High-Performance Computing Applications and Networking ... · Becoming technological advanced...
Transcript of High-Performance Computing Applications and Networking ... · Becoming technological advanced...
![Page 1: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/1.jpg)
Joe-Air Jiang, Ph.D., P.Eng.Distinguished Professor, Department of Bio-Industrial Mechatronics Engineering, NTUDirector, Education and Research Center for Bio-Industrial Automation, NTUChairman, Taiwan Institute of Biological Mechatronics
Becoming technological advanced — Case studies of IoT applications in smart areas
Date:2016/12/5
High-Performance Computing Applications and Networking Technology
Southeast Asia International Joint Research and
Collaboration Program (SEAIP2016)
![Page 2: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/2.jpg)
Colony CollapseDisorder
Remote Monitoring System of Pest
Coverage Optimization
AutomaticGreenhouse
System integration
Electrical engineering
Optimization
Wireless communication
InstrumentDegradationIdentification
Remote Visual Monitoring System (tern)
Plant Factory
Smart City Smart Gird
Smart Agriculture
Remote MonitoringSystem of Pests Field info. Monitoring Automatic pest sensor Pest population forecast IPM applications
Remote Visual Monitoring System Surveillance of critically
endangered species of tern Real-time imaging for tern Energy-efficiency Autonomous operation Optimal scheduling
Automatic Greenhouse Adaptive environment control Pest monitoring Increase productivity of orchid Productivity forecast
Colony Collapse Disorder Platform for CCD studies In-out analysis of honey bees Bee behavior analysis
Smart Grid
Monitoring of EHV Power Transmission Lines System Wide-range measurement of
conductor temperature, line sag, etc.
Conductor temperature forecast Dynamic thermal rating Ampacity margin estimation of
EHV lines
Smart CityMicro-Sensing Network Sensors for intelligent space Location awareness technology Disease transmission forecast
Air-quality Surveillance High spatiotemporal resolution Pollution spread analysis
Tech. Contents
Instrument Degradation Identification & Data Healing Big data mining Anomaly detection Autonomous distributed operation Critical data recovering
Coverage Optimization Improve quality of surveillance Scheduling of sensors for saving
energy
Localization Indoor/outdoor RF-based
localization with high accuracy
Mechatronics
Signal processing Machine
learning Data miningStatistics
RF-based Localization
Microwave Pest Killing Machine Low-cost, high efficiency magnetic
applicator Non-destructive operation Dosage analysis
Optimal Scheduling
Department of Bio-lndustrial Mechatronics Engineering,National Taiwan University
Measured Data Healing
Background & Research Scope
Bioelectromagnetics Lab.
Lab leader: Joe-Air Jiang, Ph.D., Distinguished Professor
![Page 3: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/3.jpg)
Outline• Introduction
– What is IoT?
• Advanced IoT/WSN InfoCom Platform• Cases studies of IoT applications in different scenarios
– IoT/WSN-based pest monitoring system
– Intelligent IoT/WSN-based monitoring system for greenhouse
– IoT/WSN-based experimental platform for colony collapse
disorder (CCD) of honey bees
– IoT/WSN-based PV monitoring system
– IoT/WSN-based grid-wide safety monitoring system for EHV
grids
– IoT/LoRa-based city-wide monitoring system for vehicles
• Q & A
![Page 4: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/4.jpg)
WHAT Is IoT?
Internet of Things
![Page 5: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/5.jpg)
Physical objects can communicate
with each other
IoT is interested in collecting data
using a large network consisted of
physical sensors and analyzing the
collected data to build a service or
flow model for target applications.
Object tags can be identified, located,
and enabled/disabled
All the objects can assess to the
Internet
Data modeling and scenario
detection are the goals.
