RF-Wear Wearable Everyday Body-Frame Tracking using Passive RFIDs - - PowerPoint PPT Presentation

▶
rf wear
SMART_READER_LITE
LIVE PREVIEW

RF-Wear Wearable Everyday Body-Frame Tracking using Passive RFIDs - - PowerPoint PPT Presentation

Jan 23, 2023 691 likes •1.45k views

RF-Wear Wearable Everyday Body-Frame Tracking using Passive RFIDs Haojian Jin Zhijian Yang Swarun Kumar Jason Hong 1 RF-Wear turns a regular clothing into a body-frame aware garment using low-cost , light weight , machine washable ,

slide-1
SLIDE 1

RF-Wear

Wearable Everyday Body-Frame Tracking using Passive RFIDs Haojian Jin Zhijian Yang Swarun Kumar Jason Hong

1

slide-2
SLIDE 2

RF-Wear turns a regular clothing into a body-frame aware garment using low-cost, light weight, machine washable, battery-free RFID tags.

2

slide-3
SLIDE 3

Commercial Tracking Wearables

3

slide-4
SLIDE 4

+

Pulse Sensor Pedometer (Accelerometer)

How do these devices track?

4

slide-5
SLIDE 5

many times, we want more than heart rate and steps….

5

slide-6
SLIDE 6

Personal Trainer in Fitness

6

slide-7
SLIDE 7

Gait Tracking in Rehabilitation

7

slide-8
SLIDE 8

Gesture Input in VR/AR

8

slide-9
SLIDE 9

how can we do body-frame today?

9

slide-10
SLIDE 10

Optitrack

10

slide-11
SLIDE 11

Kinect Leap Motion

Infrastructure-based sensing

Openpose (CMU)

11

slide-12
SLIDE 12

inertial sensors

Neuron Wearable Electronics

12

slide-13
SLIDE 13

Smart fabrics

Google jacquard [UIST 2016]

13

slide-14
SLIDE 14

RF-Wear

mobile, ad-hoc v.s. infrastructure solutions washable, durable, low cost v.s. wearable electronics continuous rich tracking v.s. smart fabrics

14

(limited gestures)

slide-15
SLIDE 15

RF-Wear

skeleton tracking for daily use. using low-cost, machine washable, lightweight, battery-free RFIDs

15

slide-16
SLIDE 16

180° 96° 80° 105° 75° 135° 45°

  • a. Shoulder
  • b. Knee
  • c. Elbow
  • d. Waist
  • e. Thigh

RF-Wear:

average joint angle tracking accuracy of 8~21°, 20~60 Hz

16

slide-17
SLIDE 17

17

slide-18
SLIDE 18

research contributions

1 A fjne-grained mobile RFID tag positioning 2 3

A RFID sensing primitive for joint tracking A practical body-worn RFID tag placement solution

4 A detailed prototype implementation and evaluation

18

slide-19
SLIDE 19

background

RFID sensing, phase measurement, triangulation

19

slide-20
SLIDE 20

RFID Reader RFID Antenna RFID Tags

RFID Sensing Confjguration

20

slide-21
SLIDE 21

RFID Antenna (Transmitter) RFID Tags (Refmector)

RFID Backscatter Communication

21

slide-22
SLIDE 22

RFID Antenna (Transmitter) RFID Tags (Refmector)

22

RFID Backscatter Communication

slide-23
SLIDE 23

θT θR θTag λ d θ1 θ2

Backscatter wave Direct wave

Phase in Backscatter Communication

LESS THAN 0.1 mm Phase Ranging resolution:

23

slide-24
SLIDE 24

Tagoram [MOBICOM 2014]

24

slide-25
SLIDE 25

d1 d2 d3 d4 A1 A2 A3 A4

Static multiple antennas at known positions Use triangulation to calculate the tag position

Stationary RFID Sensing

25

slide-26
SLIDE 26

d1 d2 d3 d4 A1 A2 A3 A4

Static multiple antennas at known positions Use triangulation to calculate the tag position

