PICARSO_Seminar - University of Adelaide

Report
Hardware
System
Process
Painting
System
Project
Outcomes
Design
Problems
Testing and
Results
Image
Processing
Future Work
Design
Specifications
Control
Software
1028:PICARSO
Sven Paschburg
2
PICARSO:
 Cable-driven robot
 Process standard image formats
 Reproduce images on vertical surface
Image
Processing
1028:PICARSO
Control
Software
Hardware
Sven Paschburg
Painting
System
3
Image Output
Original
Processed
Painted
End-Effector
1028:PICARSO
Sven Paschburg
4
MATLAB
Motor 3
Image Processing Toolbox
Motor 2
RS232
Cables
Cables
Vertical
Wall
End-Effector
(Mathworks 2010)
Graphical User Interface Design Environment
Motor 1 (GUIDE)
1028:PICARSO
Sven Paschburg
5
Hardware
Design
Problem
Painting
System
Project
Outcomes
Design
Specifications
Testing and
Results
Image
Processing
Future Work
System
Architecture
Control
Software
1028:PICARSO
Samuel Oosterholt
6
Goals:
 Develop mechanical system
 Scalable work space size
 Up to 3×3m
 Manipulate and stabilise the end-effector
 Mounting the system:
 in operation and
 in testing
1028:PICARSO
Samuel Oosterholt
7
Hektor (Franke & Lehni 2002)
Viktor (Lehni & Rich 2008)
Four actuators
Two
actuators
 Image producing robots
 ‘X’
‘V’ Configuration
 Scalable workspace
sizeon gravity
 Relies
 Cable driven

Hektor’s actuator
configuration
1028:PICARSO
PICARSO’s actuator
configuration
Samuel Oosterholt
Viktor’s actuator
configuration
8
Full System

Three motors

‘Y’ configuration

Cables

Upper motors
control position

Lower stabilises

Reduces cost
Render of PICARSO system
1028:PICARSO
Samuel Oosterholt
9
Motor Mount
Base Plate
Motor
Motor Controller
Spooling System
Cable feeding
system
1028:PICARSO
Samuel Oosterholt
10
Motor Mount

Motors parallel to
painting surface

Plate mounts to
painting surface

Double spool and
bearing

Two cables
 Reduce yaw
and pitch
 Minimise
kickback
1028:PICARSO
Samuel Oosterholt
11
Motor Mount
180°

Reorientated by
Cables:
eyebolts
and (Braid
 Spiderwire
fishing line)
pulleys
 Ø = 0.30mm

 Tmax
= 13.6kg
Can
move
pulleys
 Small
elasticity
and
eyebolts


Proximity to canvas
 Stability
10m
of cable
 7×7m workspace
Lower mount uses
fairleadsattach to
 Cables
 180° sweep
turnbuckles

1028:PICARSO
Samuel Oosterholt
12

Hardware selected from
Maxon Motors
 Numerous operation
modes
 Modular components
 Discounted cost & support
250W EC45 Motors
and Maxon EPOS2
70/10 Motor Controller
1028:PICARSO
Samuel Oosterholt
13
Not feasible
Canvas
implemented
to wall mount for testing
 Simulate
 3.6
× 3.2msurface
easel (working area: 2.8 × 2.8m)
 Reduces ripple
in wind environment
 Constrained
by testing

PICARSO’s easel in the
Vibrations Laboratory
PICARSO’s easel in the
FSAE shed
PICARSO’s canvas
1028:PICARSO
Samuel Oosterholt
14
Hardware
Design
Problem
Painting
System
Project
Outcomes
Design
Specifications
Testing and
Results
Image
Processing
Future Work
System
Architecture
Control
Software
1028:PICARSO
Neil Yeoh
15
Goals:
 Design an appropriate painting system
which:





1028:PICARSO
Produces circular patterns (< 3cm)
Is fast (> 1Hz) and light (< 3kg)
Does not cause instabilities
Houses suitable paint capacity
Is reliable, durable, and repeatable
Neil Yeoh
16
 Three types of painting mechanisms
*1
*2
Airbrush
Spray Can
*3
Spray Gun
Circular Patterns
( <3cm)
Fast ( >1Hz)
Light ( <3kg)
No Instabilities
Paint Capacity
Reliable
Durable
Repeatable
** Image references at end of slides
1028:PICARSO
Neil Yeoh
17
Electrical Air
Compressor
Automatic
Pressure-fed
Spray Gun
Paint Line
Air Regulator
Compressed
Air Line
*7
Personal
Computer
*5
*4
Air Line
Paint Regulator
Solenoid
Pressurised
Paint Canister
*8
Electrical Signal
*9
*6
** Image references at end of slides
1028:PICARSO
Neil Yeoh
18