![Page 6: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/6.jpg)
MSP-basedEmbedded-based
IPC-based
Sensing layerData aggregation
layerNetwork
layerData analysis & cloud
service layer
Advanced IoT/WSN InfoCom Platform
Intelligent pest monitoring andmodeling
Ambient temp./humidityillumination
Pest number
Low power consumption ecological monitoring
Ambient temp./humidity
Orchid greenhouse temp./humidity
Intelligent and automatic greenhouse monitoring system
Ambient temp./humidity Beehive temp./humidity
CCD research based on IoT/WSN
Bee countsincoming/outgoing
Movable bencheslocalization
Ecological image
Bee behavior monitoring system
CCD research based on IoT/WSN
Intelligent and automatic greenhouse monitoringsystem
Orchid quality image monitoring Orchid greenhouse pest monitoring
Low power consumption ecological monitoring
Matsu tern monitoring
Pest outbreak
Data savingFunctions Intelligent pest monitoring
and modeling
Pest analysis of monitoring area
GSM
![Page 7: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/7.jpg)
IoT/WSN-based Pest Monitoring SystemDevelopment of a remote agro-ecolocigal monitoring system
>>
ENTER
![Page 8: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/8.jpg)
STATUS QUO
Huge agricultural damage
Oriental Fruit Fly (Bactrocera dorsalis (Hende))
Tobacco Cutworm (Spodoptera litura (Fabricius))
About 570 million USDannually in Taiwan
Main targets
Traditional pest monitoring techniques heavily depend on manual measurement
Expensive, labor-intensive
Time consuming (10 days) & inefficient counting
No synchronized measurements on weather factors
![Page 9: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/9.jpg)
9
WSN-based Monitoring Station
Gateway
GPSSatellite
3G Base Station
Web &
analysis
serverInternet
FarmersGovernment
Early warning of
pest outbreaks
7-day forecast for pest
population dynamics
Mediterranean
fruit fly
Oriental fruit fly Tobacco cutworm
IoT/WSN-based Remote Pest Monitoring System
Objectives:• Weather durability
• Scalability
• Reliability
• Remote control
• Real-time warning
• 7-day forecast
Challenges:• Network reliability
• Remote accessibility
• Network scalability
![Page 10: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/10.jpg)
|2011 today|
94 sensors for TC
151 sensors for OFF
2 standalone
stations
22 WSN-based
monitoring
stations
245 sensors
All for OFF
10 Gateways for TC
12 Gateways for OFF
Oriental Fruit Fly Tobacco Cutworm
Wild Field Monitoring System
![Page 11: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/11.jpg)
Sensinglayer
Transmissionlayer
Applicationlayer
• Sensor node device
12
Wild Field Monitoring System
![Page 12: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/12.jpg)
• Gateway platform
13
Sensinglayer
Transmission layer
Application layer
Wild Field Monitoring System
Embeddedsystem-based
IPC-based
Gateway
MSP430-based
![Page 13: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/13.jpg)
• Pest analysis ofmonitoring area– Farmers can bagging the fruit
based on the analysis results
14
Sensinglayer
Transmissionlayer
Applicationlayer
Wild Field Monitoring System
![Page 14: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/14.jpg)
• Pest early alertingservice
15
Sensinglayer
Transmissionlayer
Applicationlayer
Wild Field Monitoring System
![Page 15: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/15.jpg)
IoT/WSN-based pest monitoring system
![Page 16: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/16.jpg)
Intelligent IoT/WSN-based GreenhouseDevelopment of an IoT/WSN-based Automated Greenhouse Monitoring System
>>
ENTER
![Page 17: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/17.jpg)
BACKGROUND INFO
Taiwan is the top 2 orchid exporting country world-wide (behind the Netherlands), and 1 out of 6 orchids in the world is produced in Taiwan.
Taiwan has a critical advantage in the orchid market of the world – orchid diversity.
Orchid is one of the essentialcommercial crops in Taiwan.
The export value of the orchidreached 195 million US dollarsin 2012 solely in Taiwan, whichaccounted for 93% of the totalexports of flowers from theentire country.
![Page 18: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/18.jpg)
STATUS QUO
Monitoring technology
The monitoring technology used in greenhouses was with low resolution in the past, which hindered the chance of achieving optimal growth conditions.