Stationary RFID Sensing Mobile/Wearable

26

slide-27
SLIDE 27

θ1 θ2

RF-Wear Key Primitives

27

slide-28
SLIDE 28

θ

reversing the tag-antenna relationship

28

θ1

past work RF-Wear

slide-29
SLIDE 29

1 2 3 4 5

θ

6

measure the radio signal time-of-arrival delay

29

slide-30
SLIDE 30

1 2 3 4 5

θ

6

the tag placement l is known

30

slide-31
SLIDE 31

31

θ1

the antenna is in the pocket the position may change when the user moves

slide-32
SLIDE 32

θ1 θ2

knee joint angle =

32

θ1 θ2

slide-33
SLIDE 33

ideally…

33

multipath

in reality…

slide-34
SLIDE 34

1 2 3 4 5

θ

6

Eigenspace method (MUSIC algorithm)

Broadside Angle (degrees)

  • 25
  • 20
  • 15
  • 10
  • 5

10 30 50 70 90 110 130 150 170

p(α)

Incoming signal power distribution across 0° to 180°

θ = α α

34

slide-35
SLIDE 35

35

θ1

Broadside Angle (degrees)

  • 40
  • 30
  • 20
  • 10

10 30 50 70 90 110 130 150 170

p(α)

α

Incoming signal power distribution across 0° to 180°

Real-world Spectrum

slide-36
SLIDE 36

θ1 θ2

36

  • 40
  • 30
  • 20
  • 10

Broadside Angle (degrees)

  • 60
  • 40
  • 20

10 30 50 70 90 110 130 150 170

p(α) p(α) 68°

α

Incoming signal power distribution across 0° to 180°

measure the offset of two spectrum to counter multipath signals

slide-37
SLIDE 37

RF-Wear on Body

37

slide-38
SLIDE 38

challenges on-body

1 2D sensing primitives to 3D space 2 3

Two Degree of Freedom Joints Fabric fmexibility

38

θ

slide-39
SLIDE 39

implementation

RFID tags, RFID readers, Software

39

slide-40
SLIDE 40

5.2 x 1.2 x 0.1 (cm) fmexible, washable 25Hz on the body (1m)

RFID Tags

40

slide-41
SLIDE 41

Software

implemented in Python computation time: 0.03s => live demo (15 Hz) raw signal rate at 20~60 Hz continuous skeleton tracking

Context: RapID [CHI’16] - 200 ms IDSense [CHI’15] - 2s discrete gesture recognition

41

slide-42
SLIDE 42

evaluation

1) Array geometry 2) Fabric fmexibility 3) Motion capture experiment

42

slide-43
SLIDE 43

1m away on the fmoor facing the same direction 30 seconds/repetition

microbenchmark

43

slide-44
SLIDE 44

repetitions

6 tag array dimensions [2x3; 2x4; 2x5; 3x3; 4x4; 5x5] X 3 aperture [3cm, 4cm, 5cm] X 6 relative angles [30°, 60°, 90°, 120°, 150°, 180°] X 3 repetitions = 324 experiments

example: 2x4 aperture: 5cm

44

slide-45
SLIDE 45

microbenchmark accuracy

45

azimuth elevation angle error (°) 3x2_5cm 4x2_5cm 4x2_4cm 5x2_4cm 3x3_5cm 3x3_4cm 3x3_3cm 4x4_5cm 4x4_4cm 5x5_5cm 5x5_4cm 5 10 15 20

best! worst!

slide-46
SLIDE 46

fabric fmexibility test

46

slide-47
SLIDE 47

repetitions

1 tag array confjguration [2x4 with an aperture at 5 cm] X 3 fabrics [cotton, wool, polyester] X 6 relative angles [30°, 60°, 90°, 120°, 150°, 180°] X 3 repetitions = 54 experiments (30 sec each data collection)

47

slide-48
SLIDE 48

fabric fmexibility test

A c c u r a c y E r r

  • r

( d e g r e e s ) Cotton pant Wool Sweater Jacket 6.56 5.24 8.69

48

context: cardboard: 4°

slide-49
SLIDE 49

motion capture

8 cameras on the ceiling sub-millimeter accuracy

49

slide-50
SLIDE 50

knee shoulder elbow

50

slide-51
SLIDE 51

walk in place (50s) walk around (50s)

