liu, etal,Anchor-free Backscatter Positioning for RFID Tags with High AccuracyIEEE INFOCOM 2014

128mm

[9] Jue Wang, etal,RF-Idaw: Virtual Touch Screen in the Air Using RF SignalsIEEE INFOCOM 2014

37mm

BackPosTianci Liu

RF-IDRawJue Wang

WE ARE HERE

7mm

[3] P.Nikitin, etalPhase based spatial identification of uhf rfid tags.IEEE RFID 2010.

680mm

[2] C.Hekimian-Williams, eltaAccurate localization of rfid tags using phase difference. IEEE RFID 2010.

280mm

State-of-the-art Techniques

RSS based Methods1

RSS is not a reliable location indicator especially for UHF tags

RSSOrientation VS

Phase based Methods2

PinIt (SIGCOMM 2013) RF-Compass (MobiCom 2013)

Needs to deploy dense reference tags

State-of-the-art Techniques

Summary of Challenges

Need mm-level localization accuracy achieved• especially for mobile tags.

Small overhead, COTS devices• infeasible for using many references for a tracking

system spanning a long pipeline.

Fast-changing environment • multipath reflection of RF signals • varied orientation of tags• Doppler effect

How Tagoram works?3Overview the basic idea

Backscatter Communication

Double distance

❶𝑩𝒂𝒕𝒕𝒆𝒓𝒚 𝒇𝒓𝒆𝒆

Continuous wave

Device diversity

COTS RFID Reader

0.0015 radians (4096 bits)

accuracy

𝑰𝒎𝒑𝒊𝒏𝒋 𝑹𝒆𝒂𝒅𝒆𝒓

Problem definition

Surveillance region

Reader antenna

{(𝜃1,1 ,𝑡 1,1 ) ,(𝜃1,2 , 𝑡1,2)⋯ , (𝜃1, 𝑁 ,𝑡 1 ,𝑁)}

{(𝜃2,1 ,𝑡 2,1 ) ,(𝜃2,2 ,𝑡 2,2)⋯ ,(𝜃2 ,𝑁 ,𝑡 2 ,𝑁 )}

{ (𝜃𝑀 ,1 , 𝑡𝑀 , 1 ) ,(𝜃𝑀 , 2 ,𝑡𝑀 ,2)⋯ , (𝜃𝑀 ,𝑁 , 𝑡𝑀 , 𝑁 )}⋮ ⋮ ⋮

𝑀×𝑁

Utilizing antennas’ locations, sampled phase values and timestamps to find out the tag’s trajectory ?

Tagoram

Controllable CaseCase 1.

Uncontrollable CaseCase 2.

Movement with Known TrackControllable case

4

Our goal is to find the

tag’s trajectory with

a known track.

Virtual Antenna Matrix

Conveyor

Inverse Synthetic Aperture Radar

𝒇 (𝒕𝟎)𝑉

𝐴1

𝐴2

𝐴1,1 𝐴1,2 𝐴1,3 𝐴1,4

𝐴2,1 𝐴2,2 𝐴2,3 𝐴2,4

Initial position

RF Hologram

𝑳

𝑾

𝑾

𝑳

𝒙𝒘 , 𝒍

Surveillance region grids

RF Hologram pixels

How to define the likelihood?

The key is

RF Hologram

Naïve Hologram

Augmented Hologram

DifferentialAugmented Hologram

Thermal noise

Device diversity

Naïve Hologram

-1.5

-1

-0.5

0

0.5

1

1.5

Chart Title

𝐴 𝑇

𝑒 𝑱 𝜃 The real signal

-1.5

-1

-0.5

0

0.5

1

1.5

Chart Title

𝐴 𝑋𝑒 𝑱 h(𝐴 , 𝑋 ) The theoretical signal

-2

0

2Chart Title

𝐴 𝑋𝑒 𝑱 (h ( 𝐴 ,𝑋 )−𝜃 )A virtual interfered signal

Naïve Hologram

Virtual interfered signal

Sum of all signals

Naïve Hologram

If is the initial location, the waves add up constructively.