Anest Iwata
 SGA-101 Automatic Pressure-fed Spray Gun
Pattern Adjust Knob
Fluid Adjust Knob
Air Line
Fitting Knob
Nozzle
Paint Line
1028:PICARSO
Neil Yeoh
19
Testing procedure
 Ideal Settings identified:

Specification
Setting
Air Pressure
2.0 bar
Paint Pressure
0.5 – 1.0 bar
Spray Distance
5 – 10 cm
Fluid Adjust Knob
Half Turn
Spray Duration
0.2 – 1.0 s
Repeatability

Spray Duration
Results:
 Consistent < 3cm black circles
1028:PICARSO
Neil Yeoh
20
1028:PICARSO
Neil Yeoh
21
Large Sleeve
Sleeve Shaft
Bearing Spacer
Small Sleeve
Eyebolts
Bearings
Spray Gun
1028:PICARSO
Neil Yeoh
22
1028:PICARSO
Neil Yeoh
23
Hardware
Design
Problem
Painting
System
Project
Outcomes
Design
Specifications
System
Architecture
Testing and
Results
Image
Processing
Future Work
Control
Software
1028:PICARSO
Ian Hooi
24
Goals:
 Develop image processing software to:
 Transform input image to user specified
settings
 Reproduce input image in binary form
 Output in Raster (pixel-by-pixel) form
Extension Goal:
 Output in Vector (line) form
1028:PICARSO
Ian Hooi
25
1.
2.
3.
4.
Input image
Resized to appropriate resolution
Stretching and refitting
Greyscale form
 Scaled from 0-1 where 0 is black, 1 is white
Original
1028:PICARSO
Resized
Ian Hooi
Greyscale
26
Fill Images: Direct conversion from
greyscale  binary
 Edge Images: outlines of the image

Edge Image
Fill Image
1028:PICARSO
Ian Hooi
27

Converted to binary form
 Threshold filter
1028:PICARSO
Original
Binary: Threshold = 0.25
Binary: Threshold = 0.5
Binary: Threshold = 0.75
Ian Hooi
28
Raster: pixel by pixel approach
 Vector: line based approach

Raster Based Output
Vector Based Output
Optimal
70
70 70
65 65
60
60 60
55 55
50
50 50
45 45
40
40 40
35 35
30
30 30
25 25
20
20 20
10
1028:PICARSO
20
30
40
50
60
Ian Hooi
10 10
20 20
30 30
40 40
50 50
60 60
29
Image output in binary form

 1 represents white
 0 represents black
Read and processed by Control Software
Slow to paint


70
60
50
40
30
20
1
1
1
0
1
0
1
0
1
1
1
0
1
1
1
1
0
1
1
1
0
0
0
1
0
0
0
1
1
10
0
20
30
Optimal
40
=
50
60
70
65
60
55
50
45
40
35
30
25
20
10
20
30
40
50
60
1028:PICARSO
Ian Hooi
30
Aim: Raster  Vector
Searching Algorithm based on Portrayer
(Benedettelli 2008) and Erik’s XY Plotter (2007)
Adjacent pixels  chains  Control software



Optimal
70
60
50
40
=
30
20
10
20
10
20
30
40
50
60
70
65
60
55
50
45
40
35
30
25
20
30
1028:PICARSO
40
50
60
Ian Hooi
31
Hardware
Design
Problem
Painting
System
Project
Outcomes
Design
Specifications
Testing and
Results
Image
Processing
Future Work
System
Architecture
Control
Software
1028:PICARSO
Joyce Phan
32
Goals:
 Software for Raster Mode
 Convert Image Processing output to:
• Control motors
• Control spray gun
Extension Goals:


Software for Vector Mode
Graphical User Interface (GUI)
1028:PICARSO
Joyce Phan
33
Image
Processing
Output
1 0
0 1
Control
Software
Positioning
Commands
x
y
Cartesian
Co-ordinates
Inverse
Kinematics
L1 , L 2 , L 3
Cable
Lengths
turns
Output
Commands
1028:PICARSO
Joyce Phan
on
off
Motor Turn
Units
Motor
Commands
Spray Gun
Commands
34
0
1
1
1
0
1
0
1
0
1
1
1
0
1
1
1
1
0
1
1
1
0
0
0
1
0
0
0
1
1
Image Processing
Output
1028:PICARSO
Raster Mode
Joyce Phan
Vector Mode
35
1028:PICARSO
Joyce Phan
36
Motor Controller 3 Slave
Motor Controller 2 Slave
Solenoid
RS232
PC Master
RS232
RS232
Digital Output
• Instructions
from Master to
Slaves via 3
Parallel RS232
links
• Outputs
controlled in
Maxon RS232
Communication
Protocol
Motor Controller 1 Slave
1028:PICARSO
Joyce Phan
37
Position Mode
Motor 3
Position Mode
Motor 2
• Position Mode
• Driven in steps
• Current Mode
• Provides tension
• Minimises instabilities
Motor 1
Current Mode
1028:PICARSO
Joyce Phan
38
• Easy access to user settings during operation
1028:PICARSO
Joyce Phan
39
Hardware
Design
Problem
Painting
System
Testing
and
Results
Design
Specifications
Image
Processing
Project
Outcomes
Future Work
System
Architecture
Control
Software
1028:PICARSO
Sven Paschburg
40
Goals:
 Scaled system – µAngelo
 Full scale system – PICARSO
 Image processing software
 Control software
 Graphical User Interface (GUI)
Extension Goals:
 Vector-based painting
 Touch screen interface
1028:PICARSO
Sven Paschburg
41
Scaled System - µAngelo
 Kinematics test bed
 Tri-motor Y-configuration proof of concept
Front view picture of
µAngelo
1028:PICARSO
Oblique angle picture
of µAngelo
Sven Paschburg
Picture of µAngelo’s
end-effector
42
Full Scale System - PICARSO
 Raster painting functionality
 Scalable across a vertical surface
• Up to 3×3m workspace area
 Complete a picture in 1 hour
Test Metrics
 Accuracy & Precision
 Pixel Size
 Reliability
 Stability
 Workspace Resolution
 Speed
1028:PICARSO
Sven Paschburg
43
Raster Painting Functionality
10 mm
100 mm
1028:PICARSO
50 mm
Specifications
Achievement
Accuracy
± 10 mm
Precision
± 5 mm
Workspace
Resolution
10 – 100 mm
Workspace
Resolution (ideal)
10 – 25 mm
25 mm
Sven Paschburg
44
Raster Painting Functionality
Parameter
Achievement
Pixel Size (min.)
8 mm
Pixel Size (ideal)
10 mm
0.2 s
0.25 s
0.5 s
1.0 s
8 mm
10 mm
16 mm
20 mm
1028:PICARSO
Sven Paschburg
45
Raster Painting Functionality
Specifications
Achievement
Stability
± 25 mm (z – axis)
z
pitch
roll
yaw
Bottom view of the end-effector
Side view of the end-effector
1028:PICARSO
Sven Paschburg
46
Scalable across a vertical surface
Specifications
Achievement
Workspace Area 0.5 × 0.5 m to
7.0 × 7.0 m
A picture showing the ability of the
end-effector to move around the
workspace
1028:PICARSO
Sven Paschburg
47
Goals:
 Scaled system – µAngelo
 Full scale system – PICARSO
 Image processing software
 Control software
 Graphical User Interface (GUI)
Oblique angle picture
of painted
µAngelo‘fills’ image
A 1.8 × 1.8m
of the Mona Lisa
1028:PICARSO
Sven Paschburg
48
Extension Goals:
 Analog communication
 Complete a picture in 1 hour
 Vector-based painting
 Touch screen interface
Future years:
 Colour painting
 Wireless communication
 Commercial product
1028:PICARSO
Sven Paschburg
49
1028:PICARSO
Sven Paschburg
50
Benedettelli, D. 2008, NXT Portrayer
Robot, viewed 11th April 2010
<http://robotics.benedettelli.com/portrayer.h
tm>
 Convict Episcopal de Luxembourg, 2007,
Erik’s XY-Plotter, viewed 11th April 2010,
<http://www.convict.lu/Jeunes/ultimate_stuf
f/Erik_s_xx_plotter/E_xy_plotter.htm>.
 Franke, U & Lehni, J 2002, Hektor, viewed
29 December 2009,
<http://www.hektor.ch/Book/Hektor.pdf/>.

