Optimal Multi-Element VLC Bulb Design with Power and Lighting - - PowerPoint PPT Presentation

Optimal Multi-Element VLC Bulb Design with Power and Lighting Quality Constraints

Sifat Ibne Mushfique and Murat Yuksel

• Project website: https://sites.google.com/site/nsfvlc/

Why Multi-Element/Stream VLC?

LED Array

Photo- detector

VLCS 2016 2

• Unused

space

Multiple data streams Narrow divergence – for higher spatial reuse Spherical structures – to retain smooth lighting Single data stream PHY solution Large divergence – for smooth lighting

VLCS 2016 3

Problem Statement

Ø Introduce an framework of a Multi-element Bulb Design that –

• Optimizes the placement of LED boards/transmitters
• Maximizing the signal-to-interference ratio (SIR)
• Maintain a minimum evenness of the lighting on the room floor
• Also, explores the optimization characteristics
• Under various constraints

Ø Challenges of a Multi-element Bulb Design

• Handle the line-of-sight (LOS) alignment management
• Implement seamless communication across the room
• Resolve inter-LED interference
• Balance the tradeoff: uniform lighting and high spatial reuse

VLCS 2016 4

Multi-element Bulb Configuration

VLCS 2016 5

System Model

VLCS 2016 6

Optimization Objective

Maximize SIR For a particular number of layers Add divergence angle to the optimization problem Make the problem constrained with Power Update the problem with the constraint on the illumination quality Two results are compared to analyze the effect of the illumination requirement on the overall optimization problem

VLCS 2016 7

Maximum SIR Problem (MAX_SIR)

Ø Variable Parameters

• ki , Number of Transmitters in Layer i
• LED boards can be placed in layers in many different ways
• ki depends on the size of the LED boards (rt) and the Bulb radius (R)
• θd, Divergence Angle of LEDs
• Large divergence angles à better lighting
• Narrow divergence angles à increased spatial reuse & higher SIR
• Different divergence angles are tried to find the configurations

yielding maximum SIR

VLCS 2016 8

Placement of the Transmitters

260 265 270 275 40 280 285 290 295 20

• 20

40 30 20 10

• 40
• 10
• 20
• 30
• 40

260 265 40 270 275 280 285 290 20 40 30

• 20

20 10

• 10
• 20
• 40
• 30
• 40

R R = 40 cm, rt = 3.5 cm

• No. of Layers = 7

R = 40 cm, rt = 5.5 cm

• No. of Layers = 4

Layers

VLCS 2016 9

Maximum SIR Problem (Contd.)

Ø Fixed Parameters

• Room size - 6m x 6m x 3m
• R, Radius of the Hemispherical Bulb – 40 cm
• rt, Radius of the LED Board/Transmitter - 3.5cm
• l, Number of Layers - Depending on R and rt
• Minimum number of layers à 2
• Maximum number of layers à calculated from R and rt
• The number of Layers and LED boards is varied between this

minimum and maximum value

VLCS 2016 10

Non-Linearity of the Objective Function

8

• No. of layers

6 4 2 20 25

Divergence angles

30 35 7 6 5 4 3 2 40

• Avg. SIR (dB)
• No. of layers

2 3 4 5 6 7

• Avg. SIR (dB)

2.5 3 3.5 4 4.5 5 5.5 6

20 degree 25 degree 30 degree

Local Maxima Local Minima

Multiple Local Maxima and Minima Saturation of the SIR Value

SIR doesn’t increase linearly

VLCS 2016 11

Updating the Maximum SIR Problem (MAX_SIR_LQ)

• Illumination Constraint (Is) is added to obtain smooth lighting
• Is is calculated by continuously taking average luminous intensity of 100

random points until the valueconverges

• Maximum allowed illumination variance, Ismax = 5
• The new objective function output is then (SIR / Is) instead of SIR

VLCS 2016 12

Results

Objective l (number of layers) k1 k2 k3 θd (in degrees) MAX_SIR 2 19 2

• 39.5

3 16 26 2 47.2 MAX_SIR_LQ 2 6 28

• 16
• MAX_SIR has much fewer LEDs in the higher layer
• MAX_SIR_LQ puts more LEDs in the higher layer to achieve

more even lighting across the room

VLCS 2016 13

Results (Contd.)

Power Consraint (Watts)

2 4 6 8 10

Objective function output (Avg. SIR / Is)

1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2

Power Constraint (Watts)

2 4 6 8 10

Illumination variance (Is)

1.2 1.4 1.6 1.8 2

• Avg. SIR (dB)

2 4 6 8

Illumination variance (Is) SIR without illumination constraint SIR with illumination constraint

SIR value saturates Lighting Quality starts to deteriorate as Is increases (SIR / Is) starts to fall from the same time

Optimum SIR is significantly lower in MAX_SIR_LQ

VLCS 2016 14

Summary and Future Work

• A framework is introduced to optimize a multi-element bulb design for

both illumination and communication quality of VLC in a room

• In the future, this can be done for different room size and bulb

parameters

• Trying with bulbs of some other shapes (Triangular, Square etc.) can be

interesting

• Multiple number of bulbs can be considered as well

QUESTIONS?