temperature diversity at different locations > 3 ℃
![Page 19: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/19.jpg)
• System architecture
M2M gatewayEthernet hub
for high-speed data
exchanging
Embedded systems for enhancing
local analyzing capability
(connected with private cloud)
Large-scale greenhouse
Fixed node
Mobile node
Storage
serviceAnalysis
service
Decision support for
greenhouse owners
Cloud
service
Sensor node localization
Dynamic network topology management
Cloud computing capability
A generic IoT/WSN framework
Automatic Greenhouse Monitoring System
![Page 20: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/20.jpg)
Intel-NTU Connected Context Computing Center
Intel-NTU Connected Context Computing Center 21
Greenhouse DeploymentsAutomatic Greenhouse Monitoring System (2013 ~)
‣ Mayshan (flowering)• From July 15 to Nov. 3, 2014
• 22 Wi-Fi cameras
(Morphological trait inspection)
• 20 sensor nodes
(Environment Monitoring)
The deployment
scenes in Taiwan
‣ Dalin (middle seedling)• Since March, 2012
• 58 Fixed Nodes
Anchor for localization
• 62 Mobile nodes
Dynamic network
topology
Target for localization
• 5 PANS
Centralized/distributed
network managements
‣ Gukang (large seedling)• From March 13 to July 15, 2014
• 48 Wi-Fi cameras
(Morphological trait inspection)
• 30 sensors nodes
(Environment Monitoring)
![Page 21: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/21.jpg)
‣ Gukeng greenhouse
(mature in 3.5 inch)• From 3/31~till now• 48 Wi-Fi cam
(Growth status monitoring )• 24 Wireless sensor nodes
(Environment monitoring)
Total: 288 plant
Real deployment statusWater wall
Fan
Office
80 m
140 m
R1
R2
R3
R4
61
Intelligent IoT/WSN-based Greenhouse
![Page 22: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/22.jpg)
23
CPU
ARM1176JZF-S
Camera
Soc
BCM2835
Wireless
LAN Card
GPIO
Relay5V Power
Supply
USB hub
Photograph
Command
MotorLED12V Power
Supply
Command
Turn on/off
Data
Data
Automated pest (e.g. Bradysia) monitoring system of orchidgreenhouse
Intelligent IoT/WSN-based Greenhouse
![Page 23: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/23.jpg)
Accuracy test of the automated pest monitoringsystem in orchid greenhouse
– Sampling time: Aug. to Nov., 2015.
– Sample : 95 monitoring images
– RMSE: 2.06
– Relative Error: 4.91 %
– Margin of error: 0.86 %
Count
Val
ue
Actual Value
y = 1.021x + 0.0968
R² = 0.9951
0
10
20
30
40
50
60
70
80
0 10 20 30 40 50 60 70 800
10
20
30
40
50
60
70
80
8/1
6
8/1
9
8/2
2
8/2
5
8/2
8
9/1
9/4
9/7
9/1
0
9/1
7
9/2
0
9/2
3
9/2
6
10
/1
10
/10
10
/13
10
/16
10
/19
10
/22
10
/26
10
/29
11
/1
11
/5
11
/8
11
/11
11
/16
11
/19
11
/22
11
/25
11
/28
Actual value Count value
Intelligent IoTWSN-based Greenhouse
![Page 24: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/24.jpg)
Gukang large seedling area
• 48 WiFi cameras
• 30 sensors nodes (tmp/hum/light)
• 15 advanced soil sensors
• 4 advanced light sensors
Mayshan flowering area
• 22 WiFi cameras
• 20 sensor nodes (tmp/hum/light)
• 6 advanced soil sensors
• 2 advanced light sensors
Real time monitoring – Time-lapse photography
63
Intelligent IoT/WSN-based Greenhouse
![Page 25: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/25.jpg)
Leading overlapping issue between each orchid plants
Using the symmetry of orchid leaf to estimate theoverlapping area
62
Leaf overlapping estimation and reconstruction
Intelligent IoT/WSN-based Greenhouse
![Page 26: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/26.jpg)
Leaf overlapping estimation and reconstruction Morphological traits + HSV + Canny edge detection
An increase in analyzed samples The estimation error of the non-overlapping image: 0.67% ~ 3.06%
The estimation error done by Huang and Lin is between 1.8% ~
4.02% in 2002, and done by Weng is between 5% ~ 8% in 2012
Overlapping
image
Processed
image of
overlapping
leaves
Non-
overlapping
image
Processed
image of non-
overlapping
leaves
Pixelrepainting
methodPixel Error
1 160379 147388 8%
2 160379 158010 1%
3 168837 163648 3%