1028:PICARSO
Sven Paschburg
51
1.
2.
3.
4.
5.
6.
7.
8.
9.
Spray can - http://comps.fotosearch.com/comp/UNN/UNN113/spraycan_~u14072509.jpg
Airbrush http://imgs.tootoo.com/e1/00/e10006f419523aa3bb036a77fb5038dd.jpg
Spray gun - http://image.made-in-china.com/2f0j00DaTQeBfgIEMJ/HVLPSpray-Gun-RP8021-2S-H827-S-.jpg
Air Compressor - http://toolmonger.com/wpcontent/uploads/2006/05/aircompressor.jpg
Festo Air Regulator - http://www.festo.com/rep/nlbe_be/assets/LR_2008_1314u.jpg
Pressurised Paint Canister http://www.smitsgroup.co.nz/images/objectimages/PT2-Paint-Tank.jpg
Festo Solenoid Valve http://www.luconda.com/artikeldetails/27/29/xx/bilder/2093457-1-Festo-CPE18M1H-3GL-1-4.jpg
Iwata Spray Gun - http://www.anestiwata.co.jp/english/products/paint/prd/fog/auto/images/ceramic_ph01.jpg
Personal Computer - http://www.ubergizmo.com/photos/2008/12/dell-xps-13.jpg
1028:PICARSO
Sven Paschburg
52
53
PICARSO
Mechanical &
Electrical
Hardware
1028:PICARSO
Software
Painting
System
Image
Processing
Samuel Oosterholt
Control
Software
54
1028:PICARSO
Sven Paschburg
55
1028:PICARSO
Joyce Phan
56
1028:PICARSO
Joyce Phan
57
1028:PICARSO
Joyce Phan
58
1028:PICARSO
Ian Hooi
59
PC Master
EPOS2 Slave (Node 1)
EPOS2 Slave (Node 2)
NI PCI-6221 card NI SB 68-LP
(Analogue Inputs) Breakout Board
1028:PICARSO
Joyce Phan
EPOS2 Slave (Node 3)
60
1028:PICARSO
Joyce Phan
61
1028:PICARSO
Joyce Phan
62
1028:PICARSO
Joyce Phan
63
1028:PICARSO
Sven Paschburg
64
1028:PICARSO
Sven Paschburg
65
1028:PICARSO
Sven Paschburg
66
1028:PICARSO
Sven Paschburg
67
Image Processing Software
 Raster processing functionality
 Process an image in less than 60s
 Vector processing functionality
Processing Settings
Options
Image Input
JPEG, PNG, GIF, BMP, TIFF
Image Resolution
10 – 500 pixels
Image Threshold
0–1
Conversion Type
Fills, Edges
Palette Type
Binary (Black & White)
Sizing
Square, Original Aspect Ratio
Image Processing Time
<3s
1028:PICARSO
Sven Paschburg
68
High Level Control Software
 Raster functionality
 Human Machine Interface (HMI)
 Vector functionality
 Touch screen interface
Specifications
Options
Control Mode
Raster Painting,
Vector Painting,
Manual Positioning,
Spray Control
Platform
Mathworks MATLAB 2007 +
Operating System
Windows XP, Vista, 7
Interface Type
Graphical (Touch Screen)
1028:PICARSO
Sven Paschburg
69
Low Level Control Software
 Position the motors
 Control the motor velocity and acceleration
 Read back the current position
 Actuate the spray gun solenoid
Specifications
Options
Control Functions
Position Control,
Set Max. Velocity & Acceleration,
Read Current Position,
Activate Digital Output
Platform
Mathworks MATLAB 2007 +
Operating System
Windows XP, Vista, 7
Positioning Accuracy (theoretical) 2.11 × 10-3 mm
1028:PICARSO
Sven Paschburg
70
Complete a picture in 1 hour
A 1.8 × 1.8m painted ‘fills’ image
of the Mona Lisa
1028:PICARSO
Parameter
Specification
Speed
2 – 250 mms-1
Acceleration
2 – 150 mms-2
Picture Time (av.)
2.5 hrs (edge image)
A 1 × 1m painted ‘edge’ image
of the University of Adelaide logo
Sven Paschburg
71

similar documents