1
Università degli Studi di FirenzeDipartimento di Elettronica e Telecomunicazioni
Virtual Macro Site Analysis using
Wireless Sensor Networks
Davide Di Palma, Francesco Chiti, Gianfranco Manes, Luca Bencini, Giovanni Collodi, Romano Fantacci, Antonio Manes
MIDRA Consortium
WISES08 Sixth Workshop on Intelligent Solutions in Embedded Systems University of Applied Sciences Regensburg, Germany, July 10/11th, 2008
2
Outline
• The old approach to WSN design: let’s start from protocols
• The new approach: application oriented WSN.
• The GoodFood project and the agro-food monitoring
• The “ad-hoc” designed WSN system
• Pre-industrial pilot sites
• New commercial pilot sites
• The macro site analysis approach
3
The old approach to WSN design: let’s start from protocols
Year: 2004
The prevalent and common opinion was finding an infrastructureprotocol, as the best trade off satisfying the widest set of
different application requirements
A wide literature of MAC protocol solutions
Many routing algorithms
Too few “long life” and “reliable” experimental test beds
4
The new approach: application oriented WSN.
Year: 2008
If we want to bring WSN out of the labsinto open fields, WSN must be studied, designed and developed
following the “application oriented approach”.
Dividing different applications into “requirement-homogeneous groups
Designing protocols (and hardware if necessary!) to fulfil only one group’s
requirements
Create an end-to-end system, from the sensor to the end-user
As a result: a real long life, high reliable pilot site can be deployed
How ?
5
The GoodFood project and the agro-food monitoring
Application: Vineyard Monitoring
Slow physical parameters variations.
Wide areas to cover
Unmanned and stand-alone operations
At least one farming season without battery change
Hostile environment
10 countries
FP6-IST-1-508774 GoodFood • Objective: to demonstrate to the agro-food sector actors the advantages driven to the complete food chain control by the use of Micro and Nano - technology inspired systems
• SubTask: to introduce Ambient Intelligence (AmI) paradigms in Agriculture using WSN technology and to implement a demonstrator for the wine chain
Application Specifications
6
The “ad-hoc” designed WSN system
Courtesy of HTA
“ad-hoc” MAC Protocol: STAR MAC “ad-hoc” Dynamic Multihop Protocol
Target Next hop
Hop counter
Phase Link quality
Battery level
Congestion level
X NX ϕ X η X BX CX Sink
Y NY ϕ Y η Y BY CY
“ad-hoc” Hardware Platform: MIDRA Mote
0 25 50 75 [m]
0 25 50 75 100 125 150 175 200 [m]
0 25 50 75 100 125 150 175 [m]
0 25 50 75 [m]
7
• Stand-alone communication platform,
providing transparent bi-directional
wireless TCP/IP connectivity for remote
monitoring.
• Powered by solar panels.
• Improved robustness.
• Reconnection with dynamic IP address
assignment.
The “ad-hoc” designed WSN system
“ad-hoc” Gateway Unit “ad-hoc” End-User Interface
8
The “End-To-End System”
GPRS Link
Server
World Wide Web
9
System Details:
• 1 GPRS Gateway Station• 12 Wireless Sensor Network Nodes• 30 Distributed Sensors• 4 Fully Monitored Plants
The First Pilot site was installed on 2005 in one hectare of Vineyard on a mild Chianti hill at San Casciano.
Pre-industrial pilot sites
10
Pre-industrial pilot sites
Pech RougeSelvapiana
Castello di Ama
11
2,5 years of running at Montepaldi 198 Sensors deployed 8 Hectares of vineyard monitored 4 GPRS Gateway (operating in different countries) Multihop and power saving strategies adopted 11 months of guaranteed battery life
Some Figures
12
New commercial pilot sites
13
Downey Mildew Risk:Apply Treatments
Water Deficit:Irrigate.Powdery Mildew Risk:
Apply Treatments
The macro site analysis approach
14
Conclusions
An End-To-End System
2,5 millions of data gathered
Application Driven Approach to WSN: Results
Site-specific Management
Enhanced environment knowledge
Real time collaboration of different experts on macro-areas
Macro-Site knowledge using detailed analysis of micro areas
…
It’s just the beginning!
15
Acknowledgments
The authors gratefully acknowledge many contributions by their colleagues in the GoodFood project, in particular the Groups of :
-Scuola Superiore Sant’Anna Italy-Synapsis Italy -HTA Switzerland-Tyndall Ireland -EADS Germany
The support from the European Commission is gratefully acknowledged.
16
…thanks for your attention!
17
ID PhaseHop
counterSequence Number
MAC PDU header
STAR MACSynchronous Transmission Asynchronous Reception
1. Completely novel approach effectively joining:
• WISE MAC (phase based transmission scheduling)
• S-MAC (duty cycle)
18
• Applied algorithms:
• Static 1 reference next hop to the gateway is a priori defined for each node
• Partial Dynamic additional back up neighbors facing faults (turning off) or congestion are a priori defined for each node
• Full Dynamic number and reliability of the next hop is dynamically updated depending on the network status changing
• Routing table for an ordinary node
Static routing algorithm
Partial dynamic routing algorithm
Target Next hop
Hop counter
Phase Link quality
Battery level
Congestion level
X NX ϕX ηX BX CX Sink
Y NY ϕY ηY BY CY
Multihop Routing
19
Current Consumption in RX (Listen)
~ 12 mA
Current Consumption in Power Saving (Sleep)
~ 0.01 mA
Current Consumption in TX ~ 30 mA
Number of cycles per hour 4
Battery Capacity 7 Ah
Battery life 11 mesi
Max Radio Coverage 200 m
0 25 50 75 [m]
0 25 50 75 100 125 150 175 200 [m]
0 25 50 75 100 125 150 175 [m]
0 25 50 75 [m] Courtesy of HTA
STAR MAC Implementation