Development and Implementation of MIMO Optical Camera … · 2020. 2. 9. · 2/7/2020 1 Development and Implementation of MIMO Optical Camera Communication Systems Ph.D. Dissertation

2/7/2020

1

Development and Implementation of MIMO Optical Camera Communication Systems

Ph.D. Dissertation Defend

by Trang Nguyen

31th May 2018

Outlines

1 Introduction

A-QL system2

3

4

5

Screen OFDM system

RoI Signaling

Selective-RoI high-rate OCC System

Annex 1 –Implemented Software Apps

Annex 3 –Screen OFDM details

Annex 2 –PHY configurations

Annex 4 –Channel measurement

Annex 5 –Technical Feasibility

2

Additional Annexes: 01 – 10: Further Details of Implementation Aspects

1

2

2/7/2020

2

IntroductionGeneral description of the contents described in this thesis

3

Screen

A-QL

Screen

OFDM

RoI-OCC

signaling

Hybrid

OCC

wavefor

m

1

2

3

4

Screen-Camera OCC

This system

implements tri-color

band modulation for

a novel 2D-sequential

A-QL code design

RoI signaling waveform

This RoI signaling

waveform is designed for

delivering the short-ID of

multiple light sources to

conventional cameras.

Screen OFDM

This system

implements 2D-OFDM

for a quite-mature

screen code design

High-rate RoI signaling

This hybrid waveform is

to support cameras in

detection and tracking

multiple light sources

while staying connect at

high-rate with

interested ones.

Thesis Contents 4

3

4

2/7/2020

3

Classification of OCC

Screen

Nyquist sampling

Oversampling

by high frame rate

Region-of-Interest

Signaling

Rolling shutter

Nyquist sampling

Distance several meters < hundred meter hundreds of meter tens of meter

Data

ratekbps-Mbps kbps 10bps up to 22 kbps tens bps~ kbps

StandardIEEE 802.15.7m

PHY VI modesNot standardized

IEEE 802.15.7m

PHY IV modes

IEEE 802.15.7m

PHY V modes

Intended

Systems

• Screen Tx

• Typical cameras Rx

• Traffic light, LEDs-

array Tx

• High-speed camera

Rx

• Car light/traffic light/LED

Signage Tx

• RoI-camera Rx

• LED panel Tx

• Rolling shutter camera

Rx

Characteristics

• Massive spatial MIMO

• Short-range OWC

system utilizing

screen.

• Spatial MIMO

• Tx detection based

on image processing

is a drawback

• Hybrid waveform

• The RoI signaling stream

allows fast, reliable

detection and tracking of

multiple light sources

simultaneously.

• The trade-off between the

distance and the data

rate makes this the most

suitable option for indoor

application.

5

MIMO System ≠ MIMO Block coding

A-QL 2D-OFDMRoI

SignalingHybrid

waveform

Massive cells for a short-range link

• implements a single luminaire that

composes of multiple (single-color band or

tri-color band) cells

• Data rate which relies much on the number

of cells can be achieved up to Mbps.

Massive cells available within Screen Tx.

• Applicable for point-to-point (P2P) short-

range communication within several

meters at which Rx can process the huge

amount of cells located in the two-dimensional plane.

Massive links in a challenging environment

• aims to support multiple multi-array-

luminaires those are spatially separated.

• Data rate for a single link relies much on the

frame rate of the camera, leaving the spatial

dimension free for other links connectivity.

Limited cells available within LED-Tx.

• Applicable for multiple points-to-multiple

points (MP2MP) long-range communication

(such as in a night scene/ vehicularenvironment).

6

5

6

2/7/2020

4

Highlights of Proposing Screen-Camera Systems

2D-OFDMA-QL

Code Area

Perspective Distortion

360o Rotation

Frame rate variation Rolling effect

Quite-mature design of MIMO

Tested performance

2D-carriers allocation Complete Features as A-QL

7

Highlights of Proposing RoI-Signaling based OCC Systems

RoI Signaling

Hybrid waveform

S2-PSK

S2-PSK

Wide Applicability to Infrastructures

(LED-types and cameras)

Comparative Performance

Fuzzy decoder

▪ High speed with mobility support

▪ Design for efficient communication performance

▪ Good-dimming resolution and performance

▪ Applicable to either typical cameras/ RoI cameras

Multiple Tx

detection & tracking

8

7

8

2/7/2020

5

A-QL systemAsynchronous –Quick Link Screen-Camera System

Screen

A-QL

Short video demo – 16x16 A-QL 10

56 sec

9

10

2/7/2020

6

Short video demo – 32x32 A-QL 11

58 sec

Overall description

The A-QL system includes

• New 2D tri-color code design

• New bit mapper (bits to intensity)

• Channel encoder with particular encoding for asynchronous system

• Color calibration with estimation sequence

12

11

12

2/7/2020

7

Technical Contributions

Rotation support

Code Detection & Extraction

Simpler but more efficient than QR

code.