Real axis

Imaginary axis

𝑒 𝑱 (𝒉 ( 𝑿 , 𝑨𝟏 ,𝟏 )−𝜽𝟏 ,𝟏)

𝑒 𝑱 (𝒉 ( 𝑿 , 𝑨𝟏 ,𝟐 )−𝜽𝟏 ,𝟐)

𝑒 𝑱 (𝒉 ( 𝑿 , 𝑨𝟒 ,𝟒 )−𝜽𝟒 ,𝟒)𝑒 𝑱 (𝒉 ( 𝑿 , 𝑨𝟒 ,𝟑 )−𝜽𝟒,𝟑 )

Naïve Hologram

Real axis

Imaginary axis

𝑒 𝑱 (𝒉 ( 𝑿 , 𝑨𝟏 ,𝟏 )−𝜽𝟏 ,𝟏)

𝑒 𝑱 (𝒉 ( 𝑿 , 𝑨𝟏 ,𝟐 )−𝜽𝟏 ,𝟐)

𝑒 𝑱 (𝒉 ( 𝑿 , 𝑨𝟒 ,𝟒 )−𝜽𝟒 ,𝟒)

𝑒 𝑱 (𝒉 ( 𝑿 , 𝑨𝟒 ,𝟑 )−𝜽𝟒,𝟑 )

If is not the initial location, the waves canceled out.

Naïve Hologram

High PSNR

−𝟕𝟎∼−𝟑𝟎𝒅𝒃𝒎

Influence from Thermal Noise

Experiment Observations CDF Modeling

varying with 𝝈=𝟎 .𝟏

𝜃∼𝒩(𝜇 ,𝜎 ) 𝐹 (𝜃 ;𝜇 ,𝜎 )

𝟎∘ 𝟒𝟎∘

𝟗𝟐𝟎∼𝟗𝟐𝟔𝑴𝑯𝒛

100  tags

How to deal with thermal noise?Augmented Hologram

Augmented Hologram

A probabilistic weight

(𝒉 ( 𝑿 , 𝑨 )−𝜽 )∼𝓝(𝟎 ,𝟎 .𝟏)

Probability of

Augmented Hologram

Ground truth Result

Shift

scattered

Influence from Tag Diversity

Experiment Observations

Random testModeling

At same position

70  tags times

Tag diversity

Pass KS-test to be verified over a uniform distribution with 0.5 significant level.

How to eliminate tag diversity?Differential Augmented Hologram

Differential Augmented Hologram

Using the difference of phase values

Differential Augmented Hologram

ResultGround truth

Movement with Unknown TrackUncontrollable case

5

Overview

Estimating Speeds, Fitting tag’s trajectory

1

Selecting the optimal trajectory2

Implementation & EvaluationPurely based on COTS devices6

Hardware & Software Introduction

ImpinJ R420 reader.

4 directional antennas

Reader Tag

Alien 2 × 2 Inlay

Alien Squiggle Inlay

Software

EPCglobal LLRP

Java

Linear Track

• A mobile object is emulated via a toy train on which a tag is attached.

• The system triggers the camera to take a snapshot on the train when firstly read.

• A camera is installed above the track to capture the ground truth.

Tracking Accuracy in Linear Track

Improve the accuracy by in comparison to RSS, OTrack, PinIt and BackPos.

DAH02468

10121416

Chart Title

DAH

𝟒

𝟏𝟑𝟏𝟒

Controllable Case with Nonlinear Track

• The track’s internal diameter is 540mm

• The distance varies from to

Nonlinear track

DAH02468

10121416

Chart Title

DAH

𝟒 .𝟗𝟖𝟐 .𝟗𝟓

𝟕 .𝟐𝟗

Under Uncontrollable Case

Achieve a median of 12.3cm accuracy with a standard deviation of 5cm.

Pilot StudyIn two airports7

Current workflow – Manual sortation

Sortation carousel Sorting baggage

It is error-prone step to find the baggage from the carousel in manual sortation.

Our system: TrackPointReader

Industrial computer

Tracking Visualization

Pilot site

Beijing Capital International Airport

Sanya Phoenix International Airport

Setup

• Each airport contains 5 TrackPoints, 4 visualization screens, and 22 RFID Printers.

• The two-year pilot study totally spent more than $600,000

Setup

• Consumed 110,000 RFID tags.

• Involved 53 destination airports, 93 air lines, and 1,094 flights.

Tracking Accuracy

Tagoram achieves a median accuracy of 6.35cm in practice.

• Provides real-time tracking of mobile tags with accuracy to mm-level.

• Limits almost all negative impacts. • Multipath effect• Doppler effect• Thermal noise• Device diversity

• Implemented purely based on COTS RFID products.

• Systematically evaluated under indoor environment and at two airports.

Conclusion

Not well-studied before

Questions?

hank you T