in a Visible Light Communication System Stefan Schmid, Linard - - PowerPoint PPT Presentation

▶
in a visible light communication system
SMART_READER_LITE
LIVE PREVIEW

in a Visible Light Communication System Stefan Schmid, Linard - - PowerPoint PPT Presentation

Feb 04, 2023 140 likes •344 views

Using Smartphones as Continuous Receivers in a Visible Light Communication System Stefan Schmid, Linard Arquint, Thomas R. Gross ETH Zurich October 4, 2016 Smartphone as a Light Receiver No light sensor Ambient light sensor cannot be

slide-1
SLIDE 1

October 4, 2016

Using Smartphones as Continuous Receivers in a Visible Light Communication System

Stefan Schmid, Linard Arquint, Thomas R. Gross ETH Zurich

slide-2
SLIDE 2

Smartphone as a Light Receiver

2

  • No light sensor
  • Ambient light sensor cannot be

directly accessed and/or be sampled fast enough

  • Additional peripheral is impractical
  • Camera is the only option left
  • Sampling rates up to 240 Hz (FPS)
  • Exploiting the rolling shutter effect
slide-3
SLIDE 3

Outline

3

libvlc & EnLighting Real-Time Decoder Rolling Shutter Motivation Conclusion Evaluation

slide-4
SLIDE 4

libvlc & EnLighting

4

EnLighting

libvlc

  • S. Schmid et al. “LED-to-LED Visible Light Communication Networks”, 2013
  • S. Schmid et al. “Continuous Synchronization for LED-to-LED Visible Light Communication Networks”, 2014
  • S. Schmid et al. “EnLighting: An Indoor Visible Light Communication System Based on Networked Light Bulbs”, 2016
slide-5
SLIDE 5

PHY Layer – Communication and Illumination

5

S1 G G S2 G D1 D2 G Communication Slot Illumination Slot ILLU COM

1000µs

ILLU COM ILLU COM ILLU COM ILLU COM ILLU COM ILLU ILLU COM COM

  • Illumination (ILLU) slots and

communication (COM) slots are alternating

  • ILLU slots are used to provide the

necessary light output for illumination

  • During COM slots, there is either

no light output while sensing or light output is enabled during data intervals while transmitting

  • Light is sensed during the

synchronization (S1, S2) and data (D1, D2) intervals S1, S2: Synchronization intervals G: Guard interval D1, D2: Data intervals C: Charge (reverse bias) LED M: Measure remaining voltage

M C C M C M M C

slide-6
SLIDE 6

PHY Layer – Constant Light Output

6

1000µs

ILLU ILLU COM ILLU COM ILLU COM ILLU COM ILLU COM ILLU COM ILLU ILLU COM COM

D D D D D D C C C C C

50% light >50% light 50% light

slide-7
SLIDE 7

PHY Layer Modes

7

SINGLE Communication Slot S1 G D1 G D2 G S2 50µs 170µs DOUBLE S1 G D1,1 D1,2 G D2,1 D2,2 G S2 85µs QUAD S1 G D1,1 D1,2 D1,3 D1,4 G D2,1 D2,2 D2,3 D2,4 G S2 42µs S1 G D1,1 OCTA D1,2 D1,3 D1,4 D1,5 D1,6 D1,7 G D1,8 D2,1 D2,2 D2,3 D2,4 D2,5 D2,6 D2,7 D2,8 G S2 21µs

  • Shorter data intervals require more precise synchronization
  • LEDs receive less light during shorter data intervals (lower signal strength)
  • PHY layer mode can be selected dynamically
slide-8
SLIDE 8

Rolling Shutter (I)

8

[Image Source: Point Grey Research Inc.]

  • Global shutter
  • Complete image sensor is exposed at the

same time

  • Captures one moment in time
  • Rolling shutter
  • Image sensor is exposed line by line
  • Captures multiple moments in time
slide-9
SLIDE 9

Rolling Shutter (II) – Recorded Frame

9

500 µs Sensor line 1 Sensor line 720

Rolling shutter extends the camera’s sampling rate (FPS)

slide-10
SLIDE 10

Intra-Frame Gap

10

  • Gap between two

consecutive frames

  • No light can be received
  • Gap duration depends on

camera and camera mode

  • Higher frame rates reduce

gap width

  • Slow-motion capturing

(240 FPS) for smallest gap

slide-11
SLIDE 11

Intra-Frame Gap Measurement

11

Light source is enabled for the duration

  • f half a frame (example):
  • 1. Light pulse is completely visible (417

pixels)

  • 2. Light pulse is partly visible (152 pixels)
  • 3. Remaining part of the light pulse is

visible in the next frame (175 pixels) 90 pixels are missing; results in a gap of approximately 450 µs (measured on an iPhone 6s at 240 FPS)

slide-12
SLIDE 12

Redundancy

12

  • Gap duration is (on average) shorter than 500 µs (240 FPS)
  • Missing data intervals can be reconstructed
slide-13
SLIDE 13

Protocol & Rolling Shutter

13

PHY mode: SINGLE PHY mode: DOUBLE

slide-14
SLIDE 14

Real-Time Decoder (iOS Application)

  • 1. Exposure adaptation
  • 2. Gradient detection
  • 3. Slot detection

14

  • 4. Inter-frame handling
  • 5. Bit processing
  • 6. Error detection / correction
slide-15
SLIDE 15

Testbed Setup

15

1. iPhone 6s (receiver) running the real-time decoder application 2. Light source (transmitter) connected to a microcontroller (running the VLC protocols) 3. White wall: illuminated by the transmitter and captured by the receiver

slide-16
SLIDE 16

Throughput – PHY mode SINGLE

16

  • Stable communication

up to 2.75 m

  • 750 b/s saturation

throughput

slide-17
SLIDE 17

Throughput – PHY mode DOUBLE

17

  • Stable communication

up to 2.75 m

  • 1300 b/s saturation

throughput

slide-18
SLIDE 18

Related Work

  • J. Ferrandiz-Lahuerta et al. “A Reliable Asynchronous Protocol for VLC

Communications Based on the Rolling Shutter Effect”, 2015

  • 700 b/s, up to 3 m, reflection
  • H.-Y. Lee et al. “RollingLight: Enabling Line-of-Sight Light-to-Camera

Communications”, 2015

  • FSK, multiple light sources (direct), 90 b/s, up to 5 m
  • P. Hu et al. “ColorBars: Increasing Data Rate of LED-to-Camera

Communication using Color Shift Keying”, 2015

  • CSK, 5.2 kb/s

18

slide-19
SLIDE 19

Conclusions

  • Integration of a smartphone (camera) as a

receiver into an existing VLC system

  • Flicker-free light source
  • Exploiting slow-motion capturing (240 FPS)

and rolling shutter

  • Real-time decoder implemented as an iOS

application

  • Communication distances up to 3 m
  • Data throughput up to 1300 b/s

19

slide-20
SLIDE 20

?

20