+

51

slide-52
SLIDE 52

52

knee joint angle trace

Walk around Walk in-place time (sec) knee joint angle (°)

RF-Wear (Raw) RF-Wear (KF) GroundTruth

slide-53
SLIDE 53

hand movement (160 sec)

53

slide-54
SLIDE 54

54

elbow joint angle trace

RF-Wear (Raw) RF-Wear (KF) GroundTruth

54

time (sec) elbow joint angle (°)

slide-55
SLIDE 55

shoulder rotation (3 x 20 sec)

55

slide-56
SLIDE 56

56

shoulder joint angle trace

time (sec) shoulder joint angle (°)

RF-Wear (Raw) RF-Wear (KF) GroundTruth

Horizontal DOF Vertical DOF

RF-Wear (Raw) RF-Wear (KF) GroundTruth

slide-57
SLIDE 57

Evaluation Summary

If we use a tag array for 4X2 with an 5cm aperture, Card board accuracy: 4° On fabric: 6°-9° On body: knee 9° (walk in place), 12° (walk around). elbow 12°, shoulder (21° and 8°) Context (Kinect): knee joint angle accuracy in a gait cycle: 28.5°

Accuracy of the Microsoft Kinect™ for measuring gait parameters during treadmill walking [Gait & Posture 2015] 57

slide-58
SLIDE 58

discussion

58

slide-59
SLIDE 59

number of tags?

64 on four limbs + 48 on the main body = 112 tags

59

slide-60
SLIDE 60
  • n the fabric

in the fabric

  • n the body (tattoo)

in the body (implant)

60

slide-61
SLIDE 61

follow-up work

61

slide-62
SLIDE 62

WiSh: Towards a Wireless Shape-aware World using Passive RFIDs (MobiSys’18)

slide-63
SLIDE 63

conclusion

63

slide-64
SLIDE 64

RF-Wear

body-frame tracking for daily use turns a regular clothing into a body-frame aware garment using low-cost, light weight, machine washable, battery-free RFID tags tracks joint angle at 8~21°, 20~60 Hz

64

slide-65
SLIDE 65

RF-Wear

Wearable Everyday Body-Frame Tracking using Passive RFIDs

65

Haojian Jin http://haojianj.in/

slide-66
SLIDE 66

66

Q & A

slide-67
SLIDE 67

θT θR θTag λ d θ1 θ2

Backscatter wave Direct wave

Phase in Backscatter Communication

67

slide-68
SLIDE 68

Phase to Super Resolution Distance

The speed of radio in the air is 3x10^8 m/s. The 900 MHz radio will have 9x10^8 cycles in one second. The wavelength (the length of a cycle) would be 33 cm. The resolution of phase reading is 0.0015 radians. The distance resolution = = 0.0079 cm. LESS THAN 0.1 mm

68

slide-69
SLIDE 69

Mobile Reader (battery up to 8 hours)

slide-70
SLIDE 70

70

Refresh rate

Hardware limit reader: 1,100 tags/second. RFID tags backscatter frequency on body: 20 Hz. Software limit: MUSIC algorithm is computing expensive: 15 Hz.

https://www.atlasrfjdstore.com/impinj-speedway-revolution-r420-uhf-rfjd-reader-4-port/

slide-71
SLIDE 71

71

Moving antenna

Each angle computation was run independently based on one

  • bservation.

we can do 30~60 Hz with commercial RFID readers given the reader moves at human speeds.

slide-72
SLIDE 72

Accuracy of the Microsoft Kinect™ for measuring gait parameters during treadmill walking [Gait & Posture 2015]

knee in a gait cycle RMSD: 28.5° hip RMSD: 11.8° Context, accuracy of Microsoft Kinect

72

slide-73
SLIDE 73

73

Privacy (radio awareness)

Traditional architecture: Stationary readers + Mobile Tags RFWear, WiSh Mobile readers + Mobile/Stationary Tags Users will have the control and awareness the reader status.

slide-74
SLIDE 74

74

Body-frame v.s. skeleton

RF-Wear tracks the body-frame by tracking the way clothes move as the body moves. Limitation: RF-Wear can only track the joints covered by clothing. Advantage: We can also track stomach spasms, belly movement. :)