Experiences and Best Practices with NI Platform in a ...
Transcript of Experiences and Best Practices with NI Platform in a ...
Experiences and Best Practices with NI Platform
in a Control Laboratory
Dr. Withit Chatlatanagulchai
Department of Mechanical Engineering
Kasetsart University, Bang Khen Campus
Bangkok, Thailand
1/36
Professor’s Responsibilities
• Teaching
• Research
• Academic services
2/36
Teaching with NI Platform
• Introduction of university, faculty of engineering (Labviewsite license)
• Introduction of laboratory (How did we learn Labview?, Labview ambassador)
• Our NI platform in teaching (USB DAQ, MyRIO)
• 208587 grad-level advanced automatic control class
• Expectation in the near future• MSME in robotics (208579 grad-level instruments and signal
processing class)• Undergrad teaching (208371 3rd-year automatic control class,
208322 3rd-year mechanical vibration class)• NI Labview Academy• MOU with NI
3/36
Research with NI Platform
• (2006-Present) Advanced techniques in automatic control
• (2006-2011) Engine control
• (2012-2016) Vibration reduction with input shaping
• (2017-Present) Deep learning, Machine learning
4/36
Academic Services with NI Platform• Present
• 3D Robo Farmer• Remote weapon station• Data center• Smart camera for elderly• Process control test-rig• Train simulator and train model• Wave-compensated seat• Cement process control (Clinker, Kiln)• Plastic injection machine retrofit• Rubber coagulation oven
• (Near) future• Precision farming from satellite image • Smart building• Electric motorcycle• Self-driving car
5/36
Teaching: Introduction of university, faculty of engineering
• Originated in 1904 (113 years)
• 4 main campuses
87%
11%2%
72,600 students
Undergrad Master PhD
Kasetsart University
Source: www.regis.ku.ac.th/cpcmns/rpt_std_ku3.php
6/36
Teaching: Introduction of university, faculty of engineering
• Originated in 1938 (79 years)
87%
11%2%
13,305 students
Undergrad Master PhDFaculty of Engineering
Source: www.regis.ku.ac.th/cpcmns/rpt_std_ku3.php
7/36
Teaching: Introduction of university, faculty of engineering
31
4
71
4
51
2
12
84
15
9
18
0
34
5
31
0
42
3
43
6
UNDERGRAD
Number of engineering students; Bang Khen campus; year 2017; total 6,369 persons
11
13
4
20
0
19
1
18
1
14
6
97
39
39
2
89
MASTER
0
31
26
36
19
8
39
19
17
17
PHD
Source: Educational service department, Faculty of engineering, Bang Khen campus
8/36
Teaching: Introduction of university, faculty of engineering
17
47
29
45
16
16
26
20
44
34
FULL-TIME FACULTY
Bang Khen campus; year 2017• Faculty education
• PhD 259
• Master 55
• Bachelor 10
• Faculty rank• Full professor 5
• Associate prof 108
• Assistant prof 109• Lecturer 102
Source: Departments’ websites
Source: http://www.person.ku.ac.th/new_personweb/stat/2560/UOC_STAFF_2_2560.PDF
9/36
Teaching: Introduction of university, faculty of engineering
NI Academic Site License: Faculty of Engineering, Bang Khen campus, year 2017
10/36
Teaching: Introduction of university, faculty of engineering
NI Academic Site License • For all faculty and students
• For teaching and research
• Access to NI Software
• Access to learning materials, online courses, supporting staffs
11/36
Teaching: Introduction of laboratory
• Control of Robot and Vibration Laboratory (CRVLAB)• PhD degrees 9
• Master degrees 30
• Field of interest• Advanced control theory
• Engine control
• Vibration control
• Machine learning
CRVLAB (since 2007)
Source: www.crvlab.com
12/36
Teaching: Introduction of laboratory
• How did we learn Labview?• Books
• 2003 Book by Robert H. Bishop
• In-program help• NI Example Finder• Show Context Help• Labview Help
• www.ni.com• Case studies, community,
discussion forums, knowledge base, webcasts
• Supporting staffs from NI Thailand
13/36
Teaching: Introduction of laboratory
• NI Labview Student Ambassador• Mr. Supparat Damyot
(2nd-year PhD student)
• Mr. Pisit Intarawirat
(2nd-year Master student)
• Responsibilities• Teach series of workshops
• Help prepare for CLAD exam
• Intermediary between university and NI