Huang, K.Y., T.C. Lin , J.N. Tsai, 2002, DISEASE DETECTION AND CLASSIFICATION IN PHALAENOPSIS SEEDLINGS USING MACHINE VISION, Agricultural Engineering Journal 11(1):11-30. (EI) 64
Intelligent IoT/WSN-based Greenhouse
![Page 27: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/27.jpg)
Intel-NTU Connected Context Computing Center
Intel-NTU Connected Context Computing Center 28
Technological Accomplishments
Impacts of environmental factors
Experimental period: April 19 ~ July 4, 2015
Region Avg. growth of leaf area (mm2/day)
1 79.41
2 75.46
3 74.15
4 57.68
Region T ± STD. (˚C) H ± STD. (%RH)
1 28.83 ± 2.58 71.81 ± 8.88
2 30.57 ± 3.45 71.43 ± 11.04
3 30.01 ± 3.18 72.17 ± 10.06
4 28.59 ± 2.76 79.95 ± 9.13
![Page 28: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/28.jpg)
Intel-NTU Connected Context Computing Center
Intel-NTU Connected Context Computing Center 29
Technological AccomplishmentsRelationship between blossom quality and exterior
characteristics
4 6 17 18 24 3 23 14 25 11 26 27 29 21 9 12 7 2 13 8 30 10 15 19 20 22 28 1 5 160
200
400
600
800
1000
Hierarchical clustering
The highest class of orchids
• The length of the stalk > 20 cm
• The number of blossoms > 8
• The area of the leaf > 35000 mm2
The lowest class of orchids
• The length of the stalk < 15 cm
• The number of blossoms < 7
• The area of the leaf < 28000 mm2
The most significant boundaries
of exterior charactersA+ A B B-
![Page 29: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/29.jpg)
Intel-NTU Connected Context Computing Center
Intel-NTU Connected Context Computing Center 30
Key lessonsAutomatic Greenhouse Monitoring System (2014 ~)
40 605030 7020100
Days
26000
36000
34000
32000
30000
28000
Le
af
Are
a (
mm
2)
24000
22000
Growth Data of R1
Growth Data of R2
Growth Data of R3
Growth Data of R4
Fitted Curve of R1Fitted Curve of R2
Fitted Curve of R3
Fitted Curve of R4
Region Blossom on phase I (A/B/G/N.A.*) Blossom on phase II (A/B/F/N.A.*)
1 23/9/24/16 33/19/4/16
2 26/8/25/13 41/16/2/13
3 11/9/33/19 25/20/8/19
4 4/0/36/32 16/11/13/32
![Page 30: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/30.jpg)
Intelligent IoT/WSN-based Greenhouse
(Courtesy of Intel-NTU CCC Center
![Page 31: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/31.jpg)
Colony Collapse DisorderDevelopment of an IoT/WSN-based experimental platform for
electrophysiological and behavioral study of honey bees>>
ENTER
![Page 32: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/32.jpg)
STATUS QUO
Colony Collapse Disorder
CCD
• 66% crops in the world are pollinated byhoneybees
• The total U.S. crop value that was whollydependent on honeybee pollination wasestimated to exceed $ 15 billions
Sudden disappearance of honeybees
http://sonomabees.org/MembersOnly/general/ http://daily-free-takeout.blogspot.tw/2012_06_01_archive.html
![Page 33: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/33.jpg)
STATUS QUO
• Malnutrition – GM corns or a single crop infarm
• Virus• Electromagnetic radiation
• Pesticide• Environmental/climate change
Possiblereasons of CCD
Our concerns
Colony Collapse Disorder
CCD
![Page 34: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/34.jpg)
Architecture of the automatic foragingbehavior monitoring system
![Page 35: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/35.jpg)
Practical setup and deployment
MAIN ACHIVEMENTS
Architecture of the counting system
![Page 36: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/36.jpg)
Bees will drag the deadbodies out from the beehive to avoid the entirehive collapse via the virus
QUANTITATIVE IMPACT
Practical operation
Architecture of the counting system
![Page 37: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/37.jpg)
Location: Tomatake Hall, NTU
Time: 2015/12/22
Weather: Temp. 21.8 ~ 29.6 C Temp. relative higher
Accuracy test of the automatic foragingbehavior monitoring system
54
Result Automatic foraging behavior monitoring system
Outgoing: 4348 Incoming: 4713
Human counting Outgoing: 4657 Incoming: 5061
Counting error Outgoing: 11.08% Incoming: 11.24%
![Page 38: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/38.jpg)
54
