AccelPrint: Imperfections of Accelerometers Make Smartphones - - PowerPoint PPT Presentation

▶
accelprint imperfections of
SMART_READER_LITE
LIVE PREVIEW

AccelPrint: Imperfections of Accelerometers Make Smartphones - - PowerPoint PPT Presentation

Jan 28, 2023 402 likes •807 views

AccelPrint: Imperfections of Accelerometers Make Smartphones Trackable Sanorita Dey, Nirupam Roy, Wenyuan Xu, Romit Roy Choudhury, Srihari Nelakuditi People use hundreds of apps Some apps are sneaky Exchanging IDs without consent is

slide-1
SLIDE 1

AccelPrint: Imperfections of Accelerometers Make Smartphones Trackable

Sanorita Dey, Nirupam Roy, Wenyuan Xu, Romit Roy Choudhury, Srihari Nelakuditi

slide-2
SLIDE 2

People use hundreds of apps

slide-3
SLIDE 3

Some apps are sneaky

  • Exchanging IDs without consent is rampant

– IMEI (device id), IMSI (subscriber id), or ICC-ID (SIM card serial number) help track users

  • One possible Solution: TaintDroid

– Realtime filtering of exchange of device IDs

slide-4
SLIDE 4

Law: Get user’s consent

  • While installing a cookie
  • While sharing location
slide-5
SLIDE 5

People use hundreds of apps

slide-6
SLIDE 6

Our findings

Accelerometers have fingerprint Sensors can also potentially track the users

slide-7
SLIDE 7

What if accelerometers have fingerprints?

slide-8
SLIDE 8

What if accelerometers have fingerprints?

slide-9
SLIDE 9

What if accelerometers have fingerprints?

slide-10
SLIDE 10

Evidence of fingerprint

slide-11
SLIDE 11

Toy Experimental Setup

…

Controlled, Identical Impetus

slide-12
SLIDE 12

Toy Experimental Setup

…

slide-13
SLIDE 13

Toy Experimental Setup

  • Six stand-alone

accelerometer chips

  • Stimulation with an

external vibration motor

  • Arduino to control

vibration and collect accelerometer readings

slide-14
SLIDE 14

Accelerometers are distinguishable

Accelerometer chips of Samsung Galaxy S3 Accelerometer chips of Nexus S Accelerometer chips of Samsung Galaxy Nexus

slide-15
SLIDE 15

Accelerometers are distinguishable

Samsung S3 Samsung S3 Galaxy Nexus Galaxy Nexus Nexus S Nexus S

slide-16
SLIDE 16

Accelerometers are distinguishable

Nexus s_1 Nexus s_2

slide-17
SLIDE 17

Why are accelerometers distinct?

slide-18
SLIDE 18

Accelerometers are based on MEMS

slide-19
SLIDE 19

Internal structure of an accelerometer

slide-20
SLIDE 20

Reasons for difference in accelerometers

  • Manufacturing

imperfections

  • Idiosyncrasies due to

QFN and LGA Packaging

  • Subtle imperfections do

not alter the rated functionality

  • Small imperfections can

potentially introduce idiosyncrasies in data

slide-21
SLIDE 21

Evaluation and External Impact Analysis

slide-22
SLIDE 22

Larger Scale Exploration

107 stand-alone chips, smartphones and tablets in total

+

36 time domain and frequency domain features

80 stand-alone accelerometer chips 27 smartphones and tablets

Bagged Decision Trees for ensemble learning

(with accelerometer traces)

+

slide-23
SLIDE 23

Feature Selection

Time domain features Frequency domain features

Extract 8 time and 10 frequency domain features from S(i) and I(i)

slide-24
SLIDE 24

Overall classification performance

slide-25
SLIDE 25

Overall classification performance

MPU 6050 ADXL 345 MMA 8452q Nexus One Samsung S3 MPU 6050

slide-26
SLIDE 26

worst case precision & recall > 76% average precision & recall > 99%

Precision and Recall

slide-27
SLIDE 27

Questions

  • Is the external vibration mandatory for

fingerprinting the accelerometers?

  • What is the impact of smartphone CPU load
  • n fingerprints?
  • Does the fingerprint manifest only at

faster sampling rates?

  • Does the system need to be aware of the

surface on which device is placed?

slide-28
SLIDE 28

Precision and Recall Without Vibration

worst case precision & recall > 66% average precision & recall > 88%

slide-29
SLIDE 29

Natural Questions

  • Is the external vibration mandatory for

fingerprinting the accelerometers?

  • What is the impact of smartphone CPU load
  • n fingerprints?
  • Does the fingerprint manifest only at

faster sampling rates?

  • Does the system need to be aware of the

surface on which device is placed?

slide-30
SLIDE 30

Is the system sensitive to CPU load?

  • CPU load matters. But up to 20% difference, high classification precision
slide-31
SLIDE 31

Natural Questions

  • Is the external vibration mandatory for

fingerprinting the accelerometers?

  • What is the impact of smartphone CPU load
  • n fingerprints?
  • Does the fingerprint manifest only at

faster sampling rates?

  • Does the system need to be aware of the

surface on which device is placed?

slide-32
SLIDE 32

Does the fingerprint manifest only at faster sampling rates?

  • Even at slower sampling rates, devices exhibit discriminating features
  • Likelihood of distinguishing devices improves with faster sampling rates
slide-33
SLIDE 33

Natural Questions

  • Is the external vibration mandatory for

fingerprinting the accelerometers?

  • What is the impact of smartphone CPU load
  • n fingerprints?
  • Does the fingerprint manifest only at

faster sampling rates?

  • Does the system need to be aware of the

surface on which device is placed?

slide-34
SLIDE 34

Does the system need to be aware of the surface on which device is placed?

  • Training on different surfaces helps but the system is surface-agnostic
slide-35
SLIDE 35

Conclusion and Future Work

  • Accelerometers possess fingerprints
  • Next step is commercial-grade evaluation
  • How to scrub fingerprint from sensor data?
slide-36
SLIDE 36

Two objects may be indistinguishable …

slide-37
SLIDE 37

… but no two objects are identical

slide-38
SLIDE 38

Thank You

http://web.engr.illinois.edu/~sdey4/