14/36
Teaching: Our NI platform in teaching
• We have 10 units
• 8 Analog input
• 2 Analog output
• 13 Digital I/O
• 1 Counter (count up only)
• We have 10 units
• 10 Analog input
• 6 Analog output
• 40 Digital I/O
• 4 Counters (quadrature encoder)
• Serial (UART, SPI, I2C)
• 8 PWM
• Wireless
• Accelerometer
USB-6001 NI myRIO
15/36
• 20 Master and PhD students
• 10 experimental sets
• Diesel engine’s throttle control
• USB-6001
Teaching: 208587 grad-level advanced automatic control class
16/36
Teaching: 208587 grad-level advanced automatic control class• PID control
• Sliding-mode control
• Fuzzy logic
• System identification
• Linear state-space methods
• Adaptive control
• Quantitative feedback theory
• Nonlinear control
• Robust multivariable control
• Model predictive control
• Linear matrix inequalities
• Input shaping
User-defined functions
17/36
• MSME in Robotics (2/2017)
• 208579 grad-level instruments and signal processing class, taught by NI engineer
• NI Labview Academy
• MOU with NI
Teaching: Expectations in the near future
18/36
Research: (2006-2011) Engine control
• NI powertrain controls systems for engine control and combustion analysis
• Diesel-dual-fuel engine (Diesel + gas)
• Already a commercial product for passenger truck
• 30% fuel cost saving
• More info at www.crvlab.com
19/36
Research: (2012-2016) Vibration reduction with input shaping
• NI PCI-6221, NI myRIO, NI CompactDAQ
• More info at www.crvlab.com
20/36
Research: (2017-Present) Deep learning, Machine learning
• Facial expression
• Face recognition
• Plant health monitoring
• Driverless car
• Big data
21/36
Academic Services: 3D RoboFarmer
• Toward fully automatic farming: land preparation, planting, caring, harvesting
• myRIO for control, machine vision, and deep learning
22/36
Academic Services: Remote weapon station
• Target seeking
• Target tracking
• Ballistic calculation
• System integration
23/36
Academic Services: Data center
• Energy consumption modeling
• Big data analytic
• Machine learning
• Plans to reduce consumption
Source: Dayarathna et al. (2016)
24/36
Academic Services: Smart camera for elderly
• Machine vision
• Deep learning
• Thermal camera
Detect body on the floor
Identify distress callDetect painful face
25/36
Academic Services: Process control test-rig
• Test-rig for mixing gases and liquids in a petroleum industry
• Connect with chemical process equipment
• Process control
26/36
Academic Services: Train simulator
Monitors
EquipmentsTrainer
computerTrain simulator
computer
LAN
27/36
Academic Services: Train model• Wireshark network
protocol analyser
• Imitate wificommands by Labview
• Bombardier signaling
Trains
28/36
Academic Services: Wave-compensated seat
• Design a new passenger seat
• Compensate pitch and roll movement
• Interface with serial IMU and AC servo motors
Seat
29/36
1h
2h
mh
1x
2x
nx
inputs
th
n
i
nodes
th
m
j
max1/ e
1z
1z
e t
1w
2w
mw
o t
cK
PID
u t
Academic Services: Cement process control
• Cement kiln
• Use deep learning for kiln control
• Model-independent adaptive (MIA) control
Labview DSC module and NI OPC servers
30/36
Academic Services: Plastic injection machine retrofit
• NI to replace outdated controller
• Work with thermocouple, heater, motor, hydraulic, LVDT, and so on
• State chart to replace PLC program
31/36
Academic Services: Rubber coagulation oven
32/36
Future academic Services: Precision farming from satellite image
• Drought prediction with deep learning and machine learning
33/36
Future Academic Services: Smart building
• BACnet, a standard communication protocol for building automation system
• BACnet/IP Protocol for Labview by OvakTechnologies
• No OPC server required
• Direct communication between BACnet devices and NI hardware/Labview
34/36
Future Academic Services: Premium electric motorcycle
35/36
Future Academic Services: Self-driving car
• Semantic segmentation with deep learning
• Path planning, obstacle avoidance, sensors, actuators
Source: Roozbeh, M. et al., CVPR 2014.
36/36