Regression analysis of automatic foraging behaviormonitoring system: accuracy test/counting error
System avg. accuracy Income: 84.92% Outgoing: 85.95%
R2
Incoming: 0.924 Outgoing: 0.937
Accuracy test of the automatic foragingbehavior monitoring system
![Page 39: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/39.jpg)
QUANTITATIVE IMPACT
Long term monitoring for the environmental factor profiles - Ankong Farm of NTU
The frequency of honey bees’incoming activities is similar tooutgoing activities
The daily difference between theoutgoing and incoming counting ofhoney bees was lesser than 300counts/day
The temperatures near at the nectarregion inside the beehive wereranged from 18 to 34 C
The average values and standarddeviations of the relative humidity Inside the beehives: 74.3 ± 11.6
%RH outside the beehives: 85.1 ±
4.88 %RH
![Page 40: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/40.jpg)
QUANTITATIVE IMPACT
Coming-in and going-out activities of honey on sunny day (March 27, 2014)
30
40
50
60
70
80
90
100
Rel
ativ
e h
um
idit
y (
%)
Beehive relative humidity Ambient relative humidity
15
20
25
30
35
40
Tem
per
atu
re (℃
)
Beehive temperature Ambient temperature
0
10
20
30
40
50
60
Nu
mb
er o
f b
ees'
in
an
d o
ut
In Out
Honey bees are more activebetween 5:00 a.m. and 6:00 p.m
The total frequency of honey bees’incoming behavior was 81,912, andthe total frequency of honey bees’outgoing behavior was 83,347
The trend of the ambienttemperature was also similar to thebeehive temperature
The trend of the ambient relativehumidity was not similar to that ofthe beehive relative humidity Inside the beehives: 74.0 % RH
to 83.0 %RH Outside the beehives: 38.0
%RH to 90.0 %RH
Ankong Farm of National Taiwan University (NTU)
![Page 41: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/41.jpg)
42
Time Delay Neural Networks
Prediction of in-and-out activities of honey bees
![Page 42: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/42.jpg)
The Results of TDNNs43
2014/9/1
Actual 29327
Predict 29661
APE 38
RMSE 0.0032
![Page 43: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/43.jpg)
44
The Results of TDNNs
2014/9/1
![Page 44: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/44.jpg)
END GOAL
Find the reasons why CCD happened
Providing a micro-climate monitoring system for investigating honeybee’s behavior
• A stable system is necessary• Using the system to investigate the impact of pesticides and
the electromagnetic radiation on honeybees
The system integrates an IoT/ WSN-based ecological monitoring system and a counting system
By the sensing data, modeling the dynamic behavior of bees is possible
• Predicting colony of honey bees if a beehive is going to be better or worse
• Finding out the causes for the CCD of honey bees
![Page 45: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/45.jpg)
Automatic foraging behavior monitoring system
![Page 46: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/46.jpg)
>>
ENTER
汎宇 曜得資訊Cloud
Computing
An IoT/WSN-based Monitoring System for PV Generation Plant
>>
ENTER
![Page 47: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/47.jpg)
In Taiwan, most of PV generation plants aresmall and medium-size (generation capacitiesare about 200 to 300 kW, plant covering areabout 2000 to 9000 m2). Using cablemonitoring system to detect multiple signals,such as voltage, current, temperature,humidity, illumination and other sensingsignals is difficult. Because each signal isindependent of individual signal lines, the costof wire and construction are very expensive.
BACKGROUND INFO
WSN has the great convenience and low costof deployment. By using different number ofsensing nodes, the sensing range can beflexible
The selection of sensors are very customized.Industries can choose many kinds of sensorsaccording to their considering issue, like usingthermometers to study the relationshipbetween module temperature and powergeneration.
![Page 48: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/48.jpg)
• System architecture
IoT/WSN-based monitoring system for PV generation
Users
Lx.
RMG
3G Base Station
Temp. & Hum.
Vol. & Cur.