Sequential color-QR (tested)Rotation support in A-QL

Tested detection rate: > 20 fps

Linear estimation of cells

Efficient code extraction using

outer bolders.

13

Technical Requirements

Technical Contributions

Asynchronous bits

Modulation and Coding

Temporal sampling correction Rolling effect removal

Preamble-PHR frame Data frame

Server two critical purposes:

▪ Temporal sampling correction

▪ Detect and Remove the rolling

effected images

PHY frame format

Header with channel estimation

Tri-color mapper

14

Reliable Communication

13

14

2/7/2020

8

Technical Contributions

effective FEC

Implementation-ready

Outer

Reed SolomonInner Convolutional

Code

16×16 32×32 Diff. A-QL Hidden A-QL

Various versions of A-QL system are already implemented

Natural combination of FEC codes

▪ Inner CC

▪ Outer RR

Implementation shows BER <10-5

A-QL codes are implemented from

various adaptions.

▪ 16x16 and 32x32 A-QL

▪ Differential A-QL

▪ Hidden A-QL

15

Implementation Aspect

Specifications

Design of A-QL code

▪ Outer bolder to support Rx in extracting the code

▪ Reference cells to support Rotation, Time-variant downsampling, Rolling effect cancellation

▪ Data cells to deliver bits through tri-color bands

Conceptual design of A-QL code

16

15

16

2/7/2020

9

Specifications

Pi +Pj +Pk ≠1

Color Shift Keying

Pi +Pj +Pk =1

Bits-to-Intensity mapping in A-QL

▪ Bit mapping rule: more freedom in the bit mapping in A-QL

17

Specifications

Practical demo shows that the Walsh training

sequence is helpful for

▪ Color calibration

▪ Binary threshold

▪ Color calibration: resolved in the PPDU format

18

17

18

2/7/2020

10

Specifications

▪ PHY PIB attributes for APP-based flexible configuration of Tx and PPDU in A-QL

19

Implementation

▪ Testing of A-QL code with BER < 10-5 at 3m distance (tested and verified by ETRI)

16×16 A-QL

32×32 A-QL

20

19

20

2/7/2020

11

Implementation

▪ Testing of mono-color differential A-QL code at 30m distance (tested and verified by Korea Testing Laboratory)

16×16 mono color A-QL

16×16 Differential A-QL

21

Implementation

▪ Demo of 16x16 Hidden A-QL code (full screen mode) by imperceptibly modulating the screen intensity

16x16

16x16 embedding data

16x16 intensity modulation

22

21

22

2/7/2020

12

2D-OFDM screen-camera systemThis is intended for delivering 50kbps (implemented) or higher rate (implementable work) by employing 2-dimensional OFDM via Screen.

Screen

OFDM

Short video demo – Screen OFDM 24

23

24

2/7/2020

13

Overall description

The Screen OFDM system includes

• New design of multiple OFDM symbols within Tx as a MIMO approach

• All mature-features as A-QL has

• Complete creation procedures of OFDM symbols with experimental verification

• Detailed processing and decoding guidance

25

Technical Contributions

MIMO Tx

Complete Features like A-QL

▪ Multi spatial-OFDM symbols

within Tx

▪ Multi subcarriers within symbol

Multiple symbols within Tx 2D-carriers Practical measurement of

channel attenuation

▪ Rotation support

▪ Fast code detection and extraction

▪ Perspective Distortion Correction

▪ Correction of temporal &Spatial

sampling error

-Temporal sampling error

-Spatial sampling error

26

New design of Screen code

25

26

2/7/2020

14

Technical Contributions

Theoretically analyze the impacts of channel imperfection

Cosine-Fourth Law Blurry Image Impact Temporal and Spatial

sampling error

Nonlinear channel response Pixel Eb/No Clipping Noise

27

Technical Contributions

Practically analyze the impacts of channel imperfection

On-focus condition Blurry condition

28

27

28