Database
MCMSWCM WCM WCM
Solar Cell Solar Cell Solar Cell
MCMS: microcontroller with multiple sensors
WCM: wireless communication module
RMG: remote monitoring gateway
![Page 49: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/49.jpg)
Wireless Sensor Node
Actual Photo
Wireless
Transmissi
on Module
Time
Module
ATmega32P-PU
Galvanometer
Illuminometer
Thermometer
VoltmeterInstalled thermometer and illuminometer with transparent shell
IoT/WSN-based monitoring system for PV generation
![Page 50: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/50.jpg)
• Real deployment status
WSN on Photovoltaic Monitoring SystemIoT/WSN-based monitoring system for PV generation
![Page 51: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/51.jpg)
Voltage Synchronizing Technique for Solar System
IoT/WSN-based monitoring system for PV generation
![Page 52: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/52.jpg)
Surface Temperature Estimation and Cooling Technology for Photovoltaic Modules
Gateway
Thermometer
Pyranometer
Anemometer
PV modules with cooling systems
Wireless node
Water tank
Pump
Spout
Backend server
User
Data mining
IoT/WSN-based monitoring system for PV generation
![Page 53: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/53.jpg)
IoT/WSN-based monitoring system for PV generation
![Page 54: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/54.jpg)
智慧型超高壓輸電線安全性監測系統Development of an IoT/WSN-based Grid-Wide Safety Monitoring
System for EHV Power Grids
>>
ENTER
![Page 55: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/55.jpg)
BACKGROUND INFO
Among the power grids with differentvoltage levels, extra-high voltage (EHV)transmission grid serves as an importantrole of the entire power system, just likemain artery
These facts further caused the power gridsto face severe challenges of reliability,security, efficiency, etc
To meet the increasing energy demand,power companies have to supply moreelectricity to end users so that their powergrids become more and more complex
Large power flow of power grid usuallycauses high line conductor temperature andsevere line sag and then might induce lineground fault
![Page 56: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/56.jpg)
STATUS QUO
Monitoring technology
To certify the availability and safety for powergrids, many utilities adopted manual-inspectionapproach to check power grids, which is aninefficient, high-cost, and labor-intensiveinspection method
![Page 57: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/57.jpg)
Real deployment
IoT/WSN-based Grid-Wide Safety Monitoring System
![Page 58: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/58.jpg)
Real deployment status
IoT/WSN-based Grid-Wide Safety Monitoring System
![Page 59: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/59.jpg)
Demo video
IoT/WSN-based Grid-Wide Safety Monitoring System
![Page 60: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/60.jpg)
IoT/LoRa-based Smart CityDevelopment of an IoT/LoRa-based city-wide
monitoring system for vehicles>>
ENTER
![Page 61: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/61.jpg)
106,078,000
In 2008 year, IBM proposed the concept of smart planet, which
then lead to smart city.
Smart city is an ideal urban vision by integrating information and
communication technology (ICT) and Internet of
Things(IoT).
Vehicles can provide real-time information about the surrounding
environment.
This research developed a vehicle monitoring system based on the
IoT/LoRa technology.
Traffic Jam
Heat Island
Air Pollution
Background & Motivation
![Page 62: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/62.jpg)
In 2016, Taipei City Government established the Taipei LoRa platform.
◎ The first citywide LoRa platform in Asia
◎ The communication range up to 15-20 kilometers
◎ Only 11 LoRa base stations needed to cover Taipei City
Taipei IoT Experimental Platform
![Page 63: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/63.jpg)
• 12 VDC• Timekeeping Chip• SD card support
(data redundancy)
• LoRa• OBD II• RS-232• USB port• Uart port *2
Technical specs
Interface
Sensor
• SO2• CO• O3• NO2
• TSP• PM 2.5• PM 10
• GPS
• Accelerometer
• Temperature• Humidity
Vehicle-box (V-box)The device integrated with LoRa and sensors
![Page 64: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/64.jpg)
0m200m600m
90 sec
1 cycle0 sec30 sec60 sec
20 packets per 200 meters
Limit: 1 packet per minutes
800m400m
Results – Packet reception ratio (PRR) test for LoRa
![Page 65: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/65.jpg)
100% 100%
80%
70%
50%
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0m 200m 400m 600m 800m
PP
R(%
)
Distance (m)
Results - Packet reception ratio (PRR)
![Page 66: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/66.jpg)
Personal Interface
Air qualityVehicle
information
Established by IBM Bluemix
GPS
Weather
Place your screenshot here
Pla
ce y
our
scre
enshot
here
User Interface for IoT/LoRa-based city-wide
monitoring system
![Page 67: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/67.jpg)
Demo video for testing IoT/LoRa-based city-
wide monitoring system
![Page 68: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/68.jpg)
Personal InterfaceFuture Smart Services & Applications by IoVs
Environment
Medicine
Traffic
• Congestion Predict• Road Pothole Detection• Traffic Planning
• Disease Spreading model• Alarm for Hospital • Medical planning
• Air Quality Standard• Alarm for Air Pollution• Environmental planning
![Page 69: High-Performance Computing Applications and Networking ... · Becoming technological advanced —Case studies of IoT applications in smart areas Date:2016/12/5 High-Performance](https://reader033.fdocuments.in/reader033/viewer/2022050222/5f677d655424d3097e4a090d/html5/thumbnails/69.jpg)
END GOAL
Internet of Things (IoT)makes every thing possible
Thanks for